Ben Depoorter, Peter Menell, David Schwartz - Research Handbook On The Economics of Intellectual Property Law - Vol 1 - Theory and Vol 2 - Ana

RESEARCH HANDBOOK ON THE ECONOMICS OF
INTELLECTUAL PROPERTY LAW
DEPOORTER_V1_9781848445369_t.indd 1 30/07/2019 15:48

RESEARCH HANDBOOKS IN LAW AND ECONOMICS
Series Editors: Richard A. Posner, Judge, United States Court of Appeals for the Seventh Circuit
and Senior Lecturer, University of Chicago Law School, USA and Francesco Parisi, Oppenheimer
Wolff and Donnelly Professor of Law, University of Minnesota, USA and Professor of Economics,
University of Bologna, Italy
Edited by highly distinguished scholars, the landmark reference works in this series offer advanced
treatments of specific topics that reflect the state-of-the-art of research in law and economics,
while also expanding the law and economics debate. Each volume’s accessible yet sophisticated
contributions from top international researchers make it an indispensable resource for students
and scholars alike.
Titles in this series include:
Research Handbook on the Economics of Property Law

Edited by Kenneth Ayotte and Henry E. Smith
Research Handbook on the Economics of Family Law

Edited by Lloyd R. Cohen and Joshua D. Wright
Research Handbook on the Economics of Antitrust Law

Edited by Einer R. Elhauge
Research Handbook on the Economics of Corporate Law

Edited by Brett McDonnell and Claire A. Hill
Research Handbook on the Economics of European Union Law

Edited by Thomas Eger and Hans-Bernd Schäfer
Research Handbook on the Economics of Criminal Law

Edited by Alon Harel and Keith N. Hylton
Research Handbook on the Economics of Labor and Employment Law

Edited by Michael L. Wachter and Cynthia L. Estlund
Research Handbook on Austrian Law and Economics

Edited by Todd J. Zywicki and Peter J. Boettke
Research Handbook on Behavioral Law and Economics

Edited by Joshua C. Teitelbaum and Kathryn Zeiler
Research Handbook on the Economics of Intellectual Property Law

Volume 1: Theory
Edited by Ben Depoorter and Peter S. Menell

Volume 2: Analytical Methods
Edited by Peter S. Menell and David L. Schwartz

Research Handbook on the
Economics of Intellectual
Property Law
Volume 1: Theory
Edited by
Ben Depoorter
University of California, Hastings College of Law, USA
and Ghent University, Belgium
Peter S. Menell
University of California, Berkeley School of Law, USA
Cheltenham, UK • Northampton, MA, USA

© The Editors and Contributors Severally 2019
All rights reserved. No part of this publication may be reproduced, stored in a retrieval system or
transmitted in any form or by any means, electronic, mechanical or photocopying, recording, or
otherwise without the prior permission of the publisher.
Published by
Edward Elgar Publishing Limited
The Lypiatts
15 Lansdown Road
Cheltenham
Glos GL50 2JA
UK
Edward Elgar Publishing, Inc.

William Pratt House
9 Dewey Court
Northampton
Massachusetts 01060
USA
A catalogue record for this book

is available from the British Library
Library of Congress Control Number: 2019945636
This book is available electronically in the

Law subject collection
DOI 10.4337/9781789903997
ISBN 978 1 84844 536 9 (2 volume set)

ISBN 978 1 78990 399 7 (eBook)
Typeset by Servis Filmsetting Ltd, Stockport, Cheshire

06

Contents
List of contributorsvii
PART I INTELLECTUAL PROPERTY AS PROPERTY
1 Intellectual property as property 2

Molly Shaffer Van Houweling
2 Anticommons, transaction costs, and patent aggregators 27
Rebecca S. Eisenberg
3 Governing intellectual property 47
Henry E. Smith
PART II IP AND INCENTIVES
4 Philosophical foundations of IP law: the law and economics paradigm 72

Robert P. Merges
5 Intellectual property law and the promotion of welfare 98
Christopher Buccafusco and Jonathan S. Masur
6 Economic models of innovation: stand-alone and cumulative creativity 119
Peter S. Menell and Suzanne Scotchmer
7 Economic analysis of network effects and intellectual property 157
Peter S. Menell
8 Intellectual property and competition 231
Herbert Hovenkamp
9 Intellectual property and the economics of product differentiation 262
Christopher S. Yoo
10 Price discrimination and intellectual property 281
Michael J. Meurer and Ben Depoorter
11 When are IP rights necessary? Evidence from innovation in IP’s
negative space 309
Kal Raustiala and Christopher Jon Sprigman
12 Open innovation and ex ante licensing 330
Michael J. Burstein
13 Prize and reward alternatives to intellectual property 350
Michael Abramowicz

vi Research handbook on the economics of IP law volume 1
PART III IP COSTS
14 Tailoring intellectual property rights to reduce uniformity cost 377

Michael W. Carroll
15 Intellectual property enforcement costs 407
Ben Depoorter
16 Economic analysis of intellectual property notice and disclosure 424
Peter S. Menell
PART IV IP AND INSTITUTIONS
17 Patent institutions: shifting interactions between legal actors 473

Arti K. Rai
18 The economics of collective management 489
Daniel Gervais
19 ‘The common law’ in the law and economics of intellectual property 508
Shyamkrishna Balganesh
20 In the shadow of the law: the role of custom in intellectual property 526
Jennifer E. Rothman
21 Infrastructure theory and IP 551
Brett Frischmann
PART V IP, DEVELOPMENT, AND INTERNATIONAL TRADE
22 Creative development: copyright and emerging creative industries 582

Sean A. Pager
23 Intellectual property and economic development: a guide for scholarly
and policy research 636
Shubha Ghosh
24 Economic development and intellectual property rights: key analytical
results from economics 656
Keith E. Maskus
Index677

Contributors
Michael Abramowicz, Professor of Law, George Washington University, USA

Shyamkrishna Balganesh, Professor of Law, University of Pennsylvania Law School, USA
Christopher Buccafusco, Professor of Law and Director of the Intellectual Property and
Information Law Program, Cardozo School of Law, Yeshiva University, New York, USA
Michael J. Burstein, Professor of Law, Cardozo School of Law, Yeshiva University, New
York, USA
Michael W. Carroll, Professor of Law and Director of the Program on Information Justice
and Intellectual Property, American University Washington College of Law, USA
Ben Depoorter, Max Radin Chair and Distinguished Professor of Law, University of
California, Hastings College of the Law and Affiliate Scholar, Stanford Law School,
Center for Internet and Society, USA; CASLE, Ghent University, Belgium
Rebecca S. Eisenberg, Robert and Barbara Luciano Professor of Law, Michigan Law,
University of Michigan, USA
Brett Frischmann, Charles Widger Endowed University Professor in Law, Business and
Economics Charles Widger School of Law, Villanova University, Villanova, Philadelphia,
USA
Daniel Gervais, Milton R. Underwood Chair in Law and Director of the Vanderbilt
Intellectual Property Program, Vanderbilt University Law School, USA
Shubha Ghosh, Crandall Melvin Professor of Law and Director, Syracuse Intellectual
Property Law Institute (SIPLI) and IP & Tech Commercialization Curricular Law
Program, Syracuse University College of Law, USA
Herbert Hovenkamp, James G. Dinan University Professor, Penn Law and the Wharton
School, University of Pennsylvania, USA
Keith E. Maskus, Arts and Sciences Professor of Distinction, Economics Department,
University of Colorado, Boulder, USA
Jonathan S. Masur, John P. Wilson Professor of Law and David and Celia Hilliard
Research Scholar at the University of Chicago Law School, USA
Peter S. Menell, Koret Professor of Law and Director of the Berkeley Center for Law and
Technology, Berkeley School of Law, University of California, USA
Robert P. Merges, Wilson Sonsini Goodrich and Rosati Professor of Law and Director
of the Berkeley Center for Law and Technology, University of California, Berkeley, USA
Michael J. Meurer, Abraham and Lillian Benton Scholar and Professor of Law, Boston
University School of Law, USA
vii

viii Research handbook on the economics of IP law volume 1
Sean A. Pager, Professor of Law and Associate Director of the Intellectual Property,
Information and Communications Law Program, Michigan State University College of
Law, USA
Arti K. Rai, Elvin R. Latty Professor of Law, Duke Law School, Faculty Director of the
Duke Law Center for Innovation Policy, and Duke Innovation and Entrepreneurship
Initiative Research Fellow, USA
Kal Raustiala, Professor and Director of the UCLA Ronald W. Burkle Center for
International Relations, UCLA Law School, USA
Jennifer E. Rothman, Professor of Law and Joseph Scott Fellow, Loyola Law School,
Loyola Marymount University, Los Angeles, USA
Suzanne Scotchmer, formerly Professor of Economics, Public Policy, and Law, University
of California, Berkeley, USA, who died in 2014
Henry E. Smith, Fessenden Professor of Law and Director of the Project on the
Foundations of Private Law, Harvard Law School, USA
Christopher Jon Sprigman, Professor, New York University School of Law and Co-Director
of the Engelberg Center on Innovation Law and Policy, USA
Molly Shaffer Van Houweling, Harold C. Hohbach Distinguished Professor of Patent
Law and Intellectual Property, Director of the Berkeley Center for Law and Technology,
and Associate Dean for J.D. Curriculum and Teaching, University of California, Berkeley,
USA
Christopher S. Yoo, John H. Chestnut Professor of Law, Communication, and Computer
and Information Science, University of Pennsylvania, USA

PART I
INTELLECTUAL PROPERTY
AS PROPERTY

1. Intellectual property as property
Molly Shaffer Van Houweling* 1
Contents
I. Introduction
II. ‘Property’ and IP
III. Distinguishing Tangible and Intellectual Resources as Objects of Property
A. Rivalry
B. Excludability
C. Costs
IV. Three Cross-Cutting Themes
A. Property and Possession
1. Possession, property origins, and the public domain
2. The challenges of non-possessory property
B. Property and Information
C. Property and Time
V. Conclusion
References
I. INTRODUCTION
First-year law students learn early on that lawyers think of property not ‘as a relationship
between a person (the owner) and a thing (that is owned)’ but rather as ‘relationships
among people with respect to things’ (Dukeminier et al., 2014, p. 51, n. 33). This corrective
appears perhaps to reorient legal thinking away from a layperson’s preoccupation with
things and toward a more sophisticated focus on people and their legal relations (see Grey,
1980, for an extreme example of this view). But, in fact, what makes property law distinc-
tive—in both its lay and expert formulations—is that the human relationships it governs
(unlike the human relationships governed by the law of torts or contracts) are always
mediated by things (see Smith, 2012). That these things carry legal implications with them
* Harold C. Hohbach Distinguished Professor of Patent Law and Intellectual Property;

Associate Dean, J.D. Curriculum and Teaching, University of California, Berkeley. Thanks to
James Hicks for excellent research assistance. This chapter is licensed under the terms of the
Creative Commons Attribution-NonCommercial 4.0 International (CC BY-NC 4.0) license, with
attribution to Molly Shaffer Van Houweling as the author and to the original publication venue:
Peter S. Menell and Ben Depoorter, eds. 2019. Research Handbook on the Economics of Intellectual
Property Law. Cheltenham, UK and Northampton, MA, USA: Edward Elgar Publishing. The
terms of the CC BY-NC 4.0 license are available at https://creativecommons.org/licenses/by-nc/4.0/
(last accessed March 19, 2019).

Intellectual property as property 3
yields the notion that property is in rem, not in personam. That these legal implications
can affect anyone who encounters the thing means that they are good against the world.
The things to which property rights attach are the locus of their owners’ rights to exclude.
The alert law student will also note that the specific nature of a thing governed by
property law can have consequences for the design of that law. The rules of acquisition
by capture might sensibly differ depending on whether the thing at issue is a fox or a
whale. The strength of the right to exclude might differ depending on whether the thing
from which outsiders are excluded is a home or a shopping mall. The differences can be
even more dramatic when the ‘thing’ at issue is not a tangible object at all, but rather an
intangible work of authorship or invention.
This chapter will explore how important aspects of property jurisprudence apply to
the fields of law that have come to be known as ‘intellectual property’ or ‘IP’—focusing
primarily on copyright and patent law.1 I aim to illustrate both the relevance of enduring
property themes to these areas of law and the ways in which the nature of the intangible
things governed by IP should force us to resist easy analogies to tangible property. I start
by tackling the contentious question of whether IP should be considered property at all,
in light of key differences between tangible resources and intellectual works. Finding the
property frame a useful one despite these differences, I proceed by exploring three themes
that have been important to the legal and economic analysis of tangible property, and that
can usefully be brought to bear on IP: the role of possession in establishing and signaling
property rights; the problem of information costs related to property rights; and, finally,
the relationship between property rights and time. In each of these cases, careful attention
to concepts that inform the design of sensible tangible property rules can help to inform
scholarship and policymaking about IP rules as well.
II. ‘PROPERTY’ AND IP
The US Constitution authorizes Congress to ‘promote the Progress of Science and

useful Arts, by securing for limited Times to Authors and Inventors the exclusive Right
to their respective Writings and Discoveries’ (U.S. Const. art. I, § 8, cl. 8). This consti-
tutional authorization for our federal copyright and patent laws is often referred to as
the ‘Intellectual Property Clause’ (Walterscheild, 1994).2 Indeed, the term ‘intellectual
property’ is now ubiquitous in discussions about patent, copyright, and related fields of
law. But the term is controversial.3
Advocates of strong rights for authors and inventors often characterize copyrights and
patents as property rights. They tend to use that characterization to argue that these rights
should be stronger and more vigorously enforced (e.g. Epstein, 2010; Epstein, 2005; Mossoff,
2005; Easterbrook, 1990). Others reject the property characterization, emphasizing that
1
This chapter builds upon and incorporates some of my earlier work at the intersection of
tangible and intangible property (Van Houweling, 2002, 2007, 2008, 2010, 2011, 2012a, 2012b,
2013, 2016a, 2016b, 2017).
2
On the history of the term ‘intellectual property,’ see Lemley (2005) and Hughes (2012).
3
For discussion of commentary critiquing and embracing the idea of IP as property, see gener-
ally Liivak and Peñalver (2013).

4 Research handbook on the economics of IP law volume 1
copyrights and patents are, and should be, more limited than the rights of owners of tangi-
ble objects; and resisting the expansion of property-based arguments beyond copyright and
patent to trademark and trade secret law (e.g. Lemley, 2005; Lemley, 1999; Lunney, 1999;
Samuelson, 1989). Advocates and scholars in this group worry about the rhetorical force of
property-based arguments, sensing that they tend to prompt facile thinking about absolute
dominion (e.g. Lemley, 1997; Bessen and Meurer, 2008; Vaidhyanathan, 2001).
A third group embraces the property paradigm but emphasizes that property rights are
always contingent, contextual, and shaped by policy concerns about market failure and
distributive justice. A nuanced view of property, according to these scholars, can enhance
rather than distort conversations about rights in intangible things (e.g. Carrier, 2004;
Chander, 2003; Cohen, 2014; Ghosh, 2007; Liivak and Peñalver, 2013; Menell, 2007b,
2011; Merges, 2011; Oliar and Stern, 2019).
Scholars of tangible property have entered this conversation as well. They tend, like
the third group of IP scholars, to offer a view of property that does not get carried away
with notions of absolute dominion. Interestingly, some of them look to IP as a model
from which tangible property law can borrow mechanisms for limiting property rights in
tangible things (e.g. Dagan, 2012; Depoorter, 2011; Singer, 2014a; Singer, 2014b).
My own view, from the perspective of a scholar working at the intersection of tangible
and intangible property, is that the law of tangible property can be an important source of
insights about both the benefits and costs of granting people rights to control the use of
valuable resources, and about the various ways those rights and corresponding remedies
can be structured (see Van Houweling 2002, 2007, 2008, 2010, 2011, 2012a, 2012b, 2013,
2016a, 2016b, 2017).
This chapter highlights some of the many insights from tangible property law and
doctrine that can helpfully be applied in the field that has come to be known as intellectual
property. This application should of course be done with care. To properly apply lessons
from tangible property law to IP requires assiduous attention to the special characteristics
of the things—intangible creations and inventions—to which this body of property law
applies, and to the specific environments in which IP rights are exercised.
III. DISTINGUISHING TANGIBLE AND INTELLECTUAL

RESOURCES AS OBJECTS OF PROPERTY
The most widely accepted rationale for property rights is that they promote efficient use
of, and investment in, resources by internalizing the costs and benefits of that use and
investment (Demsetz, 1967). This is a corollary of the tragedy of the commons, the notion
that in the absence of property rights everyone with access to a resource will use it without
regard to the full costs of that use while failing to make investments that would benefit
other users (Hardin, 1968). This logic is most compelling when applied to resources
that are scarce (thus making overuse problematic), that are difficult to produce and/or
maintain (thus making underinvestment likely), and for which self-help cannot easily be
deployed to avoid overuse and undercultivation (thus justifying the costs of establishing
and enforcing property rights). On each of these dimensions, there are relevant differences
between tangible resources and intellectual creations that should be kept firmly in mind
when using a tangible property frame to guide understanding of IP.

A. Rivalry
Start with scarcity and the specter of overuse. Tangible resources are generally rivalrous:
if fully consumed by one person, they cannot be consumed by another. Their stocks are
therefore subject to depletion if consumption is left unchecked. Intellectual resources, by
contrast, tend to be nonrivalrous (see, e.g., Boyle, 2003). That is, their consumption by
one person does not interfere with their consumption by others; it does not contribute
to scarcity. This distinction leads many commentators to argue that IP rights are on
shakier normative ground than tangible property rights, because they cannot be justified
by the overuse aspect of the tragedy of the commons (e.g. Breyer, 1970; Carrier, 2004;
Ghosh, 2007; Lemley, 2005; Menell, 2007a, 2011; Rose, 2003; Sterk, 1996) and because,
as Henry Smith observes, ‘excluding others from information when they could use it at
zero marginal cost seems wasteful’ (Smith, 2009b, p. 2084).
This distinction and its normative implications should not be overstated, however (as
noted by Nachbar, 2007). On the one hand, some intellectual resources can become less
valuable when overused. Trademarks cannot achieve their function of reducing consumer
confusion if they are used indiscriminately on different products from different produc-
ers. So too, the value of a celebrity endorsement could be undermined by indiscriminate
association with many goods. These are intangible objects of IP protection the value of
which could be diminished by overuse (see Landes and Posner 2003, pp. 485–7). Looking
to the other side of the comparison, tangible resources are not always rivalrous as a
practical matter. They can often be enjoyed simultaneously by multiple users who neither
interfere with each other nor diminish the availability of the resource for future users (see
Rose, 1986).
Across resource types, nonrivalry undermines the logic of the tragedy of overuse. In
the tangible realm as well as the intellectual realm, commons can produce comedy (to use
Carol Rose’s (1986) term) instead of tragedy. As I will explain below, this helps to explain
doctrinal rules that promote the growth of the unowned public domain.
B. Excludability
Property rights are often lauded as promoting investment in resources: they help investors
secure rewards on their investments by excluding others from the benefits of those invest-
ments. It is often especially difficult for investors in intangible works to exclude other
people. This characteristic is referred to as non-excludability, and is often offered as an
especially powerful justification for IP.
Note, however, that the starkness of the term non-excludability can be misleading (as
noted by Parchomovsky and Siegelman, 2002). It is not as difficult to exclude people
from enjoying the benefits of a work of creativity or invention as it is to exclude people
from enjoying the benefits of a lighthouse, the classic nonexcludable public good. The
author of a book could keep the book in her reading room, for example, to be read only
by the people to whom she grants permission to enter.4 This rather ungainly type of
4
It was common during the Middle Ages for monasteries to charge fees for permission to
copy manuscripts in their collections. Once a manuscript was copied, its owner lost control of

exclusion has its drawbacks, of course. It limits the author’s ability to realize a return on
her investment to the meager returns from her reading room invitees; and it limits the
number of readers who can benefit from her work. The heart of the problem is not so
much that intellectual works are nonexcludable. It is that exclusion is difficult to combine
with wide dissemination, and therefore with extracting the full benefits of intellectual
investments for both the author and the public (see generally Arrow, 1962). On the
other side of the comparison, the classic lighthouse example reveals that the benefits of
investing in tangible resources can also be nonexcludable. Both tangible and intellectual
resources can present a public goods problem in which we worry that the market will
not provide sufficient incentives to invest in a resource. In both cases, granting private
property rights is a possible solution to the problem, albeit an imperfect one that entails
its own costs and should be considered in light of alternatives (see generally Kapczynski,
2012).
C. Costs
There are also several distinctions between tangible and intellectual resources that are
relevant to the costs of establishing and maintaining a property system (see generally
Menell, 2011). One is the relative difficulty of delineating the boundaries that are
important to the healthy functioning of a legal system premised on the right to exclude.
Intellectual creations are notoriously difficult to define in ways that give clear notice that
can help promote transactions and avoid inadvertent trespasses (Smith, 2009b). On this
front, rights that attach to intellectual resources are similar to non-possessory rights that
attach to tangible resources—think of servitudes and future interests. These rights can be
similarly difficult to notice and comprehend in that they do not correspond to possession
of a physical parcel or tangible object but rather interfere with the rights of people in
possession (Van Houweling, 2008).
The costs of property rights also include the distributive injustices perpetrated by the
allocation of resources based on ability to pay (see generally, e.g., Ghosh, 2007; Van
Houweling, 2005; Chander, 2003). Of course, neither property rights in tangible nor
intellectual resources necessarily require or produce market-based allocation of those
resources, at least as an initial matter. But free alienability is one hallmark of private
property (albeit subject to important exceptions as noted by, e.g., Calabresi and Melamed
(1972), Fennell (2009), and Radin (1987)). Thus property tends overwhelmingly to be
distributed via the price mechanism. Here, tangible and intellectual property share the
potential for tragic consequences when the market denies basic necessities to those unable
to pay. But the tragedy is especially pronounced for nonrivalrous resources that could be
distributed to everyone at little or no marginal cost (Kapczynski, 2012).
Both tangible and intellectual property rights affect the distribution of physical
resources, some of which are embedded with intellectual content (e.g. patented tools and
medicine, books). IP rights are noteworthy in their potential to affect the distribution
the text embodied in it (Rose, 1995). This proto-copyright was valuable, however, in an age before
mechanical reproduction, when an owner could charge a premium based on the superior quality of
his manuscript compared to error-ridden copies (Van Houweling, 2010).

of inventive and creative opportunities as well, insofar as IP rights can limit cumulative
invention and creativity by those who do not acquire permission to build upon what has
come before. Once again, the distinction can be overstated. In the tangible realm, too,
distribution of resources can dramatically impact opportunities for creativity, invention,
and all other aspects of self-realization.
IV. THREE CROSS-CUTTING THEMES
Despite their differences, tangible and intellectual resources are both managed—in
part—by legal institutions that can usefully be characterized as property regimes. The
characteristics that make these regimes property—in rem rights that are good against the
world and allow their owners to exclude others from valuable resources—pose some recur-
ring questions and challenges that cut across the resource realms in which they operate. I
proceed by exploring these questions and challenges as embedded in three themes: the role
of possession in establishing and signaling property rights; the problem of information
costs related to property rights; and, finally, the relationship between property rights
and time. These themes have long been important to the legal and economic analysis of
property law. All can usefully be brought to bear on IP as well, with the distinctions above
kept carefully in mind.
A. Property and Possession
1. Possession, property origins, and the public domain

Possession operates as both a source and a signal of property rights in land and physical
objects. The importance of possession has been traced back at least to Roman law and
across legal traditions around the world (Fraley, 2011; Lueck, 1995); it has been explained
as psychologically embedded (Barros, 2011; Friedman and Neary, 2009; Merrill and
Smith, 2007) and even the product of evolution (Stake, 2004). Ambiguity about the
meaning of possession is likely as old and widespread (Drassinower, 2006; Rose, 1985;
Smith, 2003).
Theorists have cited a number of different rationales in an effort to help explain the
importance and meaning of possession as applied to particular property controversies.5
Some of these rationales are particularly helpful for understanding and perhaps improv-
ing the mechanisms by which authors and inventors seize ‘possession’ of the intangible
subject matter of IP.
One rationale for possession as an origin of property rights is that acquiring possession
typically requires labor, which is both a moral justification for property rights from a
Lockean labor-desert perspective and also worth incentivizing in order to promote the
productive use of resources from a utilitarian perspective. A difficulty that often arises
5
The classic case of Pierson v. Post, 3 Cai. 175 (N.Y. 1805) is the most typical case study.
The competing views offered by the majority, dissent, and innumerable commentators deploy
arguments emphasizing labor, investment, notice, custom, and more, in an effort to explain what
constitutes possession adequate to establish initial property rights in a wild fox (see, e.g., Rose,
1985; Epstein, 1979).

in the tangible property context is that multiple people can claim that they labored in a
deserving and socially productive way to reduce something to possession, but the rule of
first possession aims to declare only one of them the winner. As Richard Epstein notes,
‘[s]ome labor goes unrequited when two pursue and one loses’ (1979, p. 1225). And yet,
a rule that does its best to decide between two competing pursuers can be justified in the
tangible property context by the benefits of individual ownership, and by the difficulty of
fairly dividing the fruits of a joint pursuit. In his qualified defense of the first possession
rule, Epstein contrasts the rule with what he sees as the only alternative, original common
ownership coupled with a system of public control to ‘decide how the rights in question
are to be packaged and divided amongst individuals’ (1979, p. 1239). In light of the chal-
lenges of establishing such a system and the potential for its abuse, Epstein argues that
‘[o]n balance the case tilts strongly for the first possession theories, whatever their infirmi-
ties’ (1979, p. 1238).
How do these insights translate to IP? Here too, rules that award ownership on the basis
of original acquisition would appear to reward and incentivize socially beneficial activity.
But in the IP context there are special difficulties with identifying the ‘things’ subject
to original acquisition and the acts that amount to possession. The intangible subject
matter of copyright and patent protection is quite different from the parcels of land and
wild animals that provide classic examples of the first possession rule in action. And yet
there are clear analogs to first possession for purposes of acquiring rights in intellectual
creations.6
Consider two examples: an original musical melody and a newfangled (i.e., novel and
nonobvious) mousetrap. These are the types of creative and inventive works that copy-
right and patent law should incentivize. They required mental labor and were perhaps the
result of costly failed experiments. At what point are these intangible creations ‘possessed’
in a way that should trigger a rule based on first possession? Arguably the mental labor
that we want to reward and incentivize has already been expended by the time the author
and inventor have captured the melody and mousetrap in their brains. But these are not
the rules of first possession that IP law has adopted. In order to qualify for protection, the
melody would have to be fixed in a tangible medium of expression (e.g. the notes written
down or captured in a sound recording) (17 U.S.C. § 102). The mousetrap would have to
be either actually or constructively ‘reduced to practice’ (e.g. Ariad Pharmaceuticals, Inc.
v. Eli Lilly & Co., 598 F.3d 1336, 1352 (2010)).
In some ways, these IP requirements are fully consistent with the rules and rationales
of first possession that emerge from the tangible property context. These rules arguably
require the type of possession that makes a resource available for a socially beneficial
purpose. Merely giving chase to a fox—as laborious as that may be—does not result in
ownership if the fox is still alive and hunting chickens. Similarly, a melody or mousetrap
that exists merely in one individual’s head does not yet benefit society. In theory, copyright
and patent both withhold their rewards until the intangible work is at least capable of
being communicated to and used by other people. Where in practice they do not—for
example, where patents are awarded based on constructive reduction to practice that does
6
On the connection between possession and intellectual property, see, e.g., Yen (1990);
Drassinower (2006); Holbrook (2009, 2006, 2016); Oliar and Stern (2019); Smith (2007).

not in fact establish that the invention works—they are subject to critique (e.g. Lemley,
2016) that very much echoes arguments regarding the proper definition of possession of
tangible property.
In patent, something else is required in addition to reduction to practice. To establish
patent rights, an inventor must file a timely patent application that discloses how to
practice her invention (35 U.S.C. § 112). Patent’s disclosure requirements are consistent
with a utilitarian rationale for first possession, under the theory that the public does
not receive the full benefit of patentable inventions unless their details are adequately
disclosed (Kewanee Oil Co. v. Bicron Corp., 416 U.S. 470 (1974)). By requiring that
the inventor describe her invention in sufficient detail to establish that she has in fact
captured it in her own mind, disclosure requirements also provide objective evidence
that she has in fact done sufficient intellectual labor (not just abstract daydreaming) to
merit reward and encouragement (Holbrook, 2006). Note, however, that not just any
disclosure will do. Disclosure via a timely patent application is required. By contrast,
disclosure via a printed publication, public use, or public sale does not establish patent
rights. Just the opposite: these disclosures by the inventor deposit the invention into the
public domain unless the inventor-discloser files a patent application within a year (35
U.S.C. § 102(a)–(b)).
Does anything within the theory of first possession justify denying property rights to
an inventor who has captured his invention by reducing it to practice and delivered its
benefits to the public by disclosing it? Without more, the theory of possession as labor
and productive use does not seem adequate to the task. Indeed, this IP example appears
to confirm Epstein’s observation about the inadequacy of such theories: just as the unsuc-
cessful chaser expended reward-worthy labor but was denied the fox, the patent-barred
inventor expends reward-worthy ingenuity but is denied his IP.
Both the tangible and IP examples suggest that the type of possession that triggers
property rights must involve more than productive labor, which suggests that another
rationale is at work. One candidate is notice: possession can provide clear notice of an
owner’s claim. This idea is central to Carol Rose’s persuasive explanation in Possession as
the Origin of Property, in which she observes that:
[p]ossession as the basis of property ownership . . . seems to amount to something like yelling
loudly enough to all who may be interested. The first to say, ‘This is mine,’ in a way that the
public understands, gets the prize, and the law will help him keep it against someone else who
says, ‘No, it is mine.’ (1985, p. 81)
For patentable inventions, merely disclosing an invention to the public via a public use,
sale, or printed publication provides some notice—that is, notice regarding the nature of
the invention and the inventor’s intellectual possession of it. As Holbrook (2006) notes:
[D]emonstrating the possession of an intangible idea is difficult. One could describe an idea but
not necessarily truly be in possession of it. For example, the idea of teleportation has existed
in science fiction, such as in Star Trek, for some time. Simply having the idea of teleportation,
however, does not mean that those authors are in possession of a teleportation device. Instead,
the key aspect of possession is whether or not the author can actually make a functioning
device. Thus, the best evidence of possession would be either the inventor physically creating
the invention or, at least, providing a description that is clear enough to enable someone else to
build it (2006, p. 146–7).

Unless that description comes in the form of a patent application, however, it is insuf-
ficient to establish property rights in the invention. Why must an invention hunter, unlike
a fox hunter, both demonstrate that she has grasped her prey and submit a bunch of
paperwork to establish her claim to it?7 Here is an instance in which the distinct nature
of the thing that is the subject matter of property has important consequences. Because of
the nonrivalrous nature of an invention, the inventor’s possession of it does not prevent
others from encountering it ‘in the wild.’ Another inventor might conceive of the same
mousetrap idea; or he might encounter one of many coexisting physical embodiments
of the invention. In either case, he may not recognize that the invention is the subject
of patent rights unless there is some indication—either on the embodiment itself or in a
centralized registry—that it has been claimed. For physical objects possession alone may
be enough to communicate the signal that ‘this is mine!’ For intangible inventions that can
be possessed by many people at the same time, something else is required.
This is not to say that the functional equivalents of possession that we observe in the
IP context—including both requirements that the IP owner has captured her intangible
creation in a concrete way and has signaled her claim to others—are sufficient to serve
possession’s purposes. The failure of patents to provide adequate notice to rival inventors
is a notorious shortcoming of the contemporary patent system (see, e.g., Bessen and
Meurer, 2008; Fromer, 2009; Long, 2004; Menell and Meurer, 2013). As for copyrights,
the requirement that claims to works of authorship be communicated to the public via
a centralized registry or marked embodiments has been abandoned (see, e.g., Ginsburg,
2010; Litman, 2016; Sprigman, 2004). Now fixation of an original work of authorship is
all that is required, and works are eligible for the longest possible duration with no renewal
paperwork (Samuelson, 2016). Even where the existence and ownership of a copyright is
clear, its boundaries are subject to the vagaries of the idea/expression dichotomy, fair use,
and other aspects of copyright scope and defenses (Fromer, 2009; Liu, 2016; Long, 2004;
Menell, 2016). Works of authorship are even easier to claim than foxes, but copyright
claims are much more difficult to observe.
Many critics of contemporary copyright law lament this development and advocate
for policy changes that would require (or at least more strongly incentivize) copyright
notice and registration (e.g., Gibson, 2005; Landes and Posner, 2003; Samuelson et al.,
2010; Samuelson, 2016; Sprigman, 2004). These calls are consistent with the notice-based
rationales that property theorists have offered for possession as a trigger for tangible
property rights; they also recognize that the unique characteristics of intangible property
might require different mechanisms of notice provision.
For the sake of argument, assume that these critics are right to argue that copyright
(like patent) should require some form of ‘claiming’ that provides notice to the public
in a way roughly equivalent to physical possession of tangible objects.8 What should the
consequences of a failure to claim property be? In tangible property, the consequences
7
Note that paperwork is not irrelevant in the tangible property context, as Rose (1985) notes
in her discussion of possession as notice. But recording of possession-based claims to tangible
property is not generally required to establish them in the first instance, as discussed in Van
Houweling (2013).
8
On the different claiming requirements of patent and copyright law, see generally Fromer
(2009); Oliar and Stern (2019).

are typically that a rival claimant has an opportunity to become a property owner. Post
failed to capture the fox, but the ‘saucy intruder’ Pierson succeeded.9
By contrast, in the copyright and patent contexts the analogs to first possession are
not merely questions of who was first (or has anyone been first yet)—that is, who was
first as between two ‘would-be’ owners, or when is the thing at issue still unpossessed
and subject to capture and ownership by someone new? In the IP context, unlike in most
tangible property contexts, it is possible for a failed or incomplete attempt at possession to
make a work of authorship or invention forever unownable by anyone. Under US patent
law, an invention that has been in public use for more than a year cannot thereafter be
patented—by the inventor or anyone else (17 U.S.C. § 102). It has permanently entered the
public domain. For the first two-plus centuries of US copyright law, a similar situation
obtained: works that were published without proper notice entered the public domain and
could not thereafter be captured for copyright purposes by their authors or anyone else
(Samuelson, 2016; Sprigman, 2004).
Can these IP doctrines be squared with first possession theory? Indeed they can, if we
are attentive to the special qualities of intangible works. Consider Epstein’s defense of first
possession: it relies on the assumption that it is desirable for tangible objects ultimately
to be owned by someone. If the identity of that someone were not determined based
on who was the first to possess an object, then the government would have to take over
and figure out some more heavy-handed way to divvy things up. Why is this divvying
necessary? Because of the tragedy of the commons, so the theory goes. Things without
owners are unlikely to be carefully husbanded; they are instead likely to be overused. By
contrast, once an intellectual work has been created, its nonrivalrous nature means it is not
subject to dissipation.10 It therefore makes perfect sense that in the IP context the result
of a failure adequately to possess something (and clearly signal that possession in a way
that communicates an ownership claim) does not necessarily make that thing subject to
ownership by one’s rivals, but might instead deliver it to the public domain. This explana-
tion also helps to make sense of some of those tangible property contexts in which we do
find a permanent public domain: they involve resources that are relatively nonrivalrous,
where the risk of tragic overuse is low (Rose, 2003, p. 96).
Note that not all forms of IP have the feature of falling inexorably into the public
domain when inadequately claimed. For example, abandoned trademarks can be
reclaimed by new owners who use them in commerce as source identifiers (which is the
rights-triggering analog to possession in the trademark context). So too for potential
trademarks that were unsuccessfully claimed by merchants who had not used them
enough to qualify for protection. In these ways trademarks are treated like nearly captured
foxes, where patents and copyrights are not. This too makes sense if we keep in mind key
distinctions on the dimension of rivalrousness. Trademarks are unusual within the IP
realm in the degree to which they are subject to congestion externalities—that is, they are
subject to overuse. Although it is possible for many people to use the same trademark,
if they do it is likely to lose its value as a trademark. Trademarks are effectively rivalrous
9
See n. 5.
10
But see Landes and Posner (2003, p. 487), arguing that the value of a creative work might be
depleted by overuse.

resources. Their value can be maintained (and rebuilt) only if they are kept out of the
public domain of indiscriminate use.
In sum, looking to the role of possession in the law of tangible property helps to explain
analogous mechanisms of rights acquisition in the realm of IP, and to diagnose their
shortcomings. Perhaps more interesting, it also helps to explain a feature that is especially
associated with (some forms of) IP: the permanent public domain.
2. The challenges of non-possessory property

Although possession is a touchstone for the acquisition and signaling of property rights,
property and possession do not always go hand in hand. Consider the challenges posed
for both tangible and IP regimes when property rights are held by non-possessors.11
These challenges arise in the tangible property context when, for example, the claims of
current possessors are challenged by prior possessors—with the results often reflecting the
enduring notion of ‘first in time, first in right,’ but sometimes prioritizing recent claims
over stale ones (as with adverse possession, marketable title acts, and protection for bona
fide purchasers). More striking, perhaps, are controversies over property rights held by
people who may never have been in possession, but who nonetheless claim the right to
control some aspect of a resource. Land servitudes are a classic example, and the anxiety
and doctrinal complexity that has marked the body of law governing them (see French,
1982; Rose, 2011; Van Houweling, 2008) serve to reinforce the importance of possession
as a touchstone of property reasoning.
To see the relevance of these types of non-possessory property rights to IP, start with a
classic passage from Justice Holmes’ 1908 concurrence in White-Smith Music Publishing
Company v. Apollo Company, in which he considers that nature of copyright as a
property right:
The notion of property starts, I suppose, from confirmed possession of a tangible object and
consists in the right to exclude others from interference with the more or less free doing with it as
one wills. But in copyright property has reached a more abstract expression. The right to exclude is
not directed to an object in possession or owned, but is in vacuo, so to speak. It restrains the spon-
taneity of men where but for it there would be nothing of any kind to hinder their doing as they
saw fit. It is a prohibition of conduct remote from the persons or tangibles of the party having
the right. It may be infringed a thousand miles from the owner and without his ever becoming
aware of the wrong. It is a right which could not be recognized or endured for more than a limited
time, and therefore, I may remark in passing, it is one which hardly can be conceived except as a
product of statute, as the authorities now agree. (209 U.S. 1, 19 (1908) (Holmes, J., concurring))
In drawing the contrast between copyright and paradigmatic possessory property rights
in tangible objects, Justice Holmes here emphasizes the non-possessory, ‘in vacuo’ nature
of copyright and the way in which copyright owners can control strangers from afar,
unconnected to any object possessed by the copyright owner. Copyright owners are thus
unlike owners of possessory fee simple interests in land, whose rights to exclude generally
impact the limited universe of people who come into contact with the physical boundaries
of the owner’s parcel.
11
This section draws on my previous work on non-possessory property rights (Van Houweling,
2002, 2007, 2008, 2010, 2011, 2012a, 2013).

Justice Holmes alludes to another apparent copyright anomaly: although copyright

owners are not necessarily possessors, the people whose spontaneity is restrained by
copyright are typically in possession of tangible objects—books, sheet music, or other
manifestations of the copyrighted work. As to these tangible objects, copyright operates
not as an instrument of freedom from interference for the possessor but rather the oppo-
site: an instrument of constraint operated by strangers (copyright owners) via remote
control. Copyright thus strikes Justice Holmes as an odd sort of property right in that
instead of liberating people to use their possessions it ‘restrains [their] spontaneity . . .
where but for it there would be nothing of any kind to hinder their doing as they saw fit’
(White-Smith Music Publ’g Co. v. Apollo Co., 209 U.S. 1, 19 (1908)).
Copyright owners’ power to control how remote strangers use objects in their posses-
sion is not as extraordinary as this passage suggests, however. Of course copyright shares
this characteristic with patent and trademark. Beyond IP, copyrights are similar in this
regard to a whole set of ‘remote control’ property interests that give their owners the right
to control use of assets possessed by other people. Servitudes are the most prominent
example.12
A servitude (which can take the form of an easement, real covenant, or equitable
servitude) is a non-possessory property interest that gives its holder the right to use an
asset (typically land) in specified ways, or to object to specified uses of it, or to insist on
specified behavior connected to it. The asset is encumbered by the servitude, such that
the servitude’s burdens ‘run with’ the asset, ‘pass[ing] automatically to successive owners
or occupiers’ (American Law Institute, 2000, § 1.1). Unlike a mere contractual agreement
to, say, refrain from operating a gas station in a residential neighborhood, a servitude
is enforceable against successors in interest. Therefore, if you grant your neighbor an
effective servitude she will be able to enforce the restriction against you and subsequent
owners of your land. The benefit of a servitude typically runs to successors as well—from
your neighbor to the next owner of her house. As Carol Rose puts it, ‘[t]he greatest overall
advantage of servitudes is that they give stability to property arrangements over both time
and space’ (Rose, 2011, p. 297). The stability that servitudes produce can be especially
valuable for land use planning. Land is, of course, immobile and enduring. It is therefore
often important for people who invest in land to be able to predict how surrounding land
will be used far into the future, in order to make investments that will coordinate rather
than conflict with adjacent activities.
In recognition of these benefits, land servitudes and other varieties of remote control
property rights have long been enforced by courts. Nonetheless, Justice Holmes’ contention
that property rights with such features could only be the product of statute rings somewhat
true. Judges have greeted most non-possessory property rights with suspicion and hemmed
them in with doctrinal limitations. Servitudes came to be classified into the three major
categories of easements, real covenants, and equitable servitudes, with each category
subject to convoluted rules limiting formation, subject matter, and enforceability (French,
1982). As I describe in prior work, tracing the evolution of modern servitude law reveals
several rationales for this type of hostility and the limiting doctrines that it produced.
12
Other examples include the future interests that accompany various types of defeasible
estates. See generally Korngold (1988).

I have organized these rationales into three broad categories: those related to notice and
information costs; those related to dead hand control and other aspects of the ‘problem of
the future’; and those related to harmful externalities (Van Houweling, 2008).13
Turning to IP, the basic anxiety and confusion associated with non-possessory property
rights is perhaps unavoidable (unless we abandon IP for some alternative regime), because
the structure of modern IP divorces ownership of physical objects embodying works
of authorship and novel inventions from ownership of the corresponding copyrights
and patents. This suggests that IP law should be acutely attentive to the problems that I
associate with non-possessory property rights. Indeed, IP has developed an entire body
of doctrine that attempts to mediate the tension between the rights of IP owners and
the rights of owners of physical objects embodying IP. This is known as IP ‘exhaustion,’
and has emerged as one of the most contested areas of IP law in an era in which more
and more objects of everyday life are burdened with non-possessory IP rights (see, e.g.,
Perzanowski and Schultz, 2016; Van Houweling, 2016a, 2016b).14
To better understand exhaustion and other limitations on remote control property
rights, I turn in the next section to one of the key categories of concerns I associate with
non-possessory property: notice and information costs.
B. Property and Information
Information costs include the costs of conveying and comprehending information and
also the costs imposed by failures to communicate. The monetary cost of a ‘no trespass-
ing’ sign that effectively communicates the location of a land boundary is one type of
property-related information cost, as is the mental energy that a passerby expends to read
and understand that sign, and the aggravation that arises from inadvertent trespass after
the sign falls down. Information costs overlap with transaction costs, a term typically used
more generally (and seldom very precisely) to describe the information costs, negotiation
costs, and enforcement costs related to voluntary exchanges.15
There are many ways in which the doctrines of property law have been shaped by
concerns about information costs. Indeed, some scholars suggest that one of the most
13
Carol Rose offers a similar but not identical categorization, identifying the concerns as
involving information or notice, renegotiability, and value (including third-party effects) (Rose,
2011).
14
The Supreme Court has addressed (and generally reaffirmed the importance of) IP exhaus-
tion repeatedly in recent years. See Impression Products v. Lexmark, Inc., 137 S. Ct. 1523 (2017);
Kirtsaeng v. John Wiley & Sons, 568 U.S. 519 (2013); Bowman v. Monsanto, 569 U.S. 278 (2013);
Quanta Computer, Inc. v. LG Electronics, Inc., 553 U.S. 617 (2008); see also Omega S.A. v. Costco
Wholesale Corp., 541 F.3d 982 (9th Cir. 2008), aff’d by an evenly divided Court, Costco Wholesale
Corp. v. Omega S.A., 562 U.S. 40 (2010).
15
This characterization is an extreme simplification of the variety of ways in which the term
‘transaction costs’ has been used, as Lee Anne Fennell documents (2013). It maps—roughly—the
categories emphasized by Coase (1960) and around which incomplete definitional consensus has
formed. As Fennell observes, ‘[t]here is broad agreement that the costs people incur to get together,
communicate with each other, and draw up and police contracts represent transaction costs. But
the status of some other elements is contested’ (2013, p. 1484). Ellickson characterizes transaction
costs within ‘three somewhat overlapping functional categories: (1) get-together costs, (2) decision
and execution costs, and (3) information costs’ (1989, p. 615).

fundamental characteristics of property rights is their potential to impose undesirably

high information costs, if not properly designed. From this perspective, property doctrine
is and should be structured to mitigate information problems.16
The argument that information costs are of special concern in property law—
compared, for example, to the law of contracts—is based in large part on the
observation that property rights are ‘in rem,’ or ‘good against the world.’ That is, the
obligations that arise from property rights (my obligation to avoid driving past a ‘no
trespassing’ sign and onto the land beyond, for example) are not based upon any special
relationship between owners and nonowners. These obligations bind complete strangers
who encounter things to which property rights attach, regardless of whether they have
agreed to be bound.
As Thomas Merrill and Henry Smith, leading proponents of an information-cost-
centric theory of property, summarize:
The in rem nature of such rights means all actors in the relevant community must recognize that
they are subject to a duty to abstain from interfering with such rights insofar as they are held
by any other member of the community. This generalized duty, in turn, creates an enormous
information cost and collective action problem. The rights must be defined in such a way that
their attributes can be easily understood by a huge number of persons of diverse experience
and intellectual skills. The identity of the persons who hold such rights must be capable of
communication by signals that can be immediately grasped and processed by an equally large
multitude (2007, p. 1853).
Merrill and Smith use these observations about the in rem nature of property rights to
generate a design principle: ‘the rights must be defined in such a way that their attributes
can be easily understood’ (2007, p. 1853). Specifically, Merrill and Smith (2000) argue
that property rights do and should conform to a limited, standardized set of easy-to-
understand forms—not ‘one size fits all,’ but a small menu that resists customization and
idiosyncrasy in order to economize on information costs.
Information costs do not always create problems worth solving through adjustments to
legal doctrine, however. And even in circumstances in which they might, standardization
of easy-to-understand property forms is only one mechanism through which to address
those problems. But thinking about the special virtues of standardization and clarity
where rights are good against the (vast, heterogenous) world is a good starting point for
thinking about the relationship between information costs and rights regimes that govern
both tangible and intangible resources. What if the nature of the resource at issue makes
it extremely difficult to specify rights to it in a clear, standardized way? Thinking about
this dynamic in relation to the nature of the resource helps us to think critically about the
design of the law.
IP scholars often contrast tangible and intangible property schemes on the basis of how
much information is readily available about the identity of property owners and the nature
of their rights. Typically, the comparison holds up tangible property—real property in
particular—as the model of successful information provision. Physical signs can provide
clues that a piece of land is owned by someone (often the person in possession). Customs
16
Note that not all information costs (or transaction costs more generally) constitute problems
to be solved (Cooter, 1982, p. 28; Fennell, 2013, p. 1473).

in the relevant neighborhood can shape how those signs are understood.17 Public records
indicate exactly who that someone is and reveal details about the physical dimensions of
the parcel, how its ownership has changed over time, and whether express encumbrances
(liens, servitudes, and so on) complicate ownership.18 These sources of information help
to prevent inadvertent trespass by those who wish to avoid invading private land; they
facilitate consensual transactions for those who seek permission to use or buy it. IP rights,
by contrast, do not so neatly correspond either to physical things in the world or to public
records signifying ownership and identifying owners.
Anxiety about the inadequacy of information regarding IP rights has increased in recent
years due to statutory changes that have made the situation worse (e.g. the elimination
of registration and notice as prerequisites for copyright protection) and to technological
changes that have raised the stakes and thickened thickets of (often hidden) rights. In
copyright, this anxiety is manifest in policy debates about the status of ‘orphan works’
whose owners cannot be identified and located (e.g. Chiang, 2016; Hansen, 2013; Loren,
2012; Urban, 2012; U.S. Copyright Office, 2006; U.S. Copyright Office, 2015). In patent,
critics are alarmed when innovators’ investments are jeopardized by allegations that they
have infringed unclear and thus difficult-to-avoid patent claims—especially in the realms
of software and Internet business methods (Bessen and Meurer, 2008; Long, 2004; Menell
and Meurer, 2013). In both the copyright and patent contexts, informational inadequacies
can contribute to inadvertent infringement and then to surprising and costly disputes.
Or fear of potential infringement—combined with the inability to identify, locate, and
negotiate with relevant rights-holders—can chill productive endeavors (see Federal Trade
Commission, 2011, p. 3; U.S. Copyright Office, 2006, p. 15).
One approach to alleviating the information cost challenges that plague IP would be
to try to replicate real property’s formal systems of centralized information provision.
Copyright reformers, in particular, have called for statutory changes modeled on the
centralized ownership information provided by land recording systems and the title-
clearing function performed by marketable title acts (e.g. Lessig, 2010, p. 29). This could
be accomplished, some argue, by ‘reformalizing copyright’ (e.g. Sprigman, 2004).
The formal, centralized, and sometimes error-prone information mechanisms associ-
ated with land titles are not the only models offered by the law of tangible property,
however. These structures coexist with other legal mechanisms—including rules about
the form of property rights and the remedies triggered by their infringement—that are
attentive to information costs concerns. As I have explained elsewhere (Van Houweling,
2013), this common law tradition features a wide variety of doctrinal tools. Even (or
perhaps especially) if the formal, centralized informational structures of the land law are
never fully replicated for intangible property, this set of tools may prove a valuable source
of ideas for addressing contemporary IP challenges.
For example, the touch and concern doctrine has traditionally constrained the subject
matter of land servitudes in a way that helps to limit the information cost burden
imposed by these potentially confusing non-possessory property rights. Like the other
17
On custom in property and IP, see, e.g., Smith (2009a) and Rothman (2007, 2012).
18
For a helpful summary of formal information infrastructures in tangible property (with
comparison to copyright), see van Gompel (2011, pp. 244–6).

roperty-standardizing doctrines discussed by Merrill and Smith (2000), touch and concern
p
polices just how idiosyncratically the sticks in the bundle of property can be arranged. By
limiting permissible servitudes to those that have a connection to the land they burden (and
typically to a neighboring benefitted parcel) the doctrine helps to ensure that servitudes will
be relatively easy to discover upon physical inspection, and that the owner of the beneficial
interest will be relatively easy to identify and locate (Van Houweling, 2016a, 2016b).
As I have argued elsewhere (Van Houweling 2008, 2011, 2013, 2016a, 2016b), IP exhaus-
tion similarly helps to limit the information cost burden imposed by non-possessory IP
rights. It generally allows owners of objects embodying copyrighted works and patented
inventions to transfer those objects and to make normal consumer uses of them, consist-
ent with reasonable expectations about what it means to own objects of personal property
(see generally Perzanowski and Schultz, 2015, 2016).
The upshot of this comparison is open to debate in light of developments in the law
of servitudes, where the trend has been toward liberalization of doctrines like touch and
concern in favor of enforcement of even idiosyncratic servitudes. Some observers take
this to suggest that IP exhaustion should be similarly liberalized in favor of enforcement
of idiosyncratic running restrictions on how IP-burdened objects may be used and
transferred (see, e.g., Robinson (2004), but see Van Houweling (2008)). But on this point
of comparison we should be careful to keep in mind key differences between tangible
property and IP. The doctrinal liberalization in the law of land servitudes has happened in
response to the establishment and improvement of recording systems that provide notice
of even unusual servitudes. The logic of this doctrinal development might suggest, ironi-
cally, quite a different evolution in the law of IP, where digital age developments appear
thus far to have exacerbated rather than alleviated information cost problems.
In sum, the nature of property rights poses a recurring set of information cost chal-
lenges that are relevant to both tangible property and IP. Although difficult-to-define
works of creativity and invention come with special information cost burdens, longstand-
ing doctrines within tangible property law suggest useful tools that might be deployed
to mitigate information cost concerns. Doctrines like intellectual property exhaustion,
whose analogs and progenitors in the land law are being discarded, remain important for
managing IP rights in an increasingly complex information environment.
C. Property and Time
This section focuses on how IP compares to other types of property on the dimension
of time—that is, how long the exclusive rights associated with copyrights and patents
last compared with the exclusive rights associated with property rights in land and other
tangible objects.19 One oft-cited distinction between copyright and patent law and the laws
governing property in tangible things like land and chattels is evident on the face of the
constitutional language limiting the duration of authors’ and inventors’ rights, which may
only be secured ‘for limited Times’ (U.S. Const. art. I, § 8, cl. 8).20 To those skeptical of
19
This section is derived in part from my previous work (Van Houweling, 2017).
20
For commentary emphasizing this distinction, see, e.g., Sterk (2005, pp. 446–59) and Bell and
Parchomovsky (2002, p. 41).

the notion of copyright and patent as forms of ‘property,’ this finite duration of authors’
and inventors’ rights is one of many ways in which those rights do not and should not
operate like property. These skeptics point to the expansion in the duration of copyright
as a way in which copyright is becoming more like tangible property and violating the
spirit if not the letter of the constitutional limitation.21 Conversely, some of those who
most fully embrace the concept of intellectual property argue that copyrights and patents
should—like other forms of property—last forever.22
All of these scholars and advocates share a similar starting assumption: tangible
property rights are potentially infinite in duration, while copyrights and patents are
constitutionally required to be for ‘limited times.’ This characterization is ripe for refine-
ment. While the duration of rights to land and other tangible objects may be theoretically
infinite, a variety of limiting doctrines operate to terminate these rights when they
threaten to prevent societally beneficial use of valuable resources. Long-lived property
claims trigger fears about ‘dead hand control,’ a label that reflects underlying anxiety
about special types of information and transaction costs that arise as owners move and
proliferate and their claims become entangled over time, about threats to the autonomy
of the living imposed by enforcing the preferences of prior generations, and about unfair
distribution of resources caused by dynastic wealth accumulation. A variety of property
doctrines are attentive to these fears. Indeed, the infamous Rule Against Perpetuities is
arguably important as a topic of study for property students more for the powerful way
that it illustrates the force of the concern with dead hand control than for its contempo-
rary doctrinal significance. The law of adverse possession is similarly important in part
for the way in which it illustrates problems caused by the assertion of stale claims.
As for IP, despite express limits on the duration of copyrights and patents, the problems
posed by stale, obsolete, and hopelessly entangled rights nonetheless loom large where
technology is advancing so rapidly, where there is no natural limit on the proliferation
of property claims, and where the public interest in access to the relevant resources often
lasts far longer than the property owners’ interest in making available the rights-related
information necessary to facilitate voluntary transactions. The issue of time is especially
pressing in the copyright context. While the duration of copyright is theoretically limited,
for many works it might as well be infinite. This is true, for example, for some so-called
‘orphan works’ whose owners cannot be located. These works may be underused during
21
For example, Lawrence Lessig laments that ‘though the founders never used that term,
“intellectual property,” . . . to us, copyright and patents are clearly property rights, and clearly
deserve all the absolute and permanent protection that ordinary property deserves’ (2001, p. 1068).
Similarly, Simon Stern observes that ‘advocates of heightened copyright protection find it hard to
resist the analogy with tangible property when challenging the limited duration of copyright’ (2012,
p. 87).
22
Justin Hughes (2003, pp. 784–5) has documented such arguments:
Motion Picture Association of America President Jack Valenti, for example, has stated pub-
licly that copyrights should be permanent, like any other property right. He has been joined by
at least one member of the U.S. Congress, Representative Mary Bono. Ms. Bono and Mr. Valenti
carry on the legacy of many nineteenth-century U.S. authors who were advocates of perpetual
copyright protection. There is some (many would say, superficial) appeal to their position: If
one views a copyright as just another form of property, it makes sense to ask why it is treated
differently than the enduring property rights in real estate, chattels, and financial instruments.

their long copyrights because permission to use them in ways subject to copyright cannot
be obtained from unfindable owners (U.S. Copyright Office, 2006, p. 15). If such underuse
includes a failure to properly preserve or duplicate existing copies of the works (fragile
books or films, for example) then their use will also be effectively restricted even after the
copyrights have expired.23 Thus the duration of copyright’s restrictions can be practically
infinite, and yet tragically worthless in the long run to both lost copyright owners and
society at large (Bibb, 2009, pp. 169–71).
The problems associated with orphan works and other unintended consequences
of ever-longer copyright terms are well-documented (e.g. Boyle, 2008; Chiang, 2016;
Hansen, 2013; Loren, 2012; Urban, 2012; U.S. Copyright Office, 2006; U.S. Copyright
Office, 2015). A wide variety of proposals has been offered to address them—including
ex ante durational limits (Khanna, 2014), more comprehensive recording to help keep
track of copyright owners (Landes and Posner, 2003), and time-sensitive application
of doctrines like fair use (Hughes, 2003; Liu, 2002). But these proposals are often met
with objections framed in terms of property rights, based on the assertion that tangible
property rights are infinite and copyrights should be infinite as well (or at least as close
to infinite as the Constitution’s ‘limited times’ language will bear) (Hughes, 2003, p. 784–5
(citing examples)).
On closer inspection, it is clear that there are many ways that the duration of property
rights in tangible things is in fact limited (Harding, 2009, pp. 292–3). This is especially
common for those property rights that—like IP—allow their owners to exercise remote
control over resources possessed by others. Return to the passage from White-Smith v.
Apollo quoted above. Justice Holmes claims that because copyright is a non-possessory,
‘in vacuo,’ property right that exerts remote control over strangers, it ‘could not be recog-
nized or endured for more than a limited time’ (White-Smith Music Publ’g Co. v. Apollo
Co., 209 U.S. 1, 19 (1908)). His observation is also aptly applied to the other forms of
‘remote control’ property discussed above.
On the one hand, servitudes and other non-possessory interests in tangible property
(e.g. future interests accompanying defeasible fees) are designed to be durable, and thus
to facilitate efficient long-term land use planning that would be difficult to accomplish
through bilateral contractual measures alone. (Copyrights, too, are more durable than
attempts to control copying of books by extracting express promises from each recipi-
ent of a copy.24) And yet, this durability can undermine efficiency when land use needs
change and transaction costs make the restrictions difficult to renegotiate. Consider, for
example, a residential use restriction imposed on a pocket of land hemmed in by noisy
new highways. Because the likelihood of changed circumstances and of transactional
difficulties increases with the passage of time, the durability of non-possessory property
23
For examples of these types of failures, see Reese (2012, pp. 292–7). Note that these are fail-
ures arguably attributable to living copyright owners. In a thought-provoking article, Eva Subotnik
has therefore rejected the ‘dead hand’ as a way of conceptualizing the problems attributable to long
(specifically, postmortem) copyright duration, identifying the problem as suboptimal stewardship
by the living (2015, pp. 118–24).
24
Contemporary mass market ‘agreements’ imposed via click-wrap and other mechanisms that
purport to attach to copies of intellectual works and thereby to establish privity of contract with
whoever obtains possession of a copy are a different story (see Van Houweling, 2008).

rights contributes to ambivalence about them.25 The specter of ‘dead hand control’
enforcing obsolete restrictions on living landowners and sacrificing socially beneficial use
of resources thus explains much of the longstanding judicial hostility to enforcement of
non-possessory property restrictions.
Although no constitutional ‘limited times’ constraint applies to ‘non-intellectual’
property rights, judges and legislators have nonetheless imposed a collection of durational
limitations on non-possessory property rights attached to land. Some of these limits
(including some versions of the Rule Against Perpetuities) operate ab inicio to invalidate
non-possessory property rights that purport to be perpetual or potentially perpetual.
Some serve as durational limits that cause property rights to expire automatically after a
certain period of time. Some allow property owners to save their rights from automatic
expiration only by recording them periodically. Others operate ex post to extinguish
property rights that have outlived their usefulness. There remain theoretically perpetual
non-possessory property rights, but they are vulnerable to invalidation on a number of
different grounds that are facially not about duration but appear to be influenced by
concerns related to dead hand control.
I will save consideration of the Rule Against Perpetuities and other limitations on
future interests for another day (see Van Houweling, 2017) and focus here on durational
limits on servitudes. A few states have enacted strict durational limits on servitudes,
winning praise from some commentators (e.g. Rose, 1982, pp. 1413–16; Sweeney, 1995,
p. 691; Winokur, 1989, p. 78). In Minnesota, for example, ‘covenants, conditions, or
restrictions’ (with some exceptions) expire after 30 years. (Minn. Stat. Ann. § 500.20(2)
(2002)). The relevant Massachusetts statute effectively forbids perpetual restrictions by
imposing a 30-year default term on any restriction that does not specify its own duration
(184 Mass. Gen. Laws 23). As the Supreme Judicial Court has explained, this provision
‘expressly precludes the imposition of perpetual restrictions on land’ and serves to
‘provide definitive endpoints to the term of such restrictions’ (Brear v. Fagan, 447 Mass.
68, 77 (2006)).
Many states impose recording and renewal requirements instead of, or in conjunction
with, durational limitations on servitudes. In many states, these requirements are imposed
by marketable title acts that apply to various types of clouds on title. These statutory
provisions generally declare that a landowner whose chain of title goes back for a specified
period of time (typically from 20 to 40 years) has marketable title to that land free and
clear of any contrary claims that have not been recorded (Van Houweling, 2017, pp. 68–9,
n. 82 (citing examples)).
Some states make all of this more explicit, with statutory provisions separate from
their general marketable title acts that expressly provide for the automatic expiration of
servitudes and/or future interests unless they are periodically recorded. For instance, the
Iowa ‘stale use statute’ applies to both future interests and servitudes that restrict the use
of land (Iowa Code Ann. § 614.24). The statute limits the duration of land restrictions
to 21 years from initial recording unless a claim to extend them is recorded before the
expiration of those 21 years. Similarly, in Massachusetts, restrictive servitudes with
express durations of more than 30 years nonetheless automatically expire after 30 years
25
Julia Mahoney (2002, 2004) is an especially vocal critic of this type of dead hand control.

unless they are recorded (with renewal recordation required every 20 years thereafter) (184
Mass. Gen. Laws 27).26
Ironically, the rise of marketable title acts and similar recording and renewal require-
ments during the second half of the twentieth century coincided with the elimination of
the renewal requirement in US copyright law. Recognition of the unintended negative
consequences of that shift have motivated calls to reinstitute some type of registration
requirement as a prerequisite for long-lasting copyrights. For example, while serving as
Register of Copyrights, Maria Pallante suggested that ‘a formal registration require-
ment near the end of term may be beneficial to the larger legal framework’ of copyright
(Pallante, 2013, p. 1419). The details of her proposal mirror the operation of some of the
state law recording schemes just described:
Congress could shift the burden of the last twenty years of protection . . . from the user to the
copyright owner, so that at least near the end of the term, the copyright owner would have to file
with the Copyright Office as a condition of continued protection. Otherwise, the work would
enter the public domain. (Pallante, 2013, p. 1419)
Pallante’s suggestion echoes other proposals to fix copyright by ‘reformalizing’ it, as

discussed above. The examples from the tangible property realm suggest that formal
requirements that help to justify and track long-lasting remote control rights would not
only restore some advantages of historical copyright practice, but also unite copyright
with the law that helps to discipline similar rights attached to land. And in the twenty-first
century, technology might ease the process of complying with formal requirements, thus
avoiding the traps and travails of the earlier era.
In addition to ex ante durational limits that set a pre-determined expiration date for
restrictive future interests and servitudes or impose an expiration date on those that have
not been properly recorded, there are also both common law doctrines and state statutory
provisions that operate ex post to terminate non-possessory use restrictions once they
have outlived their usefulness. Underlying these ex post termination rules is the idea that
circumstances may change in the years following imposition of a restriction in ways that
render what was once a beneficial land use control into an unjustifiable hindrance that is
nonetheless difficult to remove through voluntary negotiations. In the case of servitudes,
the relevant doctrine transparently reflects this logic: restrictive servitudes can be termi-
nated (or modified, or enforced only with damages as opposed to injunctions) due to
‘changed conditions.’ As the Ninth Circuit put it, ‘[t]he doctrine of changed conditions
operates to prevent the perpetuation of inequitable and oppressive restrictions on land
use and development that would merely harass or injure one party without benefiting the
other’ (Cortese v. United States, 782 F.2d 845, 850–51 (9th Cir. 1986)).
What would it mean to borrow the changed circumstances concept and apply it to
copyright? The most straightforward way to incorporate this idea would be for judges
applying the case-by-case analysis of the fair use doctrine to include in that inquiry a
consideration of the age and continued importance of the copyright at issue. Several
scholars have suggested this approach, tying it persuasively to fair use’s statutory factors
26
This is in addition to the Massachusetts provision discussed above, which imposes a 30-year
limitation on any use restriction that does not specify its own duration.

(Hughes, 2003; Liu, 2002). Here we see that this approach has a property law pedigree
as well.
In sum, courts and legislatures have deployed a range of techniques to limit the
problems caused by long-lasting remote control property rights. These techniques include
invalidation of some purportedly perpetual or long-lasting interests, mandatory term
limits, periodic recording requirements, and ex post invalidation of obsolete restrictions.
These are in addition to duration-related doctrines that apply to both possessory and
remote control property interests (namely, adverse possession and statutes of limitation).
Proposals for addressing problems caused by long-lived IP rights are met with objec-
tions often framed in terms of tangible property rights, with some vocal copyright and
patent owners insisting that their property should not be uniquely burdened with dura-
tional limits, recording obligations, or doctrines that consider the possibility that their
rights have outlived their usefulness. But looking carefully at the law of tangible property
reveals that such burdens would be far from unique, especially when we examine the law’s
treatment of non-possessory remote control property rights that are most analogous to IP.
There is a long property tradition of looking ambivalently on remote control property—
recognizing its potential to serve goals associated with long-term investment in valuable
resources but guarding against its potential to unjustifiably constrain resource use long
after those goals are accomplished or obsolete.
V. CONCLUSION
Property rights are uniquely powerful and problematic because of the way they attach to,
run with, and exclude people from things. The powers and the problems follow patterns,
but also vary depending on the nature of the thing at issue. Keeping this variance in mind
as we explore analogies between tangible property and IP allows us to use those analogies
to distinguish and diagnose the novel and yet familiar challenges of contemporary IP.
As explored in this chapter, attention to both the history of property possession and the
special challenges of possession of IP helps to understand and diagnose the shortcomings
of doctrines of IP acquisition and signaling. Looking at the special information cost
challenges posed by rights that are good against the world helps us to understand how
IP might harness more of the doctrinal tools that have traditionally helped mitigate these
costs. And examining how the law governing non-possessory rights to tangible property is
sensitive to how those rights can cause problems over time helps us to see how durational
limits in IP are not confiscatory but rather consistent with property tradition. In sum,
understanding IP as property can illuminate rather than obscure the special nature of
rights in intellectual creations.
REFERENCES
American Law Institute. 2000. Restatement (Third) of Property: Servitudes. Philadelphia, PA: American Law
Institute.
Arrow, Kenneth J. 1962. ‘Economic Welfare and the Allocation of Resources for Invention,’ in Richard R.
Nelson, ed., The Rate and Direction of Inventive Activity: Economic and Social Factors. Cambridge, MA:
National Bureau of Economic Research.

Barros, Benjamin. 2011. ‘The Biology of Possession,’ 20 Widener Law Journal 291–317.
Bell, Abraham, and Gideon Parchomovsky. 2002. ‘Pliability Rules,’ 101 Michigan Law Review 1–79.
Bessen, James, and Michael J. Meurer. 2008. Patent Failure: How Judges, Bureaucrats, and Lawyers Put
Innovators at Risk. Princeton, NJ: Princeton University Press.
Bibb, Megan L. 2009. ‘Note, Applying Old Theories to New Problems: How Adverse Possession Can Help Solve
the Orphan Works Crises,’ 12 Vanderbilt Journal of Entertainment & Technology Law 149–81.
Boyle, James. 2003. ‘The Second Enclosure Movement and the Construction of the Public Domain,’ 66 Law &
Contemporary Problems 33–74.
Boyle, James. 2008. The Public Domain. New Haven, CT: Yale University Press.
Breyer, Stephen. 1970. ‘The Uneasy Case for Copyright: A Study of Copyright in Books, Photocopies, and
Computer Programs,’ 84 Harvard Law Review 281–351.
Calabresi, Guido, and A. Douglas Melamed. 1972. ‘Property Rules, Liability Rules, and Inalienability: One View
of the Cathedral,’ 85 Harvard Law Review 1089–128.
Carrier, Michael A. 2004. ‘Cabining Intellectual Property through a Property Paradigm,’ 54 Duke Law Journal
1–145.
Chander, Anupam. 2003. ‘The New, New Property,’ 81 Texas Law Review 715–97.
Chiang, Tun-Jen. 2016. ‘Trolls and Orphans,’ 96 Boston University Law Review 691–715.
Coase, R.H. 1960. ‘The Problem of Social Cost,’ 3 Journal of Law and Economics 1–44.
Cohen, Julie E. 2014. ‘What Kind of Property is Intellectual Property?,’ 52 Houston Law Review 691–707.
Cooter, Robert. 1982. ‘The Cost of Coase,’ 11 Journal of Legal Studies 1–33.
Dagan, Hanoch. 2012. ‘Pluralism and Perfectionism in Private Law,’ 112 Columbia Law Review 1409–45.
Demsetz, Harold. 1967. ‘Toward a Theory of Property Rights,’ 57 American Economic Review 347–59.
Depoorter, Ben. 2011. ‘Fair Trespass,’ 111 Columbia Law Review 1090–135.
Drassinower, Abraham. 2006. ‘Capturing Ideas: Copyright and the Law of First Possession,’ 54 Cleveland State
Law Review 191–204.
Dukeminier, Jesse, James E. Krier, Gregory S. Alexander, Lior J. Strahilevitz, and Michael Schill. 2014. Property.
New York, NY: Wolters Kluwer Law & Business.
Easterbrook, Frank H. 1990. ‘Intellectual Property Is Still Property,’ 13 Harvard Journal of Law & Public Policy
108–18.
Ellickson, Robert C. 1989. ‘The Case for Coase and against “Coaseanism”,’ 99 Yale Law Journal 611–30.
Epstein, Richard. 1979. ‘Possession as the Root of Title,’ 13 Georgia Law Review 1221–43.
Epstein, Richard. 2005. ‘Liberty versus Property? Cracks in the Foundation of Copyright Law,’ 42 San Diego
Law Review 1–28.
Epstein, Richard. 2010. ‘The Disintegration of Intellectual Property? A Classical Liberal Response to a
Premature Obituary,’ 62 Stanford Law Review 455–521.
Federal Trade Commission. 2011. The Evolving IP Marketplace: Aligning Patent Notice and Remedies with
Competition.Washington, D.C.
Fennell, Lee Anne. 2009. ‘Adjusting Alienability,’ 122 Harvard Law Review 1403–65.
Fennell, Lee Anne. 2013. ‘The Problem of Resource Access,’ 126 Harvard Law Review 1471–531.
Fraley, Jill M. 2011. ‘Finding Possession: Labor, Waste, and the Evolution of Property,’ 39 Capital University
Law Review 51–83.
French, Susan F. 1982. ‘Toward a Modern Law of Servitudes: Reweaving the Ancient Strands,’ 55 Southern
California Law Review 1261–319.
Friedman, Ori, and Karen Neary. 2009. ‘First Possession beyond the Law: Adults’ and Young Children’s
Intuitions about Ownership,’ 83 Tulane Law Review 679–90.
Fromer, Jeanne. 2009. ‘Claiming Intellectual Property,’ 76 University of Chicago Law Review 719–95.
Ghosh, Shubha. 2007. ‘The Fable of the Commons: Exclusivity and the Construction of Intellectual Property
Markets,’ 40 U.C. Davis Law Review 855–90.
Gibson, James. 2005. ‘Once and Future Copyright,’ 81 Notre Dame Law Review 167–243.
Ginsburg, Jane. 2010. ‘The U.S. Experience with Mandatory Copyright Formalities: A Love/Hate Relationship,’
33 Columbia Journal of Law & the Arts 311–48.
Grey, Thomas C. 1980. ‘The Disintegration of Property,’ in Roland J. Pennock and John W. Chapman, eds.,
Nomos XXII: Property. New York: New York Univsity Press.
Hansen, David R., Kathryn Hashimoto, Gwen Hinze, Pamela Samuelson, and Jennifer M. Urban.
2013. ‘Solving the Orphan Works Problem for the United States,’ 37 Columbia Journal of Law & the Arts
1–54.
Hardin, Garrett. 1968. ‘The Tragedy of the Commons,’ 162 Science 1243–8.
Harding, Sarah. 2009. ‘Perpetual Property,’ 61 Florida Law Review 285–327.
Holbrook, Timothy R. 2006. ‘Possession in Patent Law,’ 58 SMU Law Review 123–76.
Holbrook, Timothy R. 2009. ‘Equivalency and Patent Law’s Possession Paradox,’ 23 Harvard Journal of Law
and Technology 1–48.

Holbrook, Timothy R. 2016. ‘Patent Anticipation and Obviousness as Possession,’ 65 Emory Law Journal
987–1050.
Hughes, Justin. 2003. ‘Fair Use across Time,’ 50 UCLA Law Review 775–800.
Hughes, Justin. 2012. ‘A Short History of “Intellectual Property” in Relation to Copyright,’ 33 Cardozo Law
Review 1293–340.
Kapczynski, Amy. 2012. ‘The Cost of Price: Why and How to Get Beyond Intellectual Property Internalism,’
59 UCLA Law Review 970–1026.
Khanna, Derek. 2014. ‘Guarding Against Abuse: The Costs of Excessively Long Copyright Terms,’ 23
CommLaw Conspectus 52–125.
Korngold, Gerald. 1988. ‘For Unifying Servitudes and Defeasible Fees: Property Law’s Function Equivalents,’
66 Texas Law Review 533–76.
Landes, William M., and Richard A. Posner. 2003. ‘Indefinitely Renewable Copyright,’ 17 University of Chicago
Lemley, Mark A. 1997. ‘Romantic Authorship and the Rhetoric of Property,’ 75 Texas Law Review 873–906.
Lemley, Mark A. 1999. ‘The Modern Lanham Act and the Death of Common Sense,’ 198 Yale Law Journal
1687–715.
Lemley, Mark A. 2005. ‘Property, Intellectual Property, and Free Riding,’ 83 Texas Law Review 1031–75.
Lemley, Mark A. 2016. ‘Ready for Patenting,’ 96 Boston University Law Review 1171–95.
Lessig, Lawrence. 2001. ‘Copyright’s First Amendment,’ 48 UCLA Law Review 1057–73.
Lessig, Lawrence. 2010. ‘For the Love of Culture: Google, Copyright, and Our Future,’ 241 The New Republic
24–30.
Liivak, Oskar, and Eduardo Peñalver. 2013. ‘The Right Not to Use in Property and Patent Law,’ 98 Cornell
Litman, Jessica. 2016. ‘What Notice Did,’ 96 Boston University Law Review 717–44.
Liu, Joseph. 2002. ‘Copyright and Time: A Proposal,’ 101 Michigan Law Review 409–81.
Liu, Joseph P. 2016. ‘Fair Use, Notice Failure, and the Limits of Copyright as Property,’ 96 Boston University
Long, Clarisa. 2004. ‘Information Costs in Patent and Copyright,’ 90 Virginia Law Review 465–549.
Loren, Lydia Pallas. 2012. ‘Abandoning the Orphans: An Open Access Approach to Hostage Works,’ 27 Berkeley
Technology Law Journal 1431–70.
Lueck, Dean. 1995. ‘The Rule of First Possession and the Design of the Law,’ 38 Journal of Law and Economics
393–436.
Lunney, Glynn S. 1999. ‘Trademark Monopolies,’ 48 Emory Law Journal 367–486.
Mahoney, Julia D. 2002. ‘Perpetual Restrictions on Land and the Problem of the Future,’ 88 Virginia Law
Review 739–87.
Mahoney, Julia D. 2004. ‘The Illusion of Perpetuity and the Preservation of Privately Owned Lands,’ 44 Natural
Resources Journal 573–600.
Menell, Peter S. 2007a. ‘The Property Rights Movement’s Embrace of Intellectual Property: True Love or
Doomed Relationship?,’ 34 Ecology Law Quarterly 713–54.
Menell, Peter S. 2007b. ‘Intellectual Property and the Property Rights Movement,’ 30 Regulation 36–42.
Menell, Peter S. 2011. ‘Governance of Intellectual Resources and the Disintegration of Intellectual Property in
the Digital Age,’ 26 Berkeley Technology Law Journal 1523–59.
Menell, Peter S. 2016. ‘Economic Analysis of Copyright Notice: Tracing and Scope in the Digital Age,’ 96 Boston
University Law Review 967–1023.
Menell, Peter S., and Michael Meurer. 2013. ‘Notice Failure and Notice Externalities,’ 5 Journal of Legal
Analysis 1–53.
Merges, Robert P. 2011. Justifying Intellectual Property. Cambridge, MA: Harvard University Press.
Merrill, Thomas W., and Henry E. Smith. 2000. ‘Optimal Standardization in the Law of Property,’ 110 Yale
Law Journal 1–69.
Merrill, Thomas W., and Henry E. Smith. 2007. ‘The Morality of Property,’ 48 William and Mary Law Review
1849–95.
Mossoff, Adam. 2005. ‘Is Copyright Property?,’ 42 San Diego Law Review 29–43.
Nachbar, Thomas B. 2007. ‘The Comedy of the Market,’ 30 Columbia Journal of Law and the Arts 453–66.
Oliar, Dotan, and James Y. Stern. 2019. ‘Right on Time: First Possession in Property and Intellectual Property,’
99 Boston University Law Review 395–458.
Pallante, Maria A. 2013. ‘The Curious Case of Copyright Formalities,’ 28 Berkeley Technology Law Journal
1415–23.
Parchomovsky, Gideon, and Peter Siegelman. 2002. ‘Towards an Integrated Theory of Intellectual Property,’ 88
Virginia Law Review 1455–528.
Perzanowski, Aaron, and Jason Schultz. 2015. ‘Reconciling Intellectual and Personal Property,’ 90 Notre Dame

Perzanowski, Aaron, and Jason Schultz. 2016. The End of Ownership: Personal Property in the Digital Economy.
Cambridge, MA: MIT Press.
Radin, Margaret Jane. 1987. ‘Market-Inalienability,’ 100 Harvard Law Review 1849–937.
Reese, R. Anthony. 2012. ‘What Copyright Owes the Future,’ 50 Houston Law Review 287–317.
Robinson, Glen O. 2004. ‘Personal Property Servitudes,’ 71 University of Chicago Law Review 1449–523.
Rose, Carol M. 1982. ‘Servitudes, Security and Assent: Some Comments on Professors French and Reichman,’
55 Southern California Law Review 1403–16.
Rose, Carol M. 1985. ‘Possession as the Origin of Property,’ 52 University of Chicago Law Review 73–88.
Rose, Carol. 1986. ‘The Comedy of the Commons: Custom, Commerce, and Inherently Public Property,’ 53
University of Chicago Law Review 711–81.
Rose, Carol M. 2003. ‘Romans, Roads, and Romantic Creators: Traditions of Public Property in the Information
Age,’ 66 Law & Contemporary Problems 89–110.
Rose, Carol. 2011. ‘Servitudes,’ in Kenneth Ayotte and Henry E. Smith, eds., Research Handbook on the
Economics of Property Law. Cheltenham, UK and Northampton, MA, USA: Edward Elgar Publishing.
Rose, Mark. 1995. Authors and Owners: The Invention of Copyright. Cambridge, MA: Harvard University Press.
Rothman, Jennifer. 2007. ‘The Questionable Use of Custom in Intellectual Property,’ 93 Virginia Law Review
1899–982.
Rothman, Jennifer. 2012. ‘Custom, the Common Law and Intellectual Property,’ in Shyamkrishna Balganesh,
ed., Intellectual Property and the Common Law. Cambridge, UK: Cambridge University Press.
Samuelson, Pamela. 1989. ‘Information as Property: Do Ruckelshaus and Carpenter Signal a Changing
Direction in Intellectual Property Law?,’ 38 Catholic University Law Review 365–400.
Samuelson, Pamela, and the Copyright Principles Project. 2010. ‘The Copyright Principles Project: Directions
for Reform,’ 25 Berkeley Technology Law Journal 1175–245.
Samuelson, Pamela. 2016. ‘Notice Failures Arising from Copyright Duration Rules,’ 96 Boston University Law
Review 667–89.
Singer, Joseph William. 2014a. ‘Justice Breyer, Grokster, and the Four Chords Song,’ 128 Harvard Law Review
483–7.
Singer, Joseph William. 2014b. ‘Property as the Law of Democracy,’ 63 Duke Law Journal 1287–335.
Smith, Henry E. 2003. ‘The Language of Property: Form, Context, and Audience,’ 55 Stanford Law Review
1105–91.
Smith, Henry E. 2007. ‘Intellectual Property as Property: Delineating Entitlements in Information,’ 116 Yale
Law Journal 1742–822.
Smith, Henry E. 2009a. ‘Community and Custom in Property,’ 10 Theoretical Inquirites in Law 5–41.
Smith, Henry. 2009b. ‘Institutions and Indirectness in Intellectual Property,’ 157 University of Pennsylvania Law
Review 2083–133.
Smith, Henry E. 2012. ‘Property as the Law of Things,’ 125 Harvard L. Rev. 1691–726.
Sprigman, Christopher. 2004. ‘Reform(aliz)ing Copyright Law,’ 57 Stanford Law Review 485–568.
Stake, Jeffrey Evans. 2004. ‘The Property “Instinct,”’ 359 Philosophical Transactions of the Royal Society B
1763–74.
Sterk, Stewart E. 1996. ‘Rhetoric and Reality in Copyright Law,’ 94 Michigan Law Review 1197–249.
Sterk, Stewart. 2005. ‘Intellectualizing Property: The Tenuous Connections between Land and Copyright,’ 83
Washington University Law Quarterly 417–70.
Stern, Simon. 2012. ‘From Author’s Right to Property Right,’ 62 University of Toronto Law Journal 29–91.
Subotnik, Eva E. 2015. ‘Copyright and the Living Dead?: Succession Law and the Postmortem Term,’ 29
Harvard Journal of Law & Technology 77–125.
Sweeney, Michael J.D. 1995. ‘The Changing Role of Private Land Restrictions: Reforming Servitude Law,’ 64
Fordham Law Review 661–96.
United States Copyright Office. 2006. Report on Orphan Works. Washington, D.C.
United States Copyright Office. 2015. Orphan Works and Mass Digitization: A Report of the Register of
Copyrights. Washington, D.C.
Urban, Jennifer M. 2012. ‘How Fair Use Can Help Solve the Orphan Works Problem,’ 27 Berkeley Technology
Vaidhyanathan, Siva. 2001. Copyrights and Copywrongs. New York: NYU Press.
Van Houweling, Molly Shaffer. 2002. ‘Cultivating Open Information Platforms: A Land Trust Model,’ 1 Journal
of Telecommunications and Hight Technology Law 309–23.
Van Houweling, Molly. 2005. ‘Distributive Values in Copyright,’ 83 Texas Law Review 1535–79.
Van Houweling, Molly Shaffer. 2007. ‘Cultural Environmentalism and the Constructed Commons,’ 70-SPG Law
and Contemporary Problems 23–50.
Van Houweling, Molly Shaffer. 2008. ‘The New Servitudes,’ 96 Georgetown Law Journal 885–950.
Van Houweling, Molly Shaffer. 2010. ‘Author Autonomy and Atomism in Copyright Law,’ 96 Virginia Law
Review 549–642.

Van Houweling, Molly Shaffer. 2011. ‘Touching and Concerning Copyright,’ 51 Santa Clara Law Review
1063–85.
Van Houweling, Molly Shaffer. 2012a. ‘Technology and Tracing Costs: Lessons from Real Property,’ in
Shyamkrishna Balganesh, ed., Intellectual Property and the Common Law. Cambridge, UK: Cambridge
University Press.
Van Houweling, Molly Shaffer. 2012b. ‘Atomism and Automation,’ 27 Berkeley Technology Law Journal
1417–501.
Van Houweling, Molly Shaffer. 2013. ‘Land Recording and Copyright Reform,’ 28 Berkeley Technology Law
Journal 1497–510.
Van Houweling, Molly Shaffer. 2016a. ‘Exhaustion and Personal Property Servitudes,’ in Irene Calboli
and Edward Lee, eds., Research Handbook on IP Exhaustion and Parallel Imports. Cheltenham, UK and
Northampton, MA, USA: Edward Elgar Publishing.
Van Houweling, Molly Shaffer. 2016b. ‘Exhaustion and the Limits of Remote-Control Property,’ 93 Denver
Van Houweling, Molly S. 2017. ‘Disciplining the Dead Hand of Copyright: Durational Limits on Remote
Control Property,’ 30 Harvard Journal of Law & Technology 53–74.
van Gompel, Stef. 2011. Formalities in Copyright Law. Alphen ann den Rijn, The Netherlands: Kluwer Law
International.
Walterscheild, Edward C. 1994. ‘To Promote the Progress of Science and Useful Arts: The Background and
Origin of the Intellectual Property Clause of the United States Constitution,’ 2 Journal of Intellectual Property
Law 1–56.
Winokur, James L. 1989. ‘The Mixed Blessings of Promissory Servitudes: Toward Optimizing Economic Utility,
Individual Liberty, and Personal Identity,’ 1989 Wisconsin Law Review 1–97.
Yen, Alfred C. 1990. ‘Restoring the Natural Law: Copyright as Labor and Possession,’ 51 Ohio State Law
Journal 517–59.
Legislative Materials
U.S. Const. art. I, § 8, cl. 8.

17 U.S.C. § 102.
35 U.S.C. § 102(a)–(b), § 112.
184 Mass. Gen. Laws 23.
184 Mass. Gen. Laws 27.
Iowa Code Ann. § 614.24.
Minn. Stat. Ann. § 500.20(2).
Cases
Ariad Pharmaceuticals, Inc. v. Eli Lilly & Co., 598 F.3d 1336 (2010).
Bowman v. Monsanto, 569 U.S. 278 (2013).
Brear v. Fagan, 447 Mass. 68 (2006).
Cortese v. United States, 782 F.2d 845 (9th Cir. 1986).
Costco Wholesale Corp. v. Omega S.A., 562 U.S. 40 (2010).
Impression Products v. Lexmark, Inc., 137 S. Ct. 1523 (2017).
Kewanee Oil Co. v. Bicron Corp., 416 U.S. 470 (1974).
Kirtsaeng v. John Wiley & Sons, 568 U.S. 519 (2013).
Omega S.A. v. Costco Wholesale Corp., 541 F.3d 982 (9th Cir. 2008).
Pierson v. Post, 3 Cai. 175 (N.Y. 1805).
Quanta Computer, Inc. v. LG Electronics, Inc., 553 U.S. 617 (2008).
White-Smith Music Publishing Company v. Apollo Company, 209 U.S. 1 (1908).

2. Anticommons, transaction costs, and patent
aggregators
Rebecca S. Eisenberg* 27
Contents
I. Theory of the Anticommons

II. Empirical Studies
A. Impact on Upstream Research
B. Impact on Downstream Product Development
III. The Rise of Patent Aggregators
IV. Changes to Mitigate Anticommons Risks
A. Limits on ‘Upstream’ Rights and Assertions
B. Limits on Injunctions
C. Lowering Information Costs
D. Lowering the Costs of Patent Challenges
References
I. THEORY OF THE ANTICOMMONS
An anticommons is a fragmented allocation of property rights in which resources are prone

to underuse because it is too costly to assemble all the necessary permissions to put the
resources to use (Heller, 1997; Heller and Eisenberg, 1998). The anticommons metaphor
harkens back to the more familiar ‘tragedy of the commons’ (Hardin, 1968), in which
resources made freely available in a commons (such as open fisheries and pastures) are
prone to overuse because too many people hold privileges of use and nobody has a right to
exclude anyone else. Privatization may solve a tragedy of the commons by giving a property
owner a right to exclude others, thus providing the means and incentive to prevent wasteful
overuse. But a poorly designed property regime that creates too many fragmented rights
to exclude could lead to wasteful underuse in a tragedy of the anticommons. In a world of
costless transactions, people could avoid both commons and anticommons tragedies by
trading their rights (Fennell, 2004). But transaction costs, strategic holdouts, and informa-
tion problems present a risk of persistent bargaining failures. The larger the number of
rights holders and the more varied their entitlements, the more challenging it is to avoid
waste through bargaining (Heller and Eisenberg, 1998; Eisenberg, 2001).
Heller (1997) offered as an example empty storefronts in post-Soviet Moscow,
alongside robust retail trade out of flimsy metal sidewalk kiosks. Heller attributed the
* Robert and Barbara Luciano Professor of Law, Michigan Law, University of Michigan.
27

underutilization of storefronts to fragmented allocation of the rights necessary to open up

shop to a diverse group of owners (including private or quasi-private enterprises, workers’
collectives, privatization agencies, and local, regional, and federal governments) in the
course of post-socialist privatization.
Heller and Eisenberg (1998) extended the metaphor to fragmented ownership of
intellectual property (IP) rights that an innovating firm must clear in order to develop a
new product.1 They noted complaints about a proliferation of ‘upstream’ IP claims to
biomedical research tools that serve as inputs to future ‘downstream’ product develop-
ment following a shift in US government policy to promote the patenting of inventions
made in the course of federally sponsored research (Eisenberg, 1996).
Both illustrations happened to involve the creation of property rights following a
transition from a more communal form of ownership (state ownership in the case of
Soviet-era storefronts; free availability in the scientific commons in the case of research
tools) to a new property rights regime. The creation of a new property regime, although
not the only way that an anticommons might arise, may be especially fraught with pos-
sibilities for poor design of initial rights. Moreover, new property owners accustomed to
a regime of communal ownership may initially lack the skills necessary to evaluate their
rights and to transfer them at reasonable cost to users who could get more value out of
them. These difficulties may eventually give way to successful bundling of rights through
bargaining, perhaps with the assistance of intermediaries who figure out strategies for
lowering transaction costs to realize potential gains from exchange (Merges, 1996; Heller
and Eisenberg, 1998; Eisenberg, 2001). Bargaining may be costly, but not necessarily
tragic (Barnett, 2015). At any given moment it can be difficult to tell the difference
between transitory and enduring bargaining obstacles (Eisenberg, 2001).
An anticommons may be more likely to persist if transaction costs are high, if het-
erogeneous rights are dispersed among diverse owners, and if information asymmetries
and cognitive biases prevent agreement on relative values (Heller and Eisenberg, 1998).
These conditions seem more likely with patent rights than with real property entitlements,
because the patent system continuously creates new rights of indeterminate value for new
claimants, with limited opportunity to establish consensus valuations of existing entitle-
ments as new rights arise in the face of ongoing technological change. On the other hand,
because patents are wasting assets that expire after 20 years and may lose value before that
time as a result of technological obsolescence, owners may feel some urgency to enter into
timely bargains before licensing options disappear (Epstein and Kuhlik, 2004).
Subsequent commentators have sharpened and formalized the model of the tragedy of
the anticommons and noted its similarities to a tragedy of the commons (Buchanan and
Yoon, 2000; Fennell, 2004; Parisi et al., 2005; Schulz et al., 2003). Both tragedies could
in theory be resolved through bargaining, and in both cases bargaining faces obstacles.
These obstacles include the costs of identifying owners and users and determining the
validity and infringement of rights, as well as the costs of evaluating heterogeneous
rights and of bargaining over terms of exchange. In both contexts individual rights
1
Other commentators have used the term ‘patent thickets’ to describe numerous and overlap-
ping patent rights that threaten to stifle rather than encourage innovation (Shapiro, 2001; Galasso
and Shankerman, 2010). Both terms describe similar phenomena.

Anticommons, transaction costs, and patent aggregators 29
holders bargaining independently are likely to charge more in the aggregate than would a
single party who held the entire bundle of complementary rights because of the ‘double
marginalization problem’: each individual rights holder will fail to internalize the full
costs and benefits of its bargaining behavior on the overall price of the bundle (Parisi et
al., 2005). In both contexts, there is a risk that an overall deal will be derailed by strategic
behavior by one or more rights holders seeking to capture a larger share of the potential
surplus (Fennell, 2004; David, 2011). Some experimental results support these intuitions
(Depoorter and Vanneste, 2006; Stewart and Bjornstad, 2002; Parente and Winn, 2012).
Other commentators maintain that market institutions can and do avoid anticommons
tragedies (Barnett, 2015; Lichtman, 2006; Epstein and Kuhlik, 2004).
Much of the anticommons literature focuses on the patent system, but some of it
considers problems arising from fragmentation of ownership of copyright (Parisi and
Sevcenko, 2001; Hunter, 2003; Loren, 2003; Depoorter, 2004; Van Houweling, 2010;
McLeod and DiCola, 2011; Lee, 2016) and associated transaction costs (Depoorter
and Parisi, 2002). US copyright law has mechanisms for avoiding an anticommons that
have no counterpart in the US patent system, including statutory compulsory licensing
provisions,2 well-established collective rights organizations (Gervais, 2015), and the
fair use doctrine (Depoorter and Parisi, 2002).3 Nonetheless, recent controversies have
highlighted salient examples of cultural innovations that have been significantly impeded
by the complexity of licensing numerous copyrights. These include mass digitization
of libraries of copyrighted works such as the Google Books project (Hofmann, 2013;
Samuelson, 2011; U.S. Copyright Office, 2015), digital music sampling (McLeod and
DiCola, 2011), and documentary film production (Aufderheide and Jaszi, 2004).
II. EMPIRICAL STUDIES
Much of the patent literature uses the term ‘anticommons’ loosely to lament various
effects of patents that are only indirectly related to the problem of clearing numerous
rights. For example, Murray and Stern (2007) characterize as ‘anticommons effects’ an
observed relative decline in the rate of citations to scientific articles after the inventions
they disclose are patented and thus removed from the public domain, even though
each observation may involve the effects of only a single patent. An extensive literature
purports to test the ‘anticommons’ by looking for negative effects of patents on the
‘upstream’ activities of biomedical research scientists, as distinguished from the more spe-
cific focus of Heller and Eisenberg (1998) on the impact of a proliferation of ‘upstream’
IP claims on ‘downstream’ product development. Despite these imprecisions in usage
2
17 U.S.C. §§ 115 (nondramatic musical compositions), 118 (public broadcasting), 111(c)
(cable retransmission), 114(d)(2) (subscription digital audio transmission), 114(d)(1) (nonsubscrip-
tion internet radio). Royalty rates are set by a Copyright Royalty Board established by statute
(Copyright Royalty and Distribution Reform Act of 2004, Pub. L. No. 108-419, 118 Stat. 2341,
codified at 17 U.S.C. §§ 801–805). For a history of the compulsory licensing provisions for nondra-
matic musical compositions and an explanation of their limitations see Peters (2004).
3
17 U.S.C. § 107. See also 17 U.S.C. § 110 (permitting unauthorized public performances
under specific circumstances).

and differences in focus, some of this literature illuminates the anticommons model in
useful ways. A smaller literature focuses more directly on the impact of ‘patent thickets’
on product development in a variety of technological fields. But the search for empirical
evidence of such an impact suffers from a lack of salience: it is challenging to measure or
observe innovations that never happened.
The following discussion considers separately studies of the impact of patents on
upstream research and on downstream product development. Section III then considers
growing concerns that in recent years have eclipsed debate about disaggregated ownership
of fragmented patents: a distinct set of problems posed by the rise of patent aggregators.
While patent aggregation seems like it would be an effective market solution to the bargain-
ing challenge presented by fragmented ownership, it seems to have done more to reduce
the costs of assertion by patent owners than to reduce the costs faced by technology users
for clearing rights. The result on balance may be a greater risk of underuse as subsequent
innovators need to incur more transaction costs to evaluate and clear rights that they might
otherwise have ignored with little risk of assertion in the absence of aggregation.
A. Impact on Upstream Research
The metaphor of the tragedy of the anticommons has had great resonance with science
policy makers whose primary concern was not the impact of IP on future product devel-
opment, but rather its impact on research science itself. This concern motivated a series of
studies using surveys of research scientists from around the world, primarily in academia,
to determine how much IP interferes with their work (Eisenberg, 2008). Although not
directly relevant to the problem of the anticommons as pictured by Heller and Eisenberg
(1998), these studies shed an interesting light on that problem.
The most interesting finding of these studies was that academic scientists for the most
part simply ignore any patents they might be infringing, and for the most part they get
away with it (Walsh et al., 2003; Eisenberg, 2001; Nicol and Nielsen, 2003; Straus et al.,
2004; American Ass’n for the Advancement of Science, 2007).4
An important exception to the obliviousness of academic scientists to patents was the
impact of patented genes on clinical research using genetic tests in academic medical
centers (Soini et al., 2008; Cho, 2003; Walsh et al., 2003). In telephone survey interviews
with directors of laboratories performing genetic tests in the US, Cho (2003) found that 25
percent had stopped performing a genetic test and 53 percent had decided not to develop
a new genetic test because of a patent or license. Of those respondents that had been
contacted about potential infringement, 71 percent of those in commercial laboratories
and 24 percent of those in universities reported that they had been prevented from
performing a test (Cho, 2003). Patents on the BRCA1 and BRCA2 genes associated with
susceptibility to breast cancer were a recurring source of complaints (Gold and Carbone,
2010), but not unique (Merz et al., 2002).
Although relatively few academic scientists reported that patents constrained their
4
A rare but salient counterexample in which a university was held liable for patent infringe-
ment is the case of Madey v. Duke University, 307 F.3d 1351 (Fed. Cir. 2002), U.S. cert. denied sub
nom. Duke University v. Madey, 539 U.S. 958 (2003).

research activities outside of genetic testing, many scientists reported that their work was
impeded by practical restrictions on access to proprietary materials and databases that
they could not readily duplicate on their own (Cohen and Walsh, 2008; Walsh et al., 2003,
2007). The distinction between patent licenses and agreements for access to materials and
databases is not entirely crisp. Mowery and Ziedonis (2007) found that materials transfer
agreements were often associated with patented inventions. Some notoriously difficult
negotiations have arisen over access to patented materials, such as genetically engineered
mice (Murray, 2006). Academic agricultural biologists surveyed by Lei et al. (2009)
associated an observed proliferation of materials transfer agreements with concerns about
preserving IP protection.
In a review of the literature, Herder (2012) noted that different authors have character-
ized quantitatively similar findings quite differently. For example, Walsh et al. (2007),
finding that 1 in 9 researchers abandoned one research project every two years because
of patents, concluded that worries about patent-related transaction costs are overstated,
while Huang and Murray (2009), finding that genetic researchers forego about 1 in 10
research projects (or, more precisely, research publications) through the causal negative
impact of a gene patent grant, concluded that there was reason to worry about the effects
of gene patents on knowledge flows.
Survey results also appear to be sensitive to the form of the questions. The agricultural
biologists surveyed by Lei et al. (2009) gave broadly similar responses to the scientists
surveyed by Cohen and Walsh (2008) and Walsh et al. (2007) on similar questions, but only
Lei et al. asked respondents to evaluate the impact of intellectual property on their field,
leading the agricultural biologists to ‘report that the IP protection of research tools is, on
balance, having a negative impact on their research areas.’ Perhaps other respondents in
the other surveys would have given similar responses to the same question.
Other scholars have used less impressionistic data to gauge the impact of patents on
subsequent research. Murray and Stern (2007) studied forward citations of patent-paper
pairs involving papers published in the journal Nature Biotechnology between 1997
and 1999 and patents associated with the same work. They found that the grant of a
patent was associated with a modest (10–20 percent) decline in future citations to the
corresponding paper relative to expected citations for papers of similar quality that
were not paired with a patent. Huang and Murray (2009) examined a larger data set of
patent-paper pairs on gene discoveries from 1988–2006 and observed a relative decline in
citations after patent issuance of about 5 percent, with the decline increasing with number
of claims in the patent. Of particular relevance to the anticommons, the authors observed
that the negative impact of patent grant was greater in the presence of patent thickets
and when ownership of cited patents was fragmented (as suggested by prior art citations
in the patent documents). It is difficult to reconcile the observations in citation studies
of a relative decline in citations after patent issuance with the observations in the survey
studies that scientists ignore patents (Herder, 2012; McManis and Yagi, 2014).
Williams (2013) used a different empirical approach to test the impact of the IP strategy
of the gene sequencing firm Celera, which relied on database access agreements as well
as patents. Celera sequenced the human genome in a race with the publicly funded
Human Genome Project (which required grantees to disclose their data in the publicly
accessible GenBank database) and made the proprietary Celera database available
only to paying subscribers during a period when the GenBank version was incomplete.

Williams compared subsequent publications and diagnostic applications of genes that

were first identified only in the commercial Celera database with subsequent research and
diagnostic applications of genes that were first identified in the freely accessible GenBank
database and found a reduction in subsequent research and product development associ-
ated with the Celera genes of 20–30 percent.
Some observers have concluded from these studies that the effects predicted by the anti-
commons model are not borne out by the available data (Caulfield et al., 2006; McManis
and Yagi, 2014). But the data are inadequate to support this conclusion because most of
these studies merely investigated encounters of research scientists with IP rights without
regard to the existence of conditions that would set the stage for anticommons effects.
With the notable exception of Huang and Murray (2009), who found a greater decline in
citations following patent issuance in the presence of patent thickets, these studies did not
attempt to measure the existence of fragmented or overlapping IP rights held by different
owners, nor did they ask about the impact of such rights on product development.
Although not directly relevant to the impact of a proliferation of IP claims on commer-
cial product development, these studies nonetheless shed an interesting light on transaction
costs in the setting of negotiations over access to upstream research tools that could have
implications in other settings that require users to negotiate with multiple rights holders.
The studies provide considerable evidence of high transaction costs. Murray (2006) docu-
mented protracted negotiations for dealing with just one counterparty over a high-value
research tool. Maurer (2006) recounts ultimately unsuccessful efforts of a community of
academic biologists to overcome bargaining and information obstacles to pool their data
on genetic mutations in a shared database with industry support. Williams (2013) found
further evidence of significant transaction costs in negotiations between Celera and its
database subscribers in the fact that the parties did not address patent licensing issues ex
ante, before subscribers had invested in research that might be covered by Celera’s patents,
but instead waited to negotiate licenses until after further inventions had been made, when
Celera was in a stronger bargaining position. This contradicts the prediction of Green and
Scotchmer (1995) that ex ante licenses will produce optimal incentives and will therefore
be negotiated. Williams (2013) attributed the use of ex post licensing to transaction costs
associated with scarcity of ideas, asymmetric information about development costs, and
the risk that Celera would engage in competitive R&D based on unpatented information
disclosed by prospective licensees in the course of negotiations.
The observation that practical excludability had a greater impact on scientists than
patents alone led Eisenberg (2008) to propose a refinement of the anticommons hypoth-
esis to consider which side bears the burden of overcoming transaction costs. In the case
of patents, rights holders must incur significant costs in order to identify infringers and
assert their rights against them. The higher these costs, the more likely that transaction
costs will protect users against detection and liability, particularly for low-value users
like academic scientists whose activities pose little threat to the commercial interests of
rights holders. In this context high transaction costs on rights holders mitigate the risk
of underuse in an anticommons. On the other hand, for ‘practically excludable’ resources
such as biological materials and databases, users themselves must incur transaction costs
to find rights holders and negotiate for access, while rights holders may avoid search
costs and wait for users to come to them. In this context high transaction costs work
against users, aggravating the risk of underuse in an anticommons. (The effects may be

quite different for high-value users such as commercial product developers, considered in
the next section.) In other words, the presence of high transaction costs does not neces-
sarily increase the risk of underuse in an anticommons. Depending on where the burden
of overcoming transaction costs falls, high transaction costs could either aggravate or
mitigate the risk of underuse.
A corollary is that aggregation of rights could lead to both greater transaction costs
and less use than disaggregated ownership if aggregators face lower assertion costs than
disaggregated owners. This is because an aggregator might be expected to assert rights
that users could otherwise safely ignore. This possibility is considered in Section III below.
B. Impact on Downstream Product Development
There are reasons to predict that the risk of an anticommons could be either greater
or lesser in the context of product development than it is in the context of scientific
research. On one hand, rights holders that are willing to ignore the activities of academic
researchers may be more likely to assert their rights when a product of commercial value
is in sight. Moreover, commercial firms that hope to bring a product to market may
be more concerned about potential infringement liability (and therefore less willing to
ignore patents) than academic scientists. Rights that might go ignored and unasserted in
the context of low-value research may therefore be more likely to require bargaining and
clearance in the context of high-value product development, setting the stage for a pos-
sible anticommons. On the other hand, the higher values at stake in product development
may provide greater motivation for the parties to reach bargains in order to avoid waste,
making bargaining breakdowns less likely.
Although the studies reviewed in the last section focused primarily on the impact of
IP on the work of scientists rather than on product development, some of the studies
included data from scientists in commercial firms and from institutional representatives
other than research scientists. Walsh et al. (2003) included interviews with 12 lawyers,
3 scientists, and 9 business managers from the pharmaceutical industry; 7 lawyers, 4
scientists, and 7 business managers from the biotechnology industry; 7 outside lawyers;
and 5 government and trade association personnel. Nicol and Nielsen (2003) included
interviews of CEOs, IP personnel, and bench scientists from private industry and direc-
tors of research groups in diagnostic testing facilities, as well as outside patent attorneys,
licensing consultants, and government and trade representatives. Although these small
numbers raise questions about whether the data are generalizable, those studies that
included commercial respondents and distinguished them from academic respondents
found that commercial scientists were more likely to report difficulties in negotiating
licenses and a need to alter their research plans because of patents (Walsh et al., 2003;
Nicol and Nielsen, 2003; Hansen et al., 2007).
Scientists responding to surveys may not be in as good a position to observe the impact of
IP on product development as those people who are more directly involved in clearing rights
and deciding which projects to develop. Interviews reported in the literature suggest that
firms incur significant costs in culling through multiple patents to determine what licenses
are necessary, and that these costs are greater in fields characterized by more extensive
patenting (Eisenberg, 2008). For example, a lawyer for a pharmaceutical firm told Walsh et
al. (2003) that lawyers in the small molecule division in his firm were responsible for eight

projects each, while those in the biotechnology division of the same firms only handled about
two projects each because of the greater complexity of the licensing issues that they had
to manage in clearing rights for their projects. Respondents told Nicol and Nielsen (2003)
that because of the increasing breadth of patent claims, it is now necessary to analyse more
patents in detail in order to determine whether they must be licensed; however, at the end
of this analysis, the number of patents that require licensing generally remains small. One
Australian respondent estimated that a typical freedom to operate search might uncover
anywhere from a dozen to 30 or 40 patents that are relevant, but that upon closer analysis
‘there may be only one or two or a few more that are blocking’ and that ‘[a]nything beyond
three is probably too many’ (Nicol and Nielsen, 2003). Walsh et al. (2003), characterizing
the reports of about 10 US industry respondents, said that an initial search would turn up
hundreds of patents that they would have to consider, a number that was surely higher than in
the past, but that after analysis the number that they would need to address would range from
zero to a dozen. As Epstein and Kuhlik (2004) note, a proliferation of patents may represent
competing alternative stepping stones to production rather than a blockade of overlapping
rights. But a firm may need to incur significant information costs in order to figure that out.
How many patents it takes to derail a project evidently depends not only on what
the patents cover, but also on the stage at which they are identified. In both the US and
Australia, respondents indicated that numerous patents would be more likely to deter a
firm from pursuing a project at the outset than they would be to cause the same firm to
abandon a project once it was already underway. A lawyer for a biotechnology firm told
Walsh et al. (2003) that the firm assesses the patent landscape early on, and if it appears
too formidable, ‘the project never gets off the ground,’ but that if patent issues arise later
they are ‘not the show stopper that you would identify early on.’ Reviewing similar stories
from Australian respondents, Nicol and Nielsen (2003) concluded ‘that it is possible that
a number of potentially anticommons-affected projects do not come onto the radar,
because such projects will have been abandoned well before any difficulty of negotiating
with multiple parties is encountered.’
As noted in the last section, some studies have shown a negative effect of patents and
licensing on the development and use of clinical genetic tests (Cho, 2003; Merz et al.,
2002; Walsh et al., 2003; Soini et al., 2008), although it is not clear that the problems
highlighted in these studies necessarily involve fragmented property rights and anticom-
mons tragedies. Some genetic tests may be vulnerable to anticommons risks, such as DNA
microarray products that simultaneously test for multiple genes or mutations covered by
multiple patents (Barton, 2006). This is similar to the challenge of negotiating multiple
copyright licenses for digital sampling of music (McLeod and DiCola, 2011).
Industries may differ in their capacity to manage complex patent landscapes. Ziedonis
(2004) studied patenting in the semiconductor industry, a capital-intensive industry with
considerable patenting. Ziedonis used patent citation data from patents issued to 67
semiconductor firms between 1975 and 1996 to estimate the extent of fragmentation of
ownership of technological inputs, and found that firms acquire patents more aggressively
when fragmentation of ownership is more extensive. The difference was significantly more
pronounced for more capital-intensive firms. Ziedonis explained that amassing a large
patent portfolio could be a strategy for safeguarding capital investments while foregoing
the costs and delays associated with ex ante contracting, by increasing the likelihood that
the firm can threaten others with reciprocal lawsuits (Ziedonis, 2004).

Barnett (2015) observes that although information and communications technology

markets would appear to be fertile ground for an anticommons to arise, in that products
require assembly of multiple components covered by patents with dispersed ownership,
R&D spending and patenting have grown steadily, while consumer prices have declined.
Barnett attributes the avoidance of anticommons effects to the capacity of markets to
arrive at patent pooling and cross-licensing arrangements, and reviews the historical evi-
dence of institutions that arose to aggregate patents on a variety of technologies that had
generated patent thickets. Barnett also sees a pattern of limited pursuit and enforcement
of patent rights within an industry as a kind of market response to overpropertization
that limits anticommons risks (Barnett, 2009).
III. THE RISE OF PATENT AGGREGATORS
If fragmentation of rights creates obstacles to licensing and threatens to interfere with

efficient use, an obvious potential solution is the aggregation of fragmented rights by
one or more owners (Epstein and Kuhlik, 2004; Barnett, 2015). Aggregation reduces
the number of entities that a product developer must bargain with and thereby reduces
the number of potential holdouts, allowing patent owners and technology users to clear
rights with a smaller number of transactions (Lemley and Melamed, 2013). Moreover,
aggregating all the necessary rights in a single owner would solve the double marginaliza-
tion problem because that owner would consider the full impact of royalty demands under
multiple patents (assuming that aggregation reduced the number of owners to one). If
transaction costs stand in the way of efficient use of patented inventions, one might expect
entrepreneurs to figure out ways to reduce these costs in order to realize gains from licens-
ing. Examples abound in the literature of patent pools, cross-licensing arrangements, and
copyright collectives that have reduced the transaction costs associated with a prolifera-
tion of IP rights and facilitated the clearing of rights (Merges, 1996; Barnett, 2015).
Patent aggregators have proliferated over the past 20 years under liberalized antitrust
rules (Gilbert, 2004; Ewing and Feldman, 2012). But this phenomenon has not been
uniformly welcomed as an efficient means of forestalling or correcting an anticommons.
Many commentators have instead lamented the rise of patent aggregators as a new source
of holdup against innovators rather than as an efficient mechanism for clearing rights
(Ewing and Feldman, 2012; Bessen et al., 2011–12; Chien, 2009). Antitrust scholars have
worried that patent pools may have anticompetitive effects if they bring into common
ownership patents that cover substitutes that would otherwise compete in the market
rather than complements that must be combined for use (Shapiro, 2001; Carrier, 2013).
So far, patent aggregators have focused primarily on patents in the information and
communications technology sectors (Hagiu and Yoffie, 2013). But Feldman and Price
(2014) find that conditions are ripe for patent aggregators to play a larger role in the bio-
technology and pharmaceutical sectors by acquiring patents currently held in university
patent portfolios.
Patent aggregators vary in their business models (Hagiu and Yoffie, 2013; Schwartz,
2014). Many firms that innovate and operate in technologically advanced fields accumulate
large patent portfolios to preserve their own freedom to operate across a broader techno-
logical domain. Patent portfolios enhance their bargaining position in negotiations over

patents held by other firms, and thus protect them from costly litigation (Parchomovsky
and Wagner, 2005; Ziedonis, 2004).
Critics have expressed more concerns about so-called non-practicing entities (NPEs)
that make a business out of buying, selling, licensing, and asserting patent rights,
sometimes on a very large scale, without practicing the underlying technology themselves
(and therefore without exposing themselves to counter-assertion of patents held by other
firms) (Bessen et al., 2011–12; Bessen and Meurer, 2014). But it is not clear in principle
that patent portfolios held by NPEs are more pernicious than those held by operating
companies (Lemley and Melamed, 2013). The distinction between NPEs and operating
companies may be illusory, as NPEs may be acting on behalf of the operating companies
that sell patents to them, obtain licenses from them, and invest in them (Ewing and
Feldman, 2012; Lemley and Melamed, 2013).
To the extent that patent aggregators lower the transaction costs of patent assertion,
they may aggravate risks of underuse in fields characterized by extensive patenting.
This is because transaction costs faced by those who assert patents (e.g., the costs of
identifying infringers and pursuing them with demand letters and lawsuits) protect users
by making assertions less likely, thus mitigating the risk of underuse (Eisenberg, 2008,
2011). Patent aggregators may be acquiring patents from owners that would otherwise be
unlikely to assert their rights (McDonough, 2006), perhaps because they lack resources
for litigation. Transfer of these rights to aggregators increases the liability risks faced by
operating companies and shifts to technology users the burden of clearing rights that
they might otherwise have safely ignored. Sale of these patents may increase the rewards
to innovators by providing a market for their rights, but it also increases the risk of an
anticommons. Indeed, Lemley and Melamed (2013) argue that the growth of a second-
ary market for patents has likely increased incentives for inventors to seek patents, thus
contributing to an increase in the total number of patents.
There are many reasons why patent aggregators may face lower assertion costs than the
prior owners whose rights they acquire. They may have greater expertise in the strategic
moves of patent assertion. At the same time, if they are NPEs they may feel undeterred
by industry norms of forbearance from infringement litigation. Litigation costs may also
be higher for operating firms that need to divert the attention of personnel who would
otherwise be available for other more valuable tasks than for firms whose primary business
is patent assertion. Aggregators may also enjoy economies of scale in patent assertion. If a
user receives a demand letter asserting a single patent, it may be worth investing in a prior
art search or opinion letter to determine whether the patent is valid or infringed, but if the
demand letter comes from an aggregator with a large number of patents in its portfolio,
the cost of such scrutiny could be prohibitive. Increasing portfolio size thus diminishes the
likelihood that patent owners will face effective pushback for asserting low quality patents
(Eisenberg, 2011) or patents that are not infringed.5
Put differently, in the absence of a robust secondary market for patents, fragmentation
5
Studies suggest that large non-practicing entities target defendants based on ability to pay
cash settlements rather than likelihood of infringement of asserted patents. See, e.g., Risch (2012)
at 484 (findings of infringement are rare in merits rulings in litigation by non-practicing entities,
consistent with other studies); Cohen et al. (2015) at 17–24 (large non-practicing entities are more
likely to target defendants that are flush with cash irrespective of infringement).

of rights may leave many patents in the hands of owners that are unlikely to overcome
the transaction costs of enforcing their rights. These patents may do little to promote
innovation if they are never asserted or licensed, but they also are unlikely to lead to
underuse of technology, because users can ignore them with little risk of liability. The
rise of patent aggregators creates a market for these patents and consolidates them in the
hands of owners that can enforce the rights at lower cost. This makes assertion more likely,
which in turn increases the transaction costs and information costs facing users who can
no longer assume that it is safe to ignore patents. As the number of patents that might be
asserted against a user grows, the burden of identifying relevant patents and determining
whether they are valid and infringed also grows.
Aggregation may diminish the number of distinct owners of the rights that a user
might be infringing, but if aggregation makes assertion more likely, it could nonetheless
increase the transaction costs burden on users compared to the situation of fragmented
ownership of the same rights. Moreover, new patents are constantly arising, and in the
current market there are likely to be multiple aggregators asserting rights against the same
operating firms. It is thus unlikely that aggregation will entirely relieve users of the risk of
strategic holdouts and royalty-stacking. On the other hand, operating firms may be more
likely than NPEs to use their patent rights for strategic purposes that result in excessive
burdens on competitors (Lemley and Melamed, 2013).
More generally, as large patent portfolios become more common and even necessary for
operating technology companies, firms will continue to pursue larger numbers of patents,
overwhelming the quality control mechanisms of the Patent and Trademark Office, disad-
vantaging small firms and new entrants, and adding to the information costs and transac-
tion costs associated with innovation and the use of new technologies (Parchomovsky
and Wagner, 2005). The growing size of patent portfolios owned by aggregators will also
increase the cost of discerning whether included patents are complements or substitutes,
calling into question the feasibility of using antitrust law to prevent the use of aggregation
for improper anticompetitive purposes (Shapiro, 2001).
If patent aggregation increases transaction costs rather than reducing them, it may
prove more difficult for markets to figure out strategies to mitigate the inefficiencies
arising from a proliferation of patent rights in a field. On the other hand, as Barnett
(2015) and others have recounted, many innovative fields have flourished in the face of
extensive patenting, including information and communications technology in the current
era. Perhaps these fields would have seen even more innovation without having to face the
transaction costs and strategic holdout problems arising from patent thickets. But the rate
of actual technological progress suggests that these problems on balance have not been
tragic in these fields.
IV. CHANGES TO MITIGATE ANTICOMMONS RISKS
The foregoing discussion suggests reasons to worry that the transaction costs of clear-
ing numerous IP rights might pose an obstacle to socially valuable innovations in some
circumstances. These costs arise not only from the burden of finding and bargaining with
multiple rights holders who hold overlapping and fragmented rights, but also from the
need to determine which rights are valid and infringed. The burden of determining validity

and infringement is particularly significant for the patent system because the scope and
validity of patent rights is often unclear (Bessen and Meurer, 2008). Aggregation of patent
rights in the hands of a smaller number of owners may reduce the burden of finding and
negotiating with multiple rights holders and reduce bargaining externalities that can lead
to excessive royalty demands but it does not diminish the costs of determining the scope
and validity of asserted rights.
A number of changes have occurred since the publication of Heller and Eisenberg
(1998) that have mitigated these risks. Some of these changes have come about through
administrative and judicial restrictions on ‘upstream’ patents, some have been byproducts
of legislative and judicial efforts to curtail the power of ‘patent trolls,’ and some have come
from initiatives to improve patent quality. This section reviews some of these changes that
may have forestalled anticommons risks.
A. Limits on ‘Upstream’ Rights and Assertions
The picture of an anticommons as the result of fragmentation suggests poorly designed

rights that are not properly scaled and have boundaries that do not correspond to likely
uses. Heller and Eisenberg (1998) highlighted the example of expressed sequence tags, or
gene fragments, that were unlikely to be useful as standalone products but would only be
used in combinations that could be difficult to assemble in the presence of disaggregated
ownership of rights. More generally, they worried that too many patents on ‘upstream’
research discoveries might impede ‘downstream’ product development.
A number of initiatives sought to curtail a proliferation of patents in the specific con-
text of genomic discoveries. The US National Institutes of Health (NIH), in cooperation
with counterparts from around the world, undertook concerted efforts to ensure that
genomic data generated in the course of the Human Genome Project would be made
promptly and freely available in the public domain (Contreras, 2014). The US Patent and
Trademark Office (PTO) further addressed the particular problem of patent applications
on gene fragments at a time when many such patent applications were pending through
implementing examiner training materials and guidelines that provided a framework for
examiners to reject these applications (US Patent and Trademark Office, 2001). In an
approach that was ultimately approved by the Court of Appeals for the Federal Circuit
(Federal Circuit), the PTO expanded upon its interpretation of the ‘utility’ requirement
for patentability so as to prevent allowance of patents on newly discovered gene fragments
without disclosure of particular biological functions associated with those genes (In re
Fisher, 421 F.3d 1365 (Fed. Cir. 2005)). Restrictions on the patenting of gene fragments
was also the focus of a trilateral study by the European Patent Office, the Japan Patent
Office, and the US PTO (Trilateral Study B3b, 2000).
More generally, Heller and Eisenberg (1998) worried that too many ‘upstream’ patents
procured by recipients of federal research funding following the Bayh-Dole Act of 1980
could paradoxically interfere with ‘downstream’ product development as a result of
aggressive patent assertions. NIH promptly sought to discourage restrictive patenting and
licensing practices on the part of its grantees through a variety of hortatory statements
(Rai and Eisenberg, 2003; Department of Health and Human Services, 1999). Leaders
in the university technology transfer community also sought to address this problem
through dissemination of best practices to promote widespread availability of research

results and to control aggressive assertions of ‘upstream’ patents that might interfere with
downstream product development (California Institute of Technology et al., 2007).
In addition to these efforts within the research community and the science agencies,
the Federal Circuit invalidated a number of prominent university patents through its
interpretations of patent law.6 In particular, the Federal Circuit fortified the statutory
requirement that a patent application must include a ‘written description’ of the invention
in a series of decisions invalidating broad university patents on basic research discoveries
with claims that potentially dominated the products of drug companies (Mueller, 1998).
More recently, the US Supreme Court has limited the availability of patents on
fundamental research discoveries through robust enforcement of traditional exclusions
from patentable subject matter for natural laws, natural products, and abstract ideas.7
The effect is to limit the power of upstream patents over downstream products, protecting
product developing firms against patent assertions and limiting the number of upstream
rights that they need to clear.
B. Limits on Injunctions
A prominent worry in the anticommons literature is that individual patent holders will
try to use their exclusionary rights to extract excessive shares of the potential gains from
aggregation, creating a risk that valuable deals will not get done (Lemley and Shapiro,
2007). This concern has been diminished by the decision of the US Supreme Court in eBay
v. MercExchange (2006), which made it more difficult for patent holders to obtain injunc-
tions against infringers. Prior to that decision, the general rule in the Federal Circuit was
that a prevailing patent holder was ordinarily entitled to a permanent injunction. After
eBay, a patent holder seeking injunctive relief must demonstrate (1) that it has suffered an
irreparable injury, (2) that money damages are inadequate to compensate for that injury,
(3) that the balance of hardships indicates that the equitable remedy of an injunction is
warranted, and (4) that the public interest would not be disserved by an injunction. In
a concurring opinion, four justices noted in particular that a near-automatic injunctive
remedy might give too much power to opportunistic patent holders to hold out for exces-
sive settlement payments.8
6
See, e.g., Regents of the University of California v. Eli Lilly, 119 F.3d 1333 (Fed. Cir. 1997);
University of Rochester v. G.D. Searle, 358 F.3d 916 (Fed. Cir. 2004); Ariad Pharmaceuticals v. Eli
Lilly, 598 F.3d 1336 (Fed. Cir. 2010).
7
See, e.g., Mayo Collaborative Services v. Prometheus Laboratories, 566 U.S. 66 (2012); Ass’n
for Molecular Pathology v. Myriad Genetics, 569 U.S. 576 (2013).
8
See eBay v. MercExchange (2006), 547 U.S. at 396–97:
An industry has developed in which firms use patents not as a basis for producing and selling
goods but, instead, primarily for obtaining licensing fees. . ..For these firms, an injunction, and
the potentially serious sanctions arising from its violation, can be employed as a bargaining tool
to charge exorbitant fees to companies that seek to buy licenses to practice the patents. [ ] When the
patent invention is but a small component of the product the companies seek to produce and the
threat of an injunction is employed simply for undue leverage in negotiations, legal damages may well
be sufficient to compensate for the infringement and an injunction may not serve the public interest.
(concurring opinion of Justice Kennedy, joined by Justice Stevens, Justice Souter, and Justice
Breyer).

This controversial decision arguably puts US patent law in violation of international

agreements that set limits on compulsory licensing of patents (Cotropia, 2008; Dinwoodie
and Dreyfuss, 2012). On the other hand, it diminishes the risks faced by unlicensed users
by weakening the holdup power of patent owners. Product developing firms might there-
fore be more willing to invest in fields characterized by patent thickets without clearing
all potentially infringed rights through ex ante licensing, if they figure that courts will
give them a better deal in setting damages than patent holders would require to relinquish
their injunctions.
C. Lowering Information Costs
Another source of costs for subsequent innovators arising from a proliferation of patents
is the sheer burden of figuring out what patent rights are relevant to a particular project
(Bessen and Meurer, 2008). Although patents are public documents that typically disclose
the identity of the inventor and initial assignee, it is costly to search for relevant patents,
to analyse validity and infringement to determine which rights are necessary to clear, and
to identify current owners (Eisenberg, 2011). These costs may be worth incurring if a
firm is about to launch a lucrative new drug, but not if the values at stake are uncertain
and speculative. Some of these costs could be lowered through law reform measures to
make the scope of patent rights clearer and to maintain updated records of ownership
and assignments.
Although the US patent statute requires that patents include claim language ‘particu-
larly pointing out and distinctly claiming the subject matter which the inventor or a joint
inventor regards as the invention,’ (35 U.S.C. § 112), claim interpretation in the courts
has become a highly indeterminate practice, leaving substantial uncertainty as to what a
patent covers until the conclusion of appellate review (Saunders, 2007; Schwartz, 2008).
The Federal Circuit left considerable room for ambiguity by minimizing the requirement
for definiteness of claim language, holding that a patent claim was invalid for lack of
definiteness only when its language was ‘not amenable to construction’ or ‘insolubly
ambiguous’ (Exxon Research and Engineering Co. v. United States, 265 F.3d 1371, 1375
(2001)).
Recent Supreme Court decisions reversed this approach, replacing the ‘insolubly
ambiguous’ standard with a new rule that ‘a patent is invalid for indefiniteness if its
claims, read in light of the specification delineating the patent, and the prosecution
history, fail to inform, with reasonable certainty those skilled in the art about the scope
of the invention’ (Nautilus v. Biosig Instruments, 134 S. Ct. 2120, 2124 (2014)). Shortly
thereafter, the Supreme Court again reversed the Federal Circuit’s approach to claim
interpretation, holding that the Federal Circuit must defer to factual determinations
made by District Courts in the course of claim interpretation (such as determinations
of the meaning of claim language to a person working in the field of the invention)
(Teva Pharmaceuticals USA v. Sandoz, 135 S. Ct. 831 (2015)). Taken together, these two
decisions promised to make claim interpretation more predictable and to provide greater
certainty at an earlier stage in litigation without the need for appellate review, thus
reducing the costs to innovators of determining whether they are infringing any patent
rights. Initial observations suggest, however, that these decisions have had minimal
impact on the claim interpretation practices of the Federal Circuit (Rantanen, 2015).

On the other hand, the Federal Circuit has been more willing than the Supreme Court
to reduce uncertainty about the scope of patent coverage through restrictions on the
doctrine of equivalents. The doctrine of equivalents is an old infringement doctrine that
predates modern patent claiming practice and allows infringement liability even when the
defendant’s technology is not covered by the literal language of the patent claims if the
differences are insubstantial (Warner-Jenkinson v. Hilton Davis Chemical, 520 U.S. 17
(1997)). The Federal Circuit has significantly curtailed the operation of the doctrine of
equivalents over the years, making the coverage of patents correspond more closely to the
claims (Allison and Lemley, 2007). But literal claim interpretation itself has remained a
stubbornly indeterminate exercise, leaving considerable uncertainty about the technologi-
cal scope of patents.
Uncertainty about the coverage of patent claims is partly due to the inherent difficulty
of defining new technologies in words. But it is also partly a result of strategic use of
ambiguous language by patent applicants in order to retain flexibility in asserting their
patents against future competitors who may use unforeseen technological variations.
This flexibility comes at a considerable cost of uncertainty. The burden and risk for
future innovators rise quickly as the number of patents that might be asserted against
them increases. To the extent that courts make clear that they are unwilling to expand
the effective reach of patents, whether through flexible interpretation of vague claim
language or through liberal use of the doctrine of equivalents, future patent applicants
may find it advantageous to do a better job of defining their inventions clearly in their
claims. If this happens, users will be able to determine more easily just what a patent
covers.
Some reformers have also argued for legal changes that would increase publicly avail-
able information about patent ownership, assignments, and licenses (US Patent and
Trademark Office, 2011). Although current law permits the parties to record a patent
assignment and thereby cut off the interest of any subsequent bona fide purchaser (35
U.S.C. § 261), assignments are rarely recorded, leaving uncertainty about the current
ownership of patents. Recording requirements are more common for interests in real
property. As secondary markets for patents have grown with the rise of patent assertion
entities, this gap in the public record has become more problematic. Future innovators
who cannot determine who owns patents they may be infringing may be unable to clear
rights ex ante, leaving them more vulnerable to holdup after they have incurred sunk
costs (Green and Scotchmer, 1995). Federal recording requirements for patents would
reduce the information costs facing potential infringers who want to clear rights (Glovsky,
2012).
D. Lowering the Costs of Patent Challenges
The Leahy-Smith America Invents Act (2011) introduced new administrative pro-
cedures for challenging the validity of issued patents before an
administrative
tribunal within the PTO.9 These proceedings are quicker and cheaper than
9
These new proceedings include post-grant review during the first nine months after patent
issuance, codified as amended at 35 U.S.C. §§ 321–9, inter partes review more than nine months

litigation10 and have been surprisingly popular and effective as a means of invalidat-
ing patents (Carniaux and Sander, 2015). They offer a number of benefits for patent
challengers relative to litigation, including an accelerated time frame for resolution (12
months from institution of the proceeding until final decision (35 U.S.C. §§ 316(a)(11),
326(a)(11)), rather than years for litigation), and a lower burden of proof (preponder-
ance of the evidence (35 U.S.C. §§ 316(e), 326(e)), rather than the clear and convincing
evidence required to establish invalidity at trial (Microsoft v. i4i Ltd, 131 S. Ct. 2238
(2011))). But even with these advantages, using these new proceedings can still cost
hundreds of thousands of dollars per patent (American Intellectual Property Law
Association, 2015).
The goal of these new proceedings was to address the problem of patent quality, while
the primary concern in the anticommons story is about excessive quantities of patents.
But quantity and quality are related. As the quantity of patent applications has increased,
PTO resource constraints make it more difficult for examiners to scrutinize each applica-
tion with care, setting the stage for allowance of patents that closer scrutiny would reveal
to be invalid. This may be tolerable if most patents are never asserted, and if users can,
at reasonable cost, evaluate those patents that are asserted for validity and infringement
before agreeing to pay royalties. But an opinion letter evaluating validity and infringement
costs approximately $15,000 per patent (American Intellectual Property Law Association,
2015). As the quantity of asserted patents increases, it becomes unrealistic for users to
provide an effective quality check at the point of assertion. If information costs are high
enough, quantity in effect becomes a substitute for quality, giving patent aggregators the
power to extract payments from risk-averse firms without regard to the merits of their
patent infringement claims.
The problems created by large numbers of patents are not limited to the costs of nego-
tiating with multiple owners and are not resolved by aggregation of patents in the hands
of a smaller number of owners. Quite the contrary, aggregation may increase the risk of
underuse by facilitating assertion of patents that could otherwise be ignored. Some of
these patents may be valid, and these rights should be licensed in order to provide a return
to their owners, but others may be invalid or irrelevant to the activities of firms against
which they are asserted. Large numbers of patents threaten to overwhelm the quality
control mechanisms that allow patent owners, technology users, and antitrust authorities
to separate the wheat from the chaff. Determinations of validity and infringement have
remained stubbornly unpredictable, defying efforts to develop lower cost strategies for
figuring out which rights need to be cleared.
after patent issuance, codified as amended at 35 U.S.C. §§ 311–19, and a transitional program in
effect until 2020 utilizing post-grant review for covered business method patents (Leahy-Smith
America Invents Act of 2011, Pub. L. No. 112-29, § 18(a), 125 Stat. 284). See U.S. Patent and
Trademark Office (2012); Menell et al. (2015) at 14-29–14-33.
10
Recent estimates from the American Intellectual Property Law Association (2015) give a
median cost of $350 million for pursuing post-grant proceedings through appeal, and a cost range
for defending a patent infringement action of $500,000 to $375,000.

REFERENCES
Allison, John R., and Mark A. Lemley. 2007. ‘The (Unnoticed) Demise of the Doctrine of Equivalents,’ 59
Stanford Law Review 955–84.
American Intellectual Property Law Association. 2015. Report of the Economic Survey, retrieved on Dec. 22,
2015, http://files.ctctcdn.com/e79ee274201/b6ced6c3-d1ee-4ee7-9873-352dbe08d8fd.pdf.
Aufderheide, Patricia, and Peter Jaszi. 2004. ‘Untold Stories: Creative Consequences of the Rights Clearance
Culture for Documentary Filmmakers,’ retrieved on Dec. 16, 2015, http://www.cmsimpact.org/sites/default/
files/UNTOLDSTORIES_Report.pdf.
Barnett, Jonathan M. 2009. ‘Property as Process: How Innovation Markets Select Innovation Regimes,’ 119
Yale Law Journal 384–457.
Barnett, Jonathan M. 2015. ‘The Anti-Commons Revisited,’ 29 Harvard Journal of Law and Technology 127–203.
Barton, John H. 2006. ‘Emerging Patent Issues in Genomic Diagnostics,’ 24 Nature Biotechnology 939–41.
Bessen, James, and Michael J. Meurer. 2014. ‘The Direct Costs from NPE Patent Disputes,’ 99 Cornell Law
Review 387–424.
Bessen, James, Jennifer Ford, and Michael J. Meurer. 2011–12. ‘The Private and Social Costs of Patent Trolls,’
Regulation 26–35.
Buchanan, James M., and Yong J. Yoon. 2000. ‘Symmetric Tragedies: Commons and Anticommons,’ 43 Journal
of Law & Economics 1–14.
California Institute of Technology, Cornell University, Harvard University, Massachusetts Institute of Technology,
Stanford University, University of California, University of Illinois, Chicago, University of Illinois, Urbana-
Champaign, University of Washington, Wisconsin Alumni Research Foundation, Yale University, and
Association of American Medical Colleges (AAMC). 2007. In the Public Interest: Nine Points to Consider in
Licensing University Technology, retrieved on Dec. 21, 2015, https://otl.stanford.edu/documents/whitepaper-10.
pdf.
Carniaux, Michelle, and Michael E. Sander. 2015. ‘Riding the PTAB Wave: An Empirical Study of the Surge
of Petitions Filed in the Patent Trial and Appeal Board,’ retrieved on Dec. 22, 2015, http://papers.ssrn.com/
sol3/papers.cfm?abstract_id=2650662.
Carrier, Michael A. 2013. ‘Patent Assertion Entities: Six Actions the Antitrust Agencies Can Take,’ 2 CPI
Antitrust Chronicle 2–12.
Caulfield, Timothy, Robert M. Cook-Deegan, F. Scott Kieff, and John P. Walsh. 2006. ‘Evidence and Anecdotes:
An Analysis of Human Gene Patenting Controversies,’ 24 Nature Biotechnology 1091–4.
Chien, Colleen V. 2009. ‘Of Trolls, Davids, Goliaths, and Kings: Narratives and Evidence in the Litigation of
High-Tech Patents,’ 87 North Carolina Law Review 1571.
Cho, Mildred K., Samantha Illangasekare, Meredith A. Weaver, Debra G.B. Leonard, and Jon F. Merz. 2003.
‘Effects of Patents and Licenses on the Provision of Clinical Genetic Testing Services,’ 5 Journal of Molecular
Diagnostics 3–8.
Cohen, Lauren, Umit G. Gurun, and Scott Duke Cominers. 2015. ‘Patent Trolls: Evidence from Targeted Firms,’
Harvard Business School Working Paper No. 15-002, retrieved on Dec. 18, 2015, http://papers.ssrn.com/sol3/
papers.cfm?abstract_id=2464303.
Cohen, Wesley M., and John P. Walsh. 2008. ‘Real Impediments to Academic Biomedical Research,’ 8 Innovation
Policy & Economics 1–30.
Contreras, Jorge. 2014. ‘Constructing the Genome Commons,’ in Brett M. Frischmann, Michael J. Madison,
and Katherine J. Strandburg, eds., Governing Knowledge Commons. Oxford: Oxford University Press.
Cotropia, Christopher A. 2008. ‘Compulsory Licensing under TRIPS and the Supreme Court of the United
States’ Decision in eBay v. MercExchange,’ in Toshiko Takenaka and Rainer Moufang, eds., Patent Law:
A Handbook of Contemporary Research. Cheltenham, UK and Northampton, MA, USA: Edward Elgar
Publishing.
David, Paul A. 2011. ‘Mitigating “Anticommons” Harms to Science and Technology Research.’ Retrieved on
July 29, 2015 from www-siepr.stanford.edu/repec/sip/10-030.pdf. Revised and corrected version of the article
previously published in 1 The WIPO Journal: Analysis of Intellectual Property Issues 59–73 (2010).
Department of Health and Human Services, National Institutes of Health. 1999. ‘Principles and Guidelines for
Recipients of NIH Research Grants and Contracts on Obtaining and Disseminating Biomedical Research
Resources: Final Notice,’ 64 Federal Register 72090–96.
Depoorter, Ben. 2004. ‘The Several Lives of Mickey Mouse: The Expanding Boundaries of Intellectual Property
Law,’ 9 Virginia Journal of Law & Technology 4–65.
Depoorter, Ben, and Francesco Parisi. 2002. ‘Fair Use and Copyright Protection: A Price Theory Explanation,’
21 International Review of Law and Economics 453–73.

Depoorter, Ben, and Sven Vanneste. 2006. ‘Putting Humpty Dumpty Back Together: Experimental Evidence of
Anticommons Tragedies,’ 3 Journal of Law, Economics & Policy 1–23.
Dinwoodie, Graeme B., and Rochelle C. Dreyfuss. 2012. A Neofederalist Vision of TRIPS: The Resilience of the
International Intellectual Property Regime. Oxford: Oxford University Press.
Eisenberg, Rebecca S. 1996. ‘Public Research and Private Development: Patents and Technology Transfer in
Government-Sponsored Research,’ 82 Virginia Law Review 1663–727.
Eisenberg, Rebecca S. 2001. ‘Bargaining over the Transfer of Proprietary Research Tools: Is This Market
Failing or Emerging?,’ in R.C. Dreyfuss, D.L. Zimmerman, and H. First, eds., Expanding the Boundaries of
Intellectual Property: Innovation Policy for the Knowledge Society. Oxford: Oxford University Press.
Eisenberg, Rebecca S. 2008. ‘Noncompliance, Nonenforcement, Nonproblem? Rethinking the Anticommons in
Biomedical Research,’ 45 Houston Law Review 1059–99.
Eisenberg, Rebecca S. 2011. ‘Patent Costs and Unlicensed Use of Patented Inventions,’ 78 University of Chicago
Law Review 53–69.
Epstein, Richard A., and Bruce N. Kuhlik. 2004. ‘Is There a Biomedical Anticommons?,’ 27 Regulation 54–8.
Ewing, Tom, and Robin Feldman. 2012. ‘The Giants Among Us,’ 2012 Stanford Technology Law Review
1–61.
Feldman, Robin, and W. Nicholson Price II. 2014. ‘Patent Trolling: Why Bio & Pharmaceuticals Are at Risk,’
17 Stanford Technology Law Review 773–808.
Fennell, Lee Ann. 2004. ‘Common Interest Tragedies,’ 98 Northwestern University Law Review 907–90.
Galasso, Albert, and Mark Shankerman. 2010. ‘Patent Thickets, Courts, and the Market for Innovation,’ 41
RAND Journal of Economics 472–3.
Gervais, Daniel. 2015. Collective Management of Copyright and Related Rights. 3d ed. Alphen aan den Rijn:
Wolters-Kluwer.
Gilbert, Richard J. 2004. ‘Antitrust for Patent Pools: A Century of Policy Evolution,’ 2004 Stanford Technology
Law Review 3–33.
Glovsky, Susan G.L. 2012. ‘Guest Post: It’s Time for a Reliable System to Determine Who Owns a Patent,’
retrieved on Dec. 21, 2015, http://patentlyo.com/patent/2012/03/patent-recordation.html.
Gold, E. Richard, and Julia Carbone. 2010. ‘Myriad Genetics: In the Eye of the Policy Storm,’ 12 Genetic
Medicine S39–S70.
Green, Jerry, and Suzanne Scotchmer. 1995. ‘On the Division of Profit in Sequential Innovation,’ 26 RAND
Journal of Economics 20–33.
Hagiu, Andrei, and David B. Yoffie. 2013. ‘The New Patent Intermediaries: Platforms, Defensive Aggregators,
and Super-Aggregators,’ 27 Journal of Economic Perspectives 45–66.
Hansen, Stephen A., Michael R. Kisielewski, and Jana L. Asher. 2007. American Ass’n for the Advancement
of Science, ‘Intellectual Property Experiences in the United States Scientific Community,’ available at http://
sippi.aaas.org/Pubs/SIPPI_US_IP_Survey.pdf.
Heller, Michael A. 1997. ‘The Tragedy of the Anticommons: Property in the Transition from Marx to Markets,’
111 Harvard Law Review 621–88.
Heller, Michael A., and Rebecca S. Eisenberg. 1998. ‘Can Patents Deter Innovation? The Anticommons in
Biomedical Research,’ 280 Science 698–701.
Herder, Matthew. 2012. ‘Choice Patents,’ 52 IDEA 309.
Hofmann, Jeanette. 2013. ‘Information Governance in Transition: Lessons to be Learned from Google Books,’
in Ian Brown, ed., Research Handbook on Governance of the Internet. Cheltenham, UK and Northampton,
MA, USA: Edward Elgar Publishing.
Huang, Kenneth G., and Fiona E. Murray. 2009. ‘Does Patent Strategy Shape the Long-Run Supply of Public
Knowledge? Evidence from Human Genetics,’ 52 Academy of Management Journal 1193–221.
Hunter, Dan. 2003. ‘Cyberspace as Place, and the Tragedy of the Digital Anticommons,’ 91 California Law
Review 439–519.
Lee, Jyh-An. 2016. ‘Copyright Divisibility and the Anticommons,’ 32 American University International Law
Review 118–63.
Lei, Zhen, Rakhi Juneja, and Brian D. Wright. 2009. ‘Patents versus Patenting: Implications of Intellectual
Property Protection for Biological Research,’ 27 Nature Biotechnology 36–40.
Lemley, Mark A., and A. Douglas Melamed. 2013. ‘Missing the Forest for the Trolls,’ 113 Columbia Law Review
2117–190.
Lemley, Mark A., and Carl Shapiro. 2007. ‘Patent Holdup and Royalty Stacking,’ 85 Texas Law Review
1991–2049.
Lichtman, Douglas. 2006. ‘Patent Holdouts in the Standard-Setting Process,’ University of Chicago Coase-
Sandor Working Paper Series in Law and Economics, No. 292, retrieved on Dec. 15, 2015, http://ssrn.com/
abstract=902646.
Loren, Lydia Pallas. 2003. ‘Untangling the Web of Music Copyrights,’ 53 Case Western Law Review 673–721.

Maurer, Stephen M. 2006. ‘Inside the Anticommons: Academic Scientists’ Struggle to Build a Commercially
Self-Supporting Human Mutations Database, 1999–2001,’ 35 Research Policy 839–53.
McDonough, James F. III. 2006. ‘The Myth of the Patent Troll: An Alternative View of the Function of Patent
Dealers in an Idea Economy,’ 56 Emory Law Journal 189–228.
McLeod, Kembrew, and Peter DiCola. 2011. Creative License: The Law and Culture of Digital Sampling.
Durham and London: Duke University Press.
McManis, Charles R., and Brian Yagi. 2014. ‘The Bayh-Dole Act and the Anticommons Hypothesis: Round
Three,’ 21 George Mason Law Review 1049–91.
Menell, Peter, et al. 2015. Patent Case Management Judicial Guide. 3d ed. Clause 8 Publishing.
Merges, Robert P. 1996. ‘Contracting into Liability Rules: Intellectual Property Rights and Collective Rights
Organizations,’ 84 California Law Review 1293–393.
Merz, Jon F., Antigone T. Kriss, Debra G.B. Leonard, and Mildred K. Cho. 2002. ‘Diagnostic Testing Fails the
Test: The Pitfalls of Patents Are Illustrated by the Case of Haemochromatosis,’ 415 Nature 577–99.
Mowery, David C., and Arvids A. Ziedonis. 2007. ‘Academic Patents and Materials Transfer Agreements:
Substitutes or Complements?,’ 32 Journal of Technology Transfer 157–72.
Mueller, Janice M. 1998. ‘The Evolving Application of the Written Description Requirement to Biotechnological
Inventions,’ 13 Berkeley Technology Law Journal 615–52.
Murray, Fiona. 2006. ‘The Oncomouse that Roared: Resistance and Accommodation to Patenting in Academic
Science,’ retrieved on March 22, 2019, http://fmurray.scripts.mit.edu/docs/THE_ONCOMOUSE_THAT_
ROARED_FINAL.pdf.
Murray, Fiona, and Scott Stern. 2007. ‘Do Formal Intellectual Property Rights Hinder the Free Flow of
Scientific Knowledge? An Empirical Test of the Anti-Commons Hypothesis,’ 63 Journal of Economic
Behavior and Organization 647–87.
Nicol, Dianne, and Jane Nielsen. 2003. ‘Patents and Medical Biotechnology: An Empirical Analysis of Issues Facing
the Australian Industry,’ retrieved on April 3, 2019, http://papers.ssrn.com/sol3/papers.cfm?abstract_id=2583508.
Parchomovsky, Gideon, and R. Polk Wagner. 2005. ‘Patent Portfolios,’ 154 University of Pennsylvania Law
Review 1–77.
Parente, Michael, and Abel M. Winn. 2012. ‘Bargaining Behavior and the Tragedy of the Anticommons,’ 84
Journal of Economic Behavior & Organization 475–90.
Parisi, Francesco, Norbert Schulz, and Ben Depoorter. 2005. ‘Duality in Property: Commons and Anticommons,’
25 International Review of Law and Economics 578–91.
Parisi, Francesco and Catherine Sevcenko. 2001–02. ‘Lessons from the Anticommons: The Economics of New
York Times Co. v. Tasini,’ 90 Kentucky Law Journal 295–328.
Peters, Marybeth. 2004. ‘Statement of Marybeth Peters the Register of Copyrights before the Subcommittee on
Courts, the Internet and Intellectual Property of the House Committee on the Judiciary,’ 108th Congress, 2d
Session, retrieved on Dec. 17, 2015, http://copyright.gov/docs/regstat031104.html.
Rai, Arti K., and Rebecca S. Eisenberg. 2003. ‘Bayh-Dole Reform and the Progress of Biomedicine,’ 66 Law &
Rantanen, Jason. 2015. ‘Teva, Nautilus, and Change without Change,’ 18 Stanford Technology Law Review
375–93.
Risch, Michael. 2012. ‘Patent Troll Myths,’ 42 Seton Hall Law Review 4267–99.
Samuelson, Pamela. 2011. ‘The Google Book Settlement as Copyright Reform,’ 2011 Wisconsin Law Review
479–562.
Saunders, Michael. 2007. ‘A Survey of Post-Phillips Claim Construction Cases,’ 22 Berkeley Technology Law
Journal 215–40.
Schulz, Norbert, Francesco Parisi, and Ben Depoorter. 2003. ‘Fragmentation of Property: Towards a General
Model,’ 159 Journal of Institutional and Theoretical Economics 594–613.
Schwartz, David L. 2008. ‘Practice Makes Perfect? An Empirical Study of Claim Construction Reversal Rates
in Patent Cases,’ 107 Michigan Law Review 223–84.
Schwartz, David L. 2014. ‘On Mass Patent Aggregators,’ 114 Columbia Law Review Sidebar 51–69.
Shapiro, Carl. 2001. ‘Navigating the Patent Thicket: Cross Licenses, Patent Pools, and Standard Setting,’ in
Adam Jaffe, Josh Lerner, and Scott Stern, eds., Innovation Policy and the Economy, vol. 1. Cambridge, MA:
MIT Press.
Soini, Sirpa, Segolene Ayme, and Gert Matthijs. 2008. ‘Patenting and Licensing in Genetic Testing: Ethical,
Legal and Social Issues,’ 16 European Journal of Human Genetics S10–S50.
Stewart, Steven, and David J. Bjornstad. 2002. ‘An Experimental Investigation of Predictions and Symmetries
in the Tragedies of the Commons and Anticommons,’ Joint Institute for Energy and Environment Report No.
JIEE 2002–07, retrieved on Apr. 3, 2019, http://isse.utk.edu/pdf/jieepubs/2002_07Experiment.pdf.
Straus, Joseph, H. Holzapfel, and M. Lindenmeir. 2004. Genetic Inventions and Patent Law: An Empirical Survey
of Selected German R&D Institutions. Munich: Verlag Medien Design.
Trilateral Study B3b. 2000. Comparative Study on Biotechnology Patent Practices: Theme – Patentability of

DNA Fragments, retrieved on Dec. 21, 2015, http://www.trilateral.net/projects/biotechnology/patentability.

pdf.
U.S. Copyright Office. 2015. Orphan Works and Mass Digitization, retrieved on Dec. 15, 2015, http://copyright.
gov/orphan/reports/orphan-works2015.pdf.
U.S. Patent and Trademark Office. 2001. Utility Examination Guidelines, 66 Federal Register 1092–9.
U.S. Patent and Trademark Office. 2011. Request for Comments on Eliciting More Complete Patent Assignment
Information, 76 Federal Register 72372–4.
U.S. Patent and Trademark Office. 2012. Changes to Implement Inter Partes Review Proceedings, Post-Grant
Review Proceedings, and Transitional Program for Covered Business Method Patents, 77 Federal Register
48680–731.
Review 549–642.
Vanneste, Sven, Alain Van Hiel, Francesco Parisi, and Ben Depoorter. 2006. ‘From “Tragedy” to “Disaster”:
Welfare Effects of Commons and Anticommons Dilemmas,’ 26 International Review of Law and Economics
104–22.
Walsh, John P., Ashish Arora, and Wesley M. Cohen. 2003. ‘Effects of Research Tool Patents and Licensing on
Biomedical Innovation,’ in Wesley M. Cohen and Stephen A. Merrill, eds., Patents in the Knowledge-Based
Economy. Washington: National Academies Press.
Walsh, John P., Wesley M. Cohen, and Charlene Cho. 2007. ‘Where Excludability Matters: Material versus
Intellectual Property in Academic Biomedical Research,’ 36 Research Policy 1184–203.
Williams, Heidi L. 2013. ‘Intellectual Property Rights and Innovation: Evidence from the Human Genome,’ 121
Journal of Political Economy 1–27.
Yeh, Brian T. 2015. ‘Copyright Licensing in Music Distribution, Reproduction, and Public Performance,’
Congressional Research Service, retrieved on Dec. 17, 2015, https://www.fas.org/sgp/crs/misc/RL33631.pdf.
Ziedonis, Rosemarie H. 2004. ‘Don’t Fence Me In: Fragmented Markets for Technology and the Patent
Acquisition Strategies of Firms,’ 50:6 Management Science 804–20.

3. Governing intellectual property
Henry E. Smith* 1
Contents
I. Property as Platform
II. Fluid Property and the Need for Governance
III. Governance of Intellectual Property
A. Patent Versus Copyright
B. Licensing
C. Governance Through Equity
D. Group Institutions
IV. Conclusions
References
The ‘property’ in intellectual property has been nothing if not controversial. Stressing the
property element in intellectual property is often assumed to amount to a call for overly
strong protection for creators and inventors. Intellectual property skeptics believe that
an overemphasis on property in this area results from rent-seeking, mistaken thinking,
or misplaced metaphors from the world of tangible property. Information is different.
Because information is nonrival, exclusion rights over it come at the cost of preventing
harmless use. As long as they do not incur the costs of exclusion rights thought to be
characteristic of property, other institutions already have a head start.
An apparent counterpoint to property talk and property thinking is the notion of
governance. Although ‘governance’ is a term with many meanings, several related
notions stand out in potential contrast to exclusion rights. A commons is often gov-
erned by a group of user-participants, and accordingly Nobel laureate Elinor Ostrom’s
(1990) landmark book exploring this institutional arrangement is titled Governing the
Commons (see also Rose, 2011). What constitutes governance more generally covers
a wide spectrum. We speak of corporate governance and governance in a public law
setting. Closest to intellectual property, information-related activities can be governed
through many types of institutions. These include public regulation and private con-
tract. They even include customs and norms. Putting the idea of governance together
with the inherent drawbacks of exclusive rights in property, and one might say that
there is a presumption in favor of commons or the public domain, which then can be
‘governed’ in various ways.
* Fessenden Professor of Law and Director of the Project on the Foundations of Private Law,
Harvard Law School. Email: hesmith@law.harvard.edu. I thank the participants at the Research
Handbook Conference at the University of California, Berkeley, School of Law for helpful com-
ments and Andrew Lewis for excellent research assistance. All errors are mine.
47

This picture, while in many ways admirably broad, is still incomplete. Far from being at
odds with property, governance is an important aspect of property itself. Exclusion and
governance are typically mixed together in regular property (Smith, 2002). In order to
understand what property does and does not offer in the area of information, we need to
be clear on how these strategies work – and how they can work together.
In this chapter, I will revisit the question of the governance of informational resources
from the point of view of property and systems theory. Instead of being contradictory or
mutually exclusive, property and governance overlap. The governance of uses is one end
of a spectrum of strategies employed within property law. What sets intellectual property
apart is the fluid nature of the resource. As in oil and gas, water, and radio spectrum, the
separation of uses is more difficult and less desirable in general than in the case of land
and tangible personal property. Indeed, property can be seen in part as a law of separa-
tion: it manages conflict and allows for specialization by separating things from their
surroundings, classes of uses from each other, and various functions from each other (e.g.
management and beneficial use) (Smith, 2017). Particularly prominent is the separation
into modules associated with thinghood and the adoption of exclusion strategies for
delineating rights. In the case of fluid resources, the problems with and difficulties of
separation and exclusion lead the law to shift more readily to governance strategies than
in the case of non-fluid property.
This chapter will explore the structure of property and governance in intellectual
property. Section I will set the stage by noting several notions of governance and the
role they play in property theory. Regular property mixes exclusion and governance,
and provides a modular structure for managing the interaction of actors with respect to
things. Section II draws out the special challenges of fluid property, of which intellectual
property is a prominent type, and will show how these special challenges call for complex
forms of governance to facilitate use by multiple parties while containing opportunism.
Governance is often required where multiple regimes overlap in fluid property, especially
when common and private rights interact, in a semicommons. In Section III, I apply this
framework to intellectual property. The role of governance in managing multiple use
explains some differences between patent and copyright, patterns in intellectual property
licensing, the role of equity in intellectual property, and the importance of collective-
rights organizations.
I. PROPERTY AS PLATFORM
Property is such a controversial notion that contention even surrounds what property is.
I will argue for the usefulness of a notion of property that takes the notion of separation
as central. Characteristic of property law is the separation of clusters of resource-related
activities for treatment in partial isolation of their context, in modular fashion. Most
basic and prominent is the separation from legal things from their background context.
This is far from the end of the story, though.
Even in the definition of a legal thing, intellectual property is partially (but only par-
tially) similar to tangible property (Madison, 2005; Smith, 2007a, 2012; Van Houweling,
this volume). What property law requires is a legal ontology – a set of persons, activities,
and things that are the units in which the law is framed. In a world of zero transaction

Governing intellectual property 49
costs (Coase, 1960), we could imagine that the notion of thing would be far less important
than it is in our world. For all possible activities of all people in society, the potential
conflicts between and among them could be handled directly by contract – directly in the
sense that the contracts would make reference to persons and activities without the need
to define more abstract and formal notions like a thing. Proceeding in this fashion would
likely be intractable (can intractability be overcome by contract in the zero transaction
world?), and so in our world separating out clusters of activities into chunks – modules
– through the definition of things is a massive shortcut (Lee and Smith, 2012). Thus, a
person can be protected in a wide variety of activities with respect to a car from a wide
variety of activities of others generally, by defining the rights and duties availing between
the possessor (or owner) and everyone else through the car (Penner, 1997). Not literally
through the car, but through a legal thing that closely corresponds to a car. Indeed, the
legal ontology tracks everyday ontology closely in the case of personal property (although
we need the law of accession to know, for example, that a calf goes to the owner of its
mother) (Smith, 2015a).1 In the case of real property, more effort at legal delineation
is required: physical boundary markers and surveys help define the legal thing we call a
parcel. There is likely to be more error and more dispute here precisely because there is
more delineation required to get from the world of actual things to legal things (Libecap
and Lueck, 2011).
Intangibles and intellectual property are, as is well known, even more difficult to deline-
ate in legal terms (Fromer, 2009; Menell, 2011). Defining things is notoriously challenging
in intellectual property, although, as we will see, these difficulties differ by area of intel-
lectual property. Thus, copyright things (expressions) are quite ethereal, and copyright
law defines rights more in terms of use than does patent law. In principle, an invention
may be easier to define than an expression – and it serves as more of a shorthand over
a wide range of uses (make, use, sell), but the delineation is quite problematic compared
to personal and even real property (Bessen and Meurer, 2008; Menell and Meurer, 2013).
The difficulty of delineating things in this area is closely related to the extended notion
of ‘possession’ required for intangibles, which must key off notions of notice (copyright)
and enablement and other disclosure (patent) (Van Houweling, this volume; Holbrook,
2006). And violations of rights do not announce themselves in the area of rights in
intangibles, the way they do with rights over rival, tangible resources. Nevertheless, what
thing-definition is doing here, however well or poorly, is to pick out sets of activities for
bulk treatment.
To see how indirect thing-definition can be in patent law, consider patents on tennis
rackets. The tennis racket is an everyday object and not a very controversial item in most
people’s ontology, even if there may be borderline cases (a squash racket used for a tennis
game?). When it comes to incremental inventions relating to tennis rackets, the story is
different. In Athletic Alternatives, Inc. v. Prince Mfg., Inc., 73 F.3d 1573 (Fed. Cir. 1996),
a string system design company sued a racket manufacturer, alleging infringement of a
patent on a splayed racket with a string pattern of offsetting attachment locations to the
1
In recent times there has been an upsurge in interest in how the law of accession intersects
with intellectual property (Lee, 2011; Merrill, 2009, pp. 468–9; Newman, 2011; Smith, 2007a,
pp. 1766–77).

racket in which the offset distance ‘varies between’ a minimum and maximum distance.
The allegedly infringing racket had two positions for string attachment that alternated,
and the Federal Circuit upheld summary judgment of no infringement because on a
narrow reading two positions do not ‘vary’ according to the specification. The case
involved a complicated prosecution history and possibly ambiguity in the claim, and the
court was perhaps penalizing ambiguity by adopting the narrower reading. The point in
terms of delineation is that little of this tracks everyday understanding, and as a method
of picking out various activities – by using the ‘invention’ as defined in part by this
exercise in construction – it is very indirect and delicate indeed.
As we will see, because thing-definition in intellectual property does not get as far as
in regular property, governance plays a larger role relative to exclusion in intellectual
property. In patent law, much of the definition has to occur ex post, where uses can be
evaluated. Prominently and controversially, the doctrine of equivalents sometimes adds
scope to prevent evasion of the literal boundaries, and in an earlier era the doctrine of
equivalents played a larger role in cutting back on scope for potentially opportunistic
patentees (Ellis, 1949, § 216).
Alternatively, one might say that thing-definition itself is more a matter of governance
in intellectual property (Collins, 2008). This is dramatically true of trademark. Is the
mark a thing, and if so in what sense? We cannot really know what the thing is unless
we know what others are doing in the ‘vicinity’ – what kinds of marks they are using for
what kinds of business in what locations. Nonetheless, it is safe to say that governance is
doing more work here than it would if it were possible to define a robust thing less tied
to the context of use.
By separating out clusters of activities, we can use a relatively easy-to-measure activity
with respect to a thing as a proxy for harm that triggers the violation of a right over those
activities. This is dramatically so in the case of exclusion strategies. Thus, crossing the
boundary of Blackacre, damaging (or sometimes harmfully touching) an item of personal
property, and making, using, or selling a patented invention, trigger violations in a way
less costly to delineate, observe, and enforce than if one had to separately track the activi-
ties they stand in for. The exclusion strategy has the advantage of relative cheapness, and it
allows the owner to become the residual claimant over those activities. Characteristically,
property law begins with things and defines rough and ready exclusion strategies to
protect their use value: torts like trespass and actions like replevin use relatively simple on/
off signals for violation of the holder’s right. But there is a big price. Because the activities
that serve as a basis for finding violations of the right are just proxies for harm, the proxies
will be over- and underinclusive with respect to the underlying activities we care about.
This is true in the case of tangible property and even more true of intellectual property.
The proxies are underinclusive in that many uses can be interfered with, without the trig-
gering of the boundary crossing: not all nuisances involve boundary crossing by tangible
objects – for example loud noises. The boundary-related proxies are overinclusive in the
sense that not all intruders barred by the law of trespass would cause actual harm to the
possessor’s uses (Ellickson, 1993). And more subtle coordination of uses, such as uniform
door color, cannot be brought about by employing exclusion strategies.
Here is where governance comes in. When a problem involving use conflicts becomes
important enough, it makes sense to shift to the more costly but more refined governance
strategies, which unlike exclusion strategies, do condition violation on proxies closely

tied to types of uses (Smith, 2004, 2007a). Rights to exclude are not absolute in regular
property (Carrier, 2004). Exclusion gives way at some point to governance, and one form
this can take is exceptions to the right to exclude (Merrill and Smith, 2010, pp. 74–84).
The strategies fall along a spectrum; in a sense, exclusion is very crude governance and
governance is fine-grained exclusion. Thus, nuisance, covenants, and zoning all zoom in
on proper use. As we will see, when it comes to fluid property, the move from exclusion to
governance happens not only more quickly, but in characteristic ways.
The goal is to find the least bad proxies in light of both benefits and costs. In the case
of regular property, many, including virtually all law-and-economics commentators, see
property as serving the interest in the use of things. (This should be construed broadly
to include preservation and existence value.) In the case of rival resources, uses conflict,
and property serves in part to resolve those conflicts. If property employs an exclusion
strategy, it is not because there is an inherent value to exclusion; exclusion is only justified
if it is the most cost-effective method of resolving conflicts over uses. Exclusion as a
strategy is indirect (Smith, 2009, 2011): we could imagine, in a zero transaction cost world,
a law that directly resolved the conflict between any two activities. In our world, this
would be prohibitively expensive, though. In addition, property in general and exclusion
strategies in particular serve to protect owners’ investments and to provide a foundation
for contract. The indirectness of the strategy is clear when we consider how these invest-
ments could be rewarded publicly or contracted over privately on a case-by-case basis.
This more direct strategy would require an elaborate system of tracing the contribution
of investment to consumer welfare, and, again proceeding this way would be prohibitively
expensive most of the time.
How does exclusion fit, if at all, into intellectual property? First, there is no alloca-
tion problem because the use of information itself is nonrival.2 Thus ‘conflicting uses’
cannot have the same meaning or role as it does in regular property. Nevertheless, how
to pick proxies in order to protect investments and allow people to contract over their
information-related activities does present a problem for which exclusion can in principle
play some role, leaving open important empirical questions. Thus, in some areas of
intellectual property, with patent law being the most prominent example, some theorists
argue that commercialization is important in addition to or instead of the ‘reward’ that
conventionally forms the rationale for intellectual property (Rich, 1942; Kieff, 2001;
Sichelman, 2010).
Exclusion has an asset-partitioning function in intellectual property (Heald, 2005). By
employing proxies that give a residual claim over the return from a cluster of attributes of
information, one can indirectly define rights to the return from rival inputs to commer-
cialization, such as labor and materials that go into marketing. Moreover, this ability to
appropriate the return from a cluster of attributes can be packaged and transferred, and
it is likewise easier to transfer this bundle than a bundle of residual claims to articulated
rival resources (which would otherwise require tracing through the commercialization
2
The nonrivalness of information counts in favor of certain alternatives to intellectual prop-
erty, if the question is rewards for invention (Abramowicz, 2003; Duffy, 2004). Commercialization
theory, as an alternative to reward theory, is a much more complex question, to which I return
below. Nonrivalness of information also presents problems in getting people to reveal their values
(Yoo, 2007).

process) (Merges, 2005; Liivak, 2013). In general, exclusion promotes alienability in

regular property law and contributes to the function of promoting transactions identified
in the literature on patents.
Exclusion’s indirectness is a double-edged sword. It is a relatively low-cost way to get
at the commercialization activities we care about. It protects residual claims and allows
contracting over them. When negotiations fail or a contract is breached, the baseline is
there. The flipside – and major downside – is that the cruder this proxy is with respect
to what is truly valuable, the greater is the distance between the exclusion strategy and a
fully articulated system of governance or unfair competition and the greater is the waste
in terms of foregone use that the intellectual property skeptics rightly point to as a cost
of intellectual property in the first place.
The question is what combination of exclusion and governance – and no protection –
will be best. This is ultimately an empirical question, but simple institutional models can
make a few predictions and allow for some rough guesses.
To start with, aside from some very narrow detachable problems, it is difficult to serve
commercialization using governance only. I will argue that the shift to governance hap-
pens more quickly in the case of copyright than in patent, but copyright still involves a
thing – protected expression – and is, however tenuously, a property regime.
Why is status as property important at all? Generally speaking, property law manages
the complexity of private interactions over resources by employing the technique of
modularity. In general, modular systems break down complex interactions in a way
that provides for components called modules that contain intense interaction, and
the modules are connected by interfaces that capture relatively sparser interactions
(Baldwin and Clark, 2000; Langlois, 2002). Because the connections between modules
are relatively sparse, work can take place within one or few modules without destabilizing
and hard-to-foresee ripple effects. Indeed, modularity allows for ‘information hiding,’ in
which module-internal information is not relevant to outside processes. In a car, much of
the internal workings of the brake system are invisible to the windshield wiper and air
conditioning systems. Modularization thus permits specialization.
Much of the characteristic modularity of property law starts with the definition
of legal things: defining a legal thing involves setting boundaries around clumps of
interactions (modules) and defining the permitted interface between them (Smith,
2012; Sichelman and Smith, 2017). The kind of separation involved in modular-
ity is typically partial: the system of interactions is not totally decomposable into
freestanding non-interacting components; instead algorithms can be used to choose
modularizations in which interaction is relatively intense within modules but relatively
sparse across interfaces (Simon, 1981; Arora and Barak, 2009). As with modularity
in general, managing complexity in this fashion allows for specialization. In property,
recognizing the owner as being in charge (up to a point) of a legal thing permits the
owner to specialize in investing in and becoming familiar with the asset. The separation
involved in property is a separation into modules and extends to ‘entity property,’ which
is based on asset definition and the separation of control from use (Merrill and Smith
2017, pp. 616–799). Such separation into functional modules facilitates specialization
of management and various monitoring functions. Modularization in property is also
vulnerable to strategic behavior when actors take regard only for what is internalized to
them through modularization.

What is internalized is not just what is internal to the legal thing according to the
exclusion regime; internalization also occurs though the governance regime at the
interface. Thus, in real property, the legal things corresponding to Blackacre not only
have boundaries protected by trespass and the like, but they also feature the rights and
duties of nuisance law, duties of lateral support, any easements and covenants, zoning
regulations, and so on. Without this interface of governance between legal things, owners
would specialize, but myopically, engaging in a lot of overall welfare-reducing behavior.
Employing legal things has to be compared with other methods of managing complex-
ity, such as defining types of activity and using notions of foreseeability and proximate
cause to resolve conflict, as in the law of torts. The question in intellectual property is
not whether thing-definition is a good first cut or lends itself to delineation compared
to regular property – it is definitely worse – but rather how these other more activity-
oriented methods of delineation for information would work. In the case of unfair
competition, they perhaps work better. So the first question at the root of intellectual
property is whether it is more advantageous to take a first cut at defining a thing and
using an exclusion strategy, rather than doing nothing or directly governing in tort-like
fashion. Likewise, in the question of whether intellectual property is property or regula-
tion (Lemley, 2014; Mossoff, 2013), simply pointing to regulation does not answer that
question: some regulation picks up where exclusion leaves off, and like exclusion keys
off the things of a property regime. Other regulation is more free-floating and makes no
special use of thing-based entitlements at all.
This question of choice of strategy cannot be answered in isolation. Two factors are
worth singling out. First, different things could form the platform for further property
institutions. A thing could be defined narrowly or broadly, for example. Second, different
thing-definitions could form a better platform for further governance than others. That
is, ultimately, exclusion and governance work in tandem. Focusing in on particular uses
to make exceptions or supplements to the exclusion regime can partially overcome the
problems of under- and (especially) over-inclusion inherent in an exclusion regime. By the
same token, governance can also make exclusion more effective when it targets strategic
behavior designed to undermine the exclusionary regime. Examples include various
equitable interventions and aspects of the law of nuisance (Kelly, 2011; Smith, 2015b).
Thus, the institutional questions of exclusion and governance arise as a matter of total
costs and benefits and at the margin (Coase, 1960). As in all institutional analysis, we
must compare feasible alternatives. Because these institutions vary along many margins,
we have to consider the total net benefit from any of them. However, when considering
small changes, especially along one relatively isolated dimension, we can make predictions
and recommendations along the one margin (Alston and Mueller, 2014; Smith, 2018).
From a dynamic point of view, modular structures promote evolvability but only
through a range. Thus, if the system is modularized a certain way, one of its great virtues
is that one module can be tweaked or even replaced without causing havoc with the rest of
the system (Simon, 1981; Baldwin and Clark, 2000). One can redesign the brakes on a car
without worrying about the implications for fuel injection or air conditioning. Likewise,
in the patent system, one could add a wider defense of research use to patent infringement
or for satire in copyright or one could get rid of business method patents or extend patents
to sports moves. All of these would be major changes in intellectual property, but because
of the modular structure of entitlements and the legal doctrines, none of them would

implicate every other corner of an area of intellectual property or intellectual property

as a whole. While there is good reason to question whether courts and legislatures get it
right when it comes to intellectual property, the modular structure facilitates change by
making it more feasible without upsetting the entire apple cart.
Modularity’s benefits in terms of facilitating change have their limits: major changes
may not be achievable by local modular tweaks and instead may require remodulariza-
tion. In general, modular systems under a wide variety of circumstances do evolve to
achieve their goals more cost-effectively, but they can become trapped at a local maximum
(Ethiraj and Levinthal, 2004). There may simply not be a path through the addition,
subtraction, and modification of existing modules to climb a higher nearby ‘hill.’ For this,
remodularization is required. Some major changes in property and intellectual property
cannot come from the type of evolution facilitated through modularity. Instead, in the
area of property, remodularizations, such as a redesign of the style of a property system,
have been achieved on rare occasions by legislatures and autocrats. Examples include the
Napoleonic Code and the 1925 legislation in England abolishing the machinery of the
estate system and setting up national registration of land titles (Chang and Smith, 2012).
These changes might be seen as out of equilibrium and are certainly harder to predict
than are more ordinary changes.
It is at the margin that predictions are most feasible. As usual, it is easier to measure
a phenomenon and to make predictions about it when some things can be treated under
the heading ‘all else equal.’ Under a variety of circumstances, there can be a Demsetzian
evolution toward governance. The Demsetz Thesis holds that as potential externalities
emerge with higher resource value, property rights will emerge to internalize them
(Demsetz, 1967). Demsetz’s prime example was the emergence of property rights among
the indigenous people of the Labrador Peninsula with the rise of the fur trade. As beavers
became more valuable, family hunting territories emerged to prevent overhunting.3
Property rights can also emerge because the cost of internalizing externalities drops, as
where the invention of barbed wire led to the fencing of open range in the American West
(Anderson and Hill, 1975). By the same token, falling resource values and rising costs of
delineation lead to an attenuation or abandonment of property rights (Anderson and
Hill, 1975; Haddock and Kiesling, 2002).
The emergence of ‘more property’ can, however, take the form of governance, and if
the modular theory is correct, we should expect change to take place first at the interface
of modules. One way for this to occur is to deal with interactions between modules not
well handled by the exclusionary structure. Hence we get the law of nuisance, various
exceptions to trespass, and the like. In intellectual property, the rise of particular problems
can lead to exceptions to intellectual property, as with the doctrine of fair use, first in the
courts and later through codification.
I do not claim that any of these changes is optimal. What the theory does suggest is that,
to the extent the process responds to efficiency, modular evolution will move locally in
3
Interestingly, the arrangement was more like a semicommons, because others still had the
privilege to take a beaver for personal food consumption (as a sort of insurance against famine).
The norm was to leave the pelt, but the access afforded by this own-consumption privilege was hard
to police (Smith, 2000). And indeed the hunting territories were of such limited effectiveness that
the Hudson Bay Company invested its own resources in conservation (McManus, 1972).

that direction. Thus, we can expect locally efficient in-group contracting over intellectual
property rights (with the possibility of out-group externalities and collusion of the sort
that is of concern in antitrust) (Ellickson, 1991). The locality of custom counsels caution
in taking it up into the law of intellectual property (Rothman, 2013).
II. FLUID PROPERTY AND THE NEED FOR GOVERNANCE
What is special about intellectual property is that it manages a fluid resource. Other fluid
resources include oil and gas, water, and radio spectrum, and, as we will see, the special
costs and benefits of delineating rights in this family of resources lead to characteristic
institutional solutions.
Just as a legal thing is not the same as a physical thing, legal fluidity is like, but only like,
physical fluidity. A fluid continually deforms under shearing stress, and fluids ‘flow’ in the
familiar way, making individual parts of the fluid hard to track. In the physical domain, a
fluid is more difficult to divide than is a solid. A legally fluid resource is one that is likewise
difficult to separate into things – and hence to subject to exclusion strategies. The proxies
one might use to delineate uses are leaky, in the sense that they do not stably pick out the
class of uses we are interested in. This difficulty of separation of uses is related to the
public good characteristics of these resources: they are hard to bound, but multiple use is
very valuable. For fully nonrival resources, the value of the uses is additive, not mutually
subtracting.
Separating classes of uses or even things in the case of fluid resources is particularly
difficult because of the potential for strategic behavior. Because of their nature, fluid
resources are difficult to separate into things. This means that rights defined in terms of
such things will be leaky. When separation works relatively well, we speak of regulating
‘access’: this is what exclusion strategies aim for. In constrast, neither narrow clusters of
uses nor clusters that depend on less strict separation are suited to preventing actors from
using the resource based on regulating access. These more-fingrained approaches to uses
have the benefit of allowing for multiple use, but they leave the door open (as it were) to
harmful uses under the guise of permitted uses (Ellickson, 1993; Smith, 2000). That is,
the more directly rights and duties are defined in terms of narrower classes of uses and
less categorically on the basis of exclusion from a thing, the more we have to police the
specific use classes more closely, which is costly. Often this cost is worth incurring, and
sometimes the strategic behavior is simply not worth curbing at all.
The forms of governance for fluid property carry a signature that stems from the
costs and benefits of delineating property in these resources (Smith, 2016). Because fluid
resources make coordinating uses more important but more difficult in characteristic
ways, we can make better predictions than simply that governance is relatively more
important in intellectual property than in regular property. There is some truth to this
simple generalization (Smith, 2007a), but it leaves a lot out of the picture. First, the shift
to governance is nonetheless true of regular property, and the problems of strategic
behavior in regular property are not wholly unrelated, if less severe than the ones that
arise in intellectual property.
One reason for the emphasis on governance in fluid property is that fluid property often
requires hybrid regimes, mixing public, common, and private elements. Where two or

more property systems apply to different aspects of the same resource, we have to worry
about arbitrage-style behavior. Thus, in a semicommons, a resource is subject to common
and private property and the two interact: actors’ behavior in the commons may be done
with a view to private advantage, and use of private attributes can damage the commons
(Smith, 2000). The prototypical example of a semicommons is the open-field system of
early modern England (and other parts of Europe). There peasants had private rights
to use land for grain-growing in long thin strips but would be required to throw these
open as a common grazing area after harvests and during fallow periods. Ownership in
the private period for grain-growing in bunches of scattered narrow strips has long been
a puzzle, and one possible explanation is that this pattern made the strategic behavior in
the commons more difficult. Individuals could not direct animals in a common herd to
benefit their parcels with manure and damage others’ parcels with excessive trampling.
The configuration in long thin scattered strips was quite inconvenient, but it made the
boundaries effectively invisible during the commons periods (which they would not have
been with compact contiguous parcels). This boundary configuration was a supplement
to the governance rules (stinting) that are frequently found in successful common pool
arrangements (Ostrom, 1990).
Fluid property is particularly likely to be subject to hybrid regimes like the semicom-
mons (Heverly, 2003; Yu, 2005; Loren, 2007; Smith, 2007a, 2007b; Frischmann, 2012;
Barnett, 2010). The virtue of a semicommons is that it allows multiple access, and different
institutions govern access on different scales, at different times, or with respect to different
attributes. The problem is that multiple access precludes heavy reliance on boundaries to
separate the uses. Thus, either we have to put up with use in one regime that is done with
damaging regard to use in another regime, or we need governance strategies to constrain
such behavior. In the case of water, both riparianism and (even) prior appropriation rely
heavily on governance to deal with strategic behavior (Smith, 2007b; Freyfogle, 1989).
Water is partially a public good: some uses are more consumptive than others, and it is
hard to keep them separate. Thus, diversion for consumptive use is not the only purpose
for water; it can be used for navigation, power generation, and so on, some of which have
partial public goods features (Rose, 1990). In the case of water, these various uses and
the rights that track them are highly intertwined. Prior appropriation is considered more
parcelized or exclusionary than riparianism, and there is some truth in this (although
riparianism is exclusionary in the sense that water rights are tied to the package of rights
in riparian land). And yet consider the governance aspects of prior appropriation. In
prior appropriation, what an appropriator for a beneficial use gets is a use right: it is
measured by use and is lost by non-use. It also does not cover non-used water. Water
that returns to the watercourse is appropriable by downstream appropriators. Even if the
downstream appropriators are junior (later in time), a shift of use by the senior upstream
appropriator or a transfer to a new user is not allowed to impact the junior appropriated
rights to the return flow. This highly interlocking and evaluative scheme of governance
allows for more intense appropriation of the water, at the cost of greater inconvenience
for transfers and greater difficulty in shifting to new uses.
If institutions respond to considerations of benefits and costs, we should expect govern-
ance where particular uses are especially important and conflicting, and where they are
relatively easy to treat in separate fashion. First and most straightforwardly, if individual
uses are particularly valuable, they are candidates for separate treatment. In the law of

nuisance, there is a shift from focusing trespass-style on obviously intrusive uses to recon-
ciling conflicting uses, through rules of thumb or some kind of balancing (Smith, 2004).
Easements and covenants pick out very specific sets of uses. And zoning often prescribes
use, historically through separation into discrete areas, and now through land-use regula-
tion or a form of contracting with developers. Although the institutional suppliers are very
different, all of these devices are more fine-grained in terms of the uses they target than
is the exclusionary regime of ad coelum and trespass.4 Again, a Demsetzian trend toward
more property can take the form of more governance rather than more exclusion.
In intellectual property, the nonrivalness of information and the cumulative benefits of
the public domain lead to picking out uses directly. Thus, to take an extreme example, fair
use is a highly complex free compulsory license for the public (Gordon, 1982; Leval, 1990).
The uses involved are very valuable and make it worth incurring high delineation costs, espe-
cially in terms of governance. This governance can take the form of affirmative rights in the
public to use information in certain ways, rather than undelineated privileges that would be
more easily displaced. Off-the-rack law is not the only source of governance. Below we will
return to the question of intellectual property licensing and collective-rights organizations.
For simple models like that in Demsetz (1967) to hold, much has to be held constant.
For one thing, the Demsetz account is a demand-side story: it treats the supply of property
rights as a black box, and any difficulties in establishing rights are undifferentiated costs.
The problematic dynamic of establishing property rights in groups or society as a whole
is left unexplored. Others have brought supply considerations in (Libecap, 1989; Ostrom,
1990; Wyman, 2005). For fluid resources, a more fine-grained look at supply is required.
Indeed, controversy has raged in intellectual property circles about whether licensing is
likely and whether potential anticommons can be overcome (see below). Further, skepti-
cism about intellectual property sometimes extends to suspicion of licensing itself: might
licensing be used not to bargain over relaxing exclusion but rather to extend the owner’s
control beyond the optimal point, or at least the one that Congress has decided upon
(compare Burk and Cohen, 2001; Cohen, 2003 with Friedman, 1998)? Individual licensing
and collective-rights organizations are thus central to whether intellectual property is a
success or a failure. These are all governance regimes, but their costs and benefits are
closely associated with their institutional sources.
Nevertheless, the value of uses does not point unambiguously to more articulated
governance. We need to compare the costs and benefits of separating out a narrow
class of uses, versus affording owners broader exclusion rights, perhaps with defined
exceptions. Exclusion rights are less good at solving use questions directly, but they
economize on delineating uses in a fine-grained fashion. Exclusion rights also do not
preclude governance by contract, individual, or group, and where contracting is lower
(higher) cost, exclusion becomes more (less) attractive. Ultimately, to put it in economic,
terms, the question is what combination of exclusion and governance, produced by which
institution, is the most efficient.
4
The ad coelum principle is based on the maxim ‘cujus est solum, ejus est usque ad coelum et
ad inferos,’ which translates as ‘one who owns the soil owns also to the sky and to the depths’), and
provides that the boundaries of a parcel at least presumptively extend upwards and downwards,
with adjustments – notably for airplane overflights (Merrill and Smith, 2017, pp. 10–16, 142–50).

III. GOVERNANCE OF INTELLECTUAL PROPERTY
Because of its fluid nature, subjecting information to property rights is both controversial
and challenging. As we will see, fluidity tends to require more extreme solutions than in
the case of regular property. Depending on the type of use and the cost of contracting, we
expect a greater role for the public domain, broader exclusion rights, or more detailed gov-
ernance, or some combination of the three. What combination will actually occur is in large
part an empirical question. Especially where multiple use is very important, hybrid rights
regimes are valuable and require extra intervention to protect them from strategic behavior.
In this section, I focus on several applications of this framework that highlight the
importance of governance in intellectual property. First, the different costs and benefits
of delineating classes of uses will help solve the puzzle of why patent and copyright are
quite different, especially in that the former is – in a sense we can now define – more
property-like than the latter. Second, the promise and perils of licensing in intellectual
property and its crucial differences from regular property licenses will receive an explana-
tion. Third, I will show how the system of rights in information is protected by equity as
a second-order intervention where the law fails. Fourth and finally, as with fluid property
regimes like water law, governance also is effected through collective-rights organizations.
A. Patent Versus Copyright
Patent is overall more property-like than copyright. Theories of governance can help us
explain why.
All intellectual property, like property in general, is aimed at protecting uses. The
activities that count as uses in the area of invention (patent) and expression (copyright)
are different, and they vary in the ease with which they can be addressed. When it comes
to inventions, the invention itself is nonrival, but the resources that go into creating
and commercializing inventions are rival. (Although the uses of the information do not
conflict, the uses of these other resources involved in commercialization would conflict.)
The need to address these more remote rival input resources can push toward some form
of property rights in information (Kitch, 1977, pp. 275–6), but this still leaves the question
of what form these rights will take. The form of property in patent and copyright is very
different: both are, like most property regimes, a mixture of exclusion and governance, but
the emphasis is more toward exclusion in the case of patent and more toward governance
for copyright (Smith, 2007a). The activities surrounding inventions – especially those
involved in commercialization – are uncertain and interacting. Correspondingly, we have
a large specialty of expert valuation in patent but less so in copyright. Indeed, now that
in patent law there is a greater emphasis on damages relative to injunctions than there
was ten years ago, the question of how to value the contribution of a patent to an overall
product cannot be avoided (Lemley and Shapiro, 2007, 2013; Golden, 2010, pp. 582–6;
Sidak, 2013; Sichelman, 2014), and the current state of the law is widely regarded as
inadequate (Taylor, 2014). That a patent, especially in the hands of a ‘troll,’ might ex
post be costly to avoid violating – factories have been built, standards have been set – and
yet the patent contributes only in a small way to the value of a product (and could have
been avoided at much lower costs ex ante than ex post) is the very problem that leads to
a suspicion of injunctions in the first place. To date no clean solution for this valuation

problem has been found (Golden, 2010, p. 508). Again, this is characteristic of patent, and
aside from works incorporating many copyrighted works, it is much less true of copyright.
Correspondingly, the role of commercialization theory is much greater in patent than in
copyright in the first place.
Because the set of these invention-related activities is broad and uncertain, and many
such activities must be coordinated, patent law bunches these uses together in a more
exclusionary regime. Rights are based on a notional thing – the invention – and making,
using, or selling the invention constitutes the violation (35 U.S.C. § 271). That is not to
say that the violation is harmful itself. Rather, the system of rights based on this system
of proxies – here the legal thing and violation defined in terms of any use of it – serves
indirectly to protect the rival resources used in the creation and commercial development
of the information.
In copyright, by contrast, individual uses are not as hard to pick out. The basic entitle-
ment in copyright is a list of use rights (17 U.S.C. §§ 106–106A) – very unlike property in
general. In property, we can say that the essence of property is indefiniteness.5 The bundle
of rights conceived as a list of use rights has come in for criticism for not capturing well
how we actually delineate most property (for a variety of views see Symposium, 2011).
But it does a reasonable job of capturing copyright. At the same time, defining a thing in
copyright is not particularly easy (Long, 2004), and relative to focusing in on the activities
we are interested in – in a governance regime – putting heavy emphasis on defining an
expression as a thing does not make sense.
Even some of the pathologies of copyright stem from this greater emphasis on govern-
ance. The story of copyright is a series of enactments for and against protection, lobbied
for by interested parties at various points (Litman, 1987). One reason this is possible is
that individual uses can be separated out and legislatively bargained over and delineated
in legislation. Examples include the compulsory licenses set forth in the copyright act (see
below). Also, if lobbying effort can coalesce upon a well-defined narrow set of uses, this
helps in organizing the lobbying groups and in making their pitch.
Within patent and copyright, the relative emphasis on exclusion and governance also
tracks the predictions of information theory. In patent law, chemical invention is less
problematic from a notice point of view than are patents in electronics or software (Bessen
and Meurer, 2008; Menell and Meurer, 2013). Exclusion regimes depend on simple and
stable proxies to define violations. Stating the invention in terms of a chemical formula
and formulating violation in terms of any use of a substance describable with that formula
is a prime example of an exclusion regime. Note that at the same time we allow such claims
some indefiniteness and breadth on the dimension of function: if an inventor comes up
with a new substance, the patent covers undisclosed and even heretofore unknown uses
of that substance.
In copyright, governance breaks down where we expect. Thus, the part of copyright
that involves the most uncertain and interlocking set of activities is the area of derivative
works. A new derivative work can incorporate many copyrighted works. Here valuation
problems start looking more like they do in patent law. And when it comes to derivative
5
Austin (1873, p. 827) (‘[I]ndefiniteness is of the very essence of the right; and implies that the
right . . . cannot be determined by exact and positive circumscription’).

works, commercialization theory starts being cited as a rationale for protection. These
notorious difficulties with derivative works stem partly from trying to solve with a govern-
ance regime the problems of multiple hard-to-foresee uses.
Some well-known differences between patent and copyright line up along the exclusion-
governance spectrum. In patent law, independent invention is usually not a defense, which
makes the proxy upon which violation is based much simpler, thus contributing to the
exclusionary aspect of the regime. In copyright, independent creation is a defense, and
it requires evaluation of access and similarity in order to tell whether a violation of the
copyright has occurred.
Likewise, the greater reliance on compulsory licenses in copyright is consistent with
an emphasis on governance. In copyright, statutory provisions provide for compulsory
licenses for secondary transmission by cable television (17 U.S.C. § 111(c)–(e)), produc-
tion and distribution of phonorecords of musical works (§ 115), use by noncommercial
broadcasters (§ 118), satellite retransmission (§ 119), and manufacturing and importing
of digital audiotape devices (Audio Home Recording Act of 1992, Pub. L. No. 102-563,
106 Stat. 4237 (codified at 17 U.S.C. §§ 1001–10)). Delineating licenses requires specify-
ing a class of covered uses, and compulsory licenses that are liability rules – one can use
the copyright and pay a statutory fee – require measurement of uses in terms of value.
That is relatively easier where the use is separable, and harder where it is uncertain and
intertwined with other uses. By contrast, compulsory licenses were (at least until recently)
rare in patent law despite some obvious immediate benefits.6
Finally, fair use is a complex governance regime, which we find only in copyright.
Despite occasional calls for fair use in patent and regular property (O’Rourke, 2000;
Depoorter, 2011), fair use is characteristic of copyright. Fair use requires a multifactor
evaluation of the use in question:
(1) the purpose and character of the use, including whether such use is of a commercial nature
or is for nonprofit educational purposes;
(2) the nature of the copyrighted work;
(3) the amount and substantiality of the portion used in relation to the copyrighted work as a
whole; and
(4) the effect of the use upon the potential market for or value of the copyrighted work.
(17 U.S.C. § 107)
Fair use in patent law would require measurement of uses in a way that has so far proved
difficult.
Although exclusion and governance can sometimes substitute at the margin, we have
to consider multiple approaches to combining exclusion and governance. Thus the
strength (and weakness) of patent law is its indirectness, which allows one to capture
the return from commercializing information. Nothing requires that this be done in one
regime. Extending the approach in the legislation on orphan drugs, Ted Sichelman (2010)
6
‘[C]ompulsory licensing is a rarity in our patent system,’ and ‘compulsory licensing of pat-
ents has often been proposed but never enacted on a broad scale’ (Dawson Chem. Co. v. Rohm &
Haas Co., 448 U.S. 176, 215 and n.21 (1980) (citations omitted)). A limited amount of compulsory
licensing is allowed under TRIPS (Agreement on Trade-Related Aspects of Intellectual Property
Rights, art. 8, 1994, 33 I.L.M. 81). One could see the decision in eBay v. MercExchange as a move
toward compulsory licensing. I return to this question below.

proposes a regime for exclusive rights targeted just to commercialization. This is more
fine-grained in two respects: it requires two regimes rather than one, and it requires an
estimate of the incentive needed to commercialize. Whether to incur these costs – rather
than to stick with a cruder regime that depends more on the indirect residual claim of
the current patent – is an empirical question. It is not obvious whether and to what
extent judges should be asked to calibrate intellectual property rights (Chiang, 2017).
Brett Frischmann (2012, pp. 253–305) proposes that in light of a framework based on
intellectual property as a semicommons, patent should look more like copyright. Again,
the advisability of this turns on whether new conditions, for example in software, call for
more governance to override the exclusion regime in patent law.
B. Licensing
Licensing illustrates the promise and limits of governance of intellectual property.

Licenses pick out specific uses and allow for modification of the basic exclusion right,
transaction costs permitting (Barnett, 2010, 2017; Feldman and Lemley, 2015). Compared
with licenses in real property, intellectual property licenses are often more robust and are
comparable in some ways to easements in land.
Like licenses in regular property, intellectual property licenses allow for a regime that
combines property and contract. Interestingly, also as in regular property, there is a lot
of confusion about what exactly a license is. Because many licenses in real property and
most licenses in intellectual property are created by contract, there is a natural tendency
to equate licenses and contracts. This is not entirely accurate, and it can matter greatly in
some contexts (Newman, 2013).
Licenses are more property-related than the contracts that create them. Thus, if a copy-
right licensee exceeds the terms of its license, the copyright holder could seek the remedies
available for a breach of copyright, including in many cases an injunction, even if the
contract itself would not be eligible for enforcement by specific performance (Newman,
2013). Further, if the copyright holder transfers the copyright to another, it is subject to
existing licenses. In real property this running with the asset would happen automatically
with easements appurtenant and would happen for some covenants that pass certain tests.
Covenants on personal property have always had an ambiguous status, especially when it
comes to whether they run (Chafee, 1928, 1956; Robinson, 2004; Van Houweling, 2008).
Licenses also raise issues particular to governance of information. Thus, one way
of thinking about easements, covenants, and licenses, is as sticks in the bundle held by
the owner and ready for distribution or retention. Some have objected that this is an
unrealistic picture of these lesser interests (or non-interest permissions in the case of real
property licenses) (Penner, 2013). Instead, the owner is exercising a power to create a new
legal relation. In the case of an easement, the owner is creating a new in rem right, which
in theory could be enforced against third parties.
Licenses, in real property and in intellectual property, are not in rem, although through
technological means they start looking that way. In the famous case of ProCD, Inc. v.
Zeidenberg, 86 F.3d 1447 (7th Cir. 1996), Judge Easterbrook found that the copyright
license was (or was like) a contract, and because it was in personam it could not be
similar enough to the in rem copyright for purposes of preemption. Nevertheless, despite
the heavy reliance in ProCD on the distinction between in rem and in personam rights,

questions remain about how they differ in practice and how governance can become a
means of achieving exclusion. One mode of governance is technological. The holder
of information can ration access using Digital Rights Management. If this method is
cheap and effective enough, does it become possible to achieve not just exclusion but in
rem rights – good against everyone else – by means of mass producing more in personam
governance relations? If one can bring it about through technological means that each
individual using the information has to agree to the ‘owner’s’ terms, does this mean that
the latter has achieved property? Does this upset the balance struck by copyright legisla-
tion? Consider fair use. If an author or publisher could restrict access to a book such that
a reviewer could not read it, would that violate a right of the reviewer’s? It does, if we take
seriously fair use as a robust affirmative regime governing use.
Governance through licenses presents a problem of information costs for third par-
ties. The more that licenses avail against third parties and the more they run with assets,
the more impact they will have from an informational point of view. In the case of real
property, even easements which are theoretically enforceable against third parties are
constrained so that they do not present much more of an informational burden than the
underlying fee in the land itself. An easement cannot easily govern behavior occurring
off the land in question. (Negative easements are strictly rationed, in part for these
reasons of open-endedness and potential lack of notice.) Thus, because of the nature
of the resource (spatial) and reinforced by certain doctrines in the case of covenants
(e.g. touch and concern), servitudes have few implications for third parties beyond the
basic message sent by boundaries and trespass law. Licenses in real property are not
enforceable against third parties and do not run, so they too present few third-party
information costs.
Further, real property servitudes and licenses do not present huge problems when
parcels are split and combined. Because servitudes and licenses apply to a given area,
they simply add up. If there is an easement over Parcel A in favor of Parcel B, and then
Parcel A is merged with Parcel C, the easement is still located where it was. Nothing about
the scope of the easement changes. The main problem is one of potential misuse. If O
owns Parcel A with an easement over Parcel B and then acquires Parcel D next to A, the
traditional rule mandated that O not use the easement to get to Parcel D. Such use would
be misuse. This is clearly not easy to police, and proposals have been made to measure use
by reasonableness, rather than location (Strang, 2008). Either way, as parcels are scaled up
and down, there is not much in the way of confusion or contradiction.
The situation is very different with intellectual property licenses. The special challenge
of servitudes in fluid property-like intellectual property is well illustrated in Molly Van
Houweling’s (2008) analysis of the ‘new servitudes’ in intellectual property. In addition
to exacerbated problems of notice in intellectual property, she shows how licenses can
conflict downstream. For example, the GPL Version 2, under which the original Linux
kernel was created and licensed, and the Wikipedia GNU Free Documentation License
require downstream users creating works incorporating the licensed material to make
available the new work on the same terms as the input work. However, the terms requir-
ing openness are detailed, and later versions of the license sometimes require conflicting
actions. A downstream work thus may incorporate works with conflicting mandates.
And like other fluid property, we want the governance rules to interlock tightly, but the
problem then becomes dynamic change. Indeed, an involved and controversial process

was required to make it possible for the Wikimedia Foundation to relicense in order to
overcome this problem (Elkin-Koren, 2011, pp. 340–41).
In addition, as with other fluid property, the high cost of exclusion – based on a legal
thing – leads to governance being built on a shakier foundation of exclusion. Indeed, as
Van Houweling points out, we should worry when intellectual property servitudes exceed
the entitlement baseline. Thus, while the Creative Commons license employs the (admit-
tedly fuzzy) copyright baseline, Microsoft’s Vista EULA, by contrast, exceeded the scope
of its copyright. Because this complicates the modularity of the legal rights, it comes at a
cost not internalized to its creator. In a sense, intellectual property poses a special problem
of standardization (numerus clausus, Merrill and Smith, 2000) that the ‘bubbles’ around
parcels of land and items of personal property prevent in the case of regular property.
First sale in copyright and exhaustion in patent respond in part to these problems of
third-party information costs. If that is their role, it suggests that current doctrine may
be too strong.
In copyright, first sale doctrine allows the purchaser of a copyrighted work to resell
it free from restrictions of the copyright holder. Similarly in patent law, the purchase of
a patented article from the patentee ‘exhausts’ the patentee’s rights, and the owner of
the good can resell it without restriction. These doctrines resonate with the traditional
suspicion of servitudes on personal property (see above). As with servitudes on personal
property, the problem is that these restrictions do not announce themselves and there is
no registry to publicize them. Marking on the good itself is not standardized. However, as
we have seen, intellectual property holders often try to accomplish through licensing what
cannot be accomplished through property law, in effect substituting elaborate governance
for exclusion or for the law’s version of governance. What if the intellectual property
holder licenses the buyer but refuses to license those contracting with the buyer, especially
if the good itself is not sold but manufactured under a license? Can the intellectual prop-
erty owner go after the licensee’s customers? After some ambiguity, in Quanta Computer
v. LG Electronics, Inc., 553 U.S. 617 (2008), the answer from the US Supreme Court is
an undertheorized no. Moreover, like first sale, exhaustion is a mandatory rather than a
default rule; it cannot be contracted around through licensing (Impression Products, Inc.
v. Lexmark International, Inc., 137 S. Ct. 1523 (2017)).
The institutional and informational approach can shed some light on the question of
why we have first sale and exhaustion doctrines and how strong they should be. (Again, we
are in the realm of empirical guesswork for now.) One possible position is the pre-Quanta
Federal Circuit’s contractarianism, in which exhaustion would not apply to a method
patent, leaving it open to a patent holder to license a firm but not the firm’s customers
and to require them to acquire their own licenses. In reversing in Quanta, the US Supreme
Court took the opposite approach, holding that exhaustion is a mandatory rule preclud-
ing restrictions of any kind further downstream from a licensee. Both positions are
problematic. Full contractarianism does not reckon with the notice problems, especially
to consumers, of licensing restrictions, and it foregoes the convenience of the anchor of a
physical object as a locus of restriction-free use. The holding in Quanta, however, precludes
more fine-grained multi-party governance regimes, which could be valuable in multiple
stages of production among parties that are well aware of the need for a license. And there
is no shortage of potential positions between these poles that could trade off notice costs
and governance benefits. For example, Van Houweling (2008, pp. 932–9) argues for the

distinction made in earlier Supreme Court cases between complex licensing that allows for
restrictions on downstream commercial producing entities on the one hand and prohibits
them for individual consumers (and perhaps bailees like repair shops) on the other. The
former can be expected to be aware of the need for a license more than the latter, who have
pre-existing expectations of being able to use the items of personal property they own.
C. Governance Through Equity
Exclusion is sometimes vulnerable to opportunism, and to counter such opportunism, a

characteristic governance regime comes into play: equity. Although not all such oppor-
tunism calls for equitable intervention, we generally have to worry about actors who are
too aware of property law’s modular structures and can exploit misalignment of proxies
and true value. Thus, in a fish catch limit, we need to address the potential strategic
behavior of fishers who might kill and throw back small fish in order to ‘meet’ their quota
with only more valuable larger fish.
In intellectual property, strategic behavior is a major problem. In general, equity can be
seen as a second-order intervention when problems like opportunism show the weakness
in the regular law (Smith, 2015b). Equity in this sense is much like the civilian doctrine
of abuse of right.
Much of the problem of strategic behavior or opportunism in intellectual property
surfaces in the area of remedies. And this problem of opportunism points to the useful-
ness of traditional equitable standards.
Opportunism can take the form of unexpected and unfair use of leverage. The problem
of excessive leverage is not unique to intellectual property, even though it may be more
severe. The problem of coupling an injunction with exclusion is that the severity of the
remedy can call forth opportunism, in which the right holder inflicts harm on a good faith
violator out of all proportion to the injury or the blameworthiness of the violation. In
real property, an injunction will not issue for a mere trespass, but repeated and continuous
violations do presumptively lead to injunctions. If, however, the violation was in good
faith and the injunction would inflict disproportionate hardship – that is, much more
harm on the violator than it would benefit the movant – a court can withhold the injunc-
tion and just give damages (a general approach in equity, see Laycock, 2011). Bad faith
violators still face injunctions. This scheme of shifting presumptions is designed to afford
protection to rights, with a safety valve for certain good faith violations, which prevents
excessive precaution on the one side and overenforcement on the other.
The problem of so-called patent trolls exhibits a similar structure. Unfortunately,
instead of applying the traditional equitable set of presumptions, modified perhaps to
make good faith easier to find where patent notice is problematic, the US Supreme Court
devised what it termed a ‘traditional’ equitable four-part test for injunctions in eBay Inc.
v. MercExchange, L.L.C., 547 U.S. 388 (2006), which is actually based on the test for
preliminary relief. Under this test, the movant must show:
(1) that it has suffered an irreparable injury; (2) that remedies available at law, such as monetary
damages, are inadequate to compensate for that injury; (3) that, considering the balance of
hardships between the plaintiff and defendant, a remedy in equity is warranted; and (4) that the
public interest would not be disserved by a permanent injunction. (547 U.S. at 391)

This test is problematic for many reasons. It makes no reference to good faith, which was
one of the main methods of targeting in the traditional test. Moreover, ‘disproportionate
hardship’ meant overwhelming disparity: it is a safety valve, not a balancing test seeking
equipoise.7 The Federal Circuit may well have been wrong to apply a near-automatic
presumption for injunctions, but the eBay test’s implicit agnosticism about violation and
the lack of presumption for an injunction in cases of violation are not necessary as long as
the (targeted) safety valve of disproportionate hardship is wide enough. Disproprtionate
hardship is one possible tool to deal with holdup.8
Indeed there is an economic case to be made for the traditional disproportionate hard-
ship defense in the area of patents (Schwartz, 1964, pp. 1045–6; Denicolò et al., 2008;
Smith, 2009, pp. 2125–33). Disproportionate hardship is a proxy for opportunism, and is
a targeted way to ‘reset’ the balance of leverage or achieve ‘proportionality’ between rights
and leverage (cf. Merges, 2011, pp. 165–9). Moreover the (actually) traditional approach
to injunctions took into account the potential for opportunism on both sides: make the
safety valve too large, and opportunistic violators would use the possibility of under-
compensation by a court as a weapon (Miller, 2007, p. 390). As with any liability rule, we
have to worry that a violator will cherry-pick undervalued assets, and take advantage of
the difficulty of proving damages – as would be the case with a patent that has not been
licensed before.
D. Group Institutions
As in other fluid property regimes, the need for governance often calls for collective-action
organizations. In regular property, common property is located between private property
and open access. Unlike open access, in common property a group manages the resource.
Unlike with private property, this group is larger than the usual groups of co-owners (and
the group does not create an artificial person to own the asset). Instead, the exclusion
regime over the thing and group membership serve as a foundation for the group to build
an often elaborate regime of governance. This is ‘governance’ in Ostrom’s (1990) sense
and also ‘governance’ in the sense that these group-initiated regimes are couched in terms
7
Thus, interestingly, from a traditional equitable point of view all sides in eBay were off the
mark. Disproportionate hardship did not mean asking which party would be would hurt more by
an injunction, as the skeptics of injunctive relief argued (eBay, Inc. v. MercExchange, L.L.C., 2006
WL 1785363 (Appellate Brief) (U.S. Jun. 26, 2006); Brief Amici Curiae of 52 Intellectual Property
Professors in Support of Petitioners (NO. 05-130), at 9). On the other hand, the Federal Circuit’s
automatic injunction rule ignores the extent of the disproportionate hardship defense, as did its
proponents in the Supreme Court, who ironically looked to building encroachments as an analogy
for automatic injunctions. Epstein et al.’s amici brief analogized patent infringement to trespass
and stated that ‘[t]he rule [for injunctions in building encroachment cases] is virtually automatic
when the infringement is deliberate, and strict even when it is not’ (eBay, Inc. v. MercExchange,
L.L.C., 2006 WL 639164 (Appellate Brief) (U.S. Mar. 10, 2006) Brief of Various Law & Economics
Professors as Amici Curiae in Support of Respondent (NO. 05-130) (Richard A. Epstein, F. Scott
Kieff & R. Polk Wagner); Epstein (2010, pp. 488–9)). However, Epstein had advocated dispropor-
tionate hardship as exception to a presumption for property rules in an earlier paper (1997, p. 2102).
8
Other possible tools include other equitable devices like estoppel (Smith, 2013), as well as
denying injunctions in a more categorical fashion (Lee and Melamed, 2016).

of use. After all, the members of the group have access; the question is how they will
coordinate their use.
The Ostrom framework is now being applied to problems of knowledge dissemination
and use, of which intellectual property is only a part. (See the case studies in Frischmann
et al., 2014.) Whether exclusion is needed as a platform for governance depends in part
on how easy the knowledge problem is to modularize in the first place. Thus, in open
source software, Wikipedia, and other realtively open, collaborative undertakings, the
overall project can be broken down into small freestanding tasks (Benkler, 2002; Vetter,
2004). Where modularization needs to be imposed, there is more of a role for intellectual
property or organizations (entity property).
Governance can also be used by groups to overcome the drawbacks of exclusion. One
potential problem with exclusion stems from the need in a larger project to use informa-
tion subject to many rights. This could happen because a final product involves many
patents. Where this occurs, it has been termed an anticommons: just as in a commons too
many use rights lead toward overuse, in an anticommons too many exclusion rights pro-
mote underuse (Heller, 1998, 2008; Heller and Eisenberg, 1998). In addition to licensing,
a large anticommons sometimes calls forth collective-rights organizations. Patent pools
and copyright organizations are famous examples (Gervais, 2011). These organizations
often overcome exclusion and achieve a shift from property rules to liability rules (Merges,
1996a), and indeed governance strategies, keyed as they are to uses, are often paired
with liability rules. Once one is delineating a use, it often makes sense to value it rather
than use a more on/off remedy like an injunction. Rights organizations are one reason
anticommons problems have not loomed as large or persisted as long as many initially
feared (Barnett, 2015; see also Eisenberg, this volume).
Often these group institutions or sets of industry norms involve a semicommons.
Information is often more common or public for some uses and more private for other
uses. In a more diffuse context, scientists have (or at least had) a norm as described by
Merges (1996b), in which scientists would be expected to share information and materi-
als with other scientists but could demand payment from commercial actors. In a more
organized setting, firms can form joint ventures in which intellectual assets are common
for some purposes and private for others (Smith, 2000). In between are standard-setting
organizations (SSOs), which separate out one function of property law, setting standards,
and treat that as common while allowing members to control their intellectual property
subject to FRAND obligations (to license on a ‘fair, reasonable, and non-discriminatory’
basis) (Lemley, 2002; Miller, 2007; Sidak, 2013; Smith, 2013; Contreras, 2017).
Finally, we can consider the possibility of governance without exclusion. In regular
property, this is closest to being true of what Carol Rose (1986) calls ‘inherently public’
property. Here no one – neither a private party nor a group nor the government –
manages the resource. Instead, the unorganized public spontaneously, possibly through
custom, uses the resource with high benefits of multiple use and a low enough level of
conflict to make the shift to more organized management undesirable. For example,
various doctrines of navigation and historic examples like maypoles illustrate inher-
ently public property. The public domain of information is another example (Rose,
2003).

IV. CONCLUSIONS
Governance is not at odds with property; it is part of the overall institution of property.
Because of their fluid nature, informational resources tend to involve a great emphasis on
governance of uses in comparison to regular property. Exploring the problem of govern-
ance of property and the hybrid regimes we actually see has become a major avenue of
work in regular property, and it is spreading to intellectual property. This focus is welcome
because the problems of use are more difficult and more ethereal in this area.
REFERENCES
Abramowicz, Michael B. 2003. ‘Perfecting Patent Prizes,’ 56 Vanderbilt Law Review 114–236.
Alston, Lee, and Bernardo Mueller. 2014. ‘Towards a More Evolutionary Theory of Property Rights,’ 100 Iowa
Anderson, Terry L., and P.J. Hill. 1975. ‘The Evolution of Property Rights: A Study of the American West,’ 18
Journal of Law & Economics 163–79.
Arora, Sanjeev, and Boaz Barak. 2009. Computational Complexity: A Modern Approach. Cambridge:
Cambridge University Press.
Austin, John. 1873. Lectures on Jurisprudence, Robert Campbell ed., 4th ed. London: John Murray, vol. 2.
Baldwin, Carliss Y., and Kim B. Clark. 2000. Design Rules: The Power of Modularity. Cambridge, Mass.: MIT
Press, vol. 1.
Barnett, Jonathan M. 2010. ‘The Illusion of the Commons,’ 25 Berkeley Technology Law Journal 1751–816.
Barnett, Jonathan. 2015. ‘The Anti-Commons Revisited,’ 29 Harvard Journal of Law & Technology 127–203.
Barnett, Jonathan. 2017. ‘Why Is Everyone Afraid of IP Licensing?,’ 30 Harvard Journal of Law & Technology
123–45.
Benkler, Yochai. 2002. ‘Coase’s Penguin, or, Linux and the Nature of the Firm,’ 112 Yale Law Journal 369–446.
Bessen, James, and Michael J. Meurer. 2008. Patent Failure. Princeton: Princeton University Press.
Burk, Dan L., and Julie E. Cohen. 2001. ‘Fair Use Infrastructure for Rights Management Systems,’ 15 Harvard
Journal of Law & Technology 41–84.
Carrier, Michael A. 2004. ‘Cabining Intellectual Property through a Property Paradigm,’ 54 Duke Law Journal
1–145.
Chafee, Zechariah, Jr. 1928. ‘Equitable Servitudes on Chattels,’ 41 Harvard Law Review 945–1013.
Chafee, Zechariah, Jr. 1956. ‘The Music Goes Round and Round: Equitable Servitudes and Chattels,’ 69 Harvard
Chang, Yun-chien, and Henry E. Smith. 2012. ‘An Economic Analysis of Civil versus Common Law Property,’
88 Notre Dame Law Review 1–55.
Chiang, Tun-Jen. 2017. ‘The Paradox of IP,’ 30 Harvard Journal of Law & Technology 9–32.
Coase, R.H. 1960. ‘The Problem of Social Cost,’ 3 Journal of Law and Economics 1–44.
Cohen, Julie E. 2003. ‘DRM and Privacy,’ 18 Berkeley Technology Law Journal 575–618.
Collins, Kevin Emerson. 2008. ‘The Reach of Literal Claim Scope into After-Arising Technology: On Thing
Construction and the Meaning of Meaning,’ 41 Connecticut Law Review 493–559.
Contreras, Jorge. 2017. ‘From Private Ordering to Public Law: The Legal Frameworks Governing Standards-
Essential Patents,’ 30 Harvard Journal of Law & Technology 211–31.
Denicolò, Vincenzo, Damien Geradin, Anne Layne-Farrar, and A. Jorge Padilla. 2008. ‘Revisiting Injunctive
Relief: Interpreting eBay in High-Tech Industries with Non-Practicing Patent Holders,’ 4 Journal of Competition
Law & Economics 571–608.
Depoorter, Ben. 2011. ‘Fair Trespass,’ 111 Columbia Law Review 1090–135.
Duffy, John F. 2004. ‘Rethinking the Prospect Theory of Patents,’ 71 University of Chicago Law Review
439–510.
Eisenberg, Rebecca S. This volume. ‘Anticommons, Transaction Costs, and Patent Aggregators.’
Elkin-Koren, Niva. 2011. ‘Tailoring Copyright to Social Production,’ 12 Theoretical Inquiries in Law 309–47.
Ellickson, Robert C. 1991. Order Without Law: How Neighbors Settle Disputes. Cambridge, Mass.: Harvard
University Press.
Ellickson, Robert C. 1993. ‘Property in Land,’ 102 Yale Law Journal 1315–400.
Ellis, Ridsdale. 1949. Patent Claims. New York: Baker, Voorhis & Co.

Epstein, Richard A. 1997. ‘A Clear View of The Cathedral: The Dominance of Property Rules,’ 106 Yale Law
Journal 2091–120.
Epstein, Richard A. 2010. ‘The Disintegration of Intellectual Property? A Classical Liberal Response to a
Ethiraj, Sendil K., and Daniel Levinthal. 2004. ‘Modularity and Innovation in Complex Systems,’ 50
Management Science 159–73.
Feldman, Robin, and Mark A. Lemley. 2015. ‘Do Patent Licensing Demands Mean Innovation?,’ 101 Iowa Law
Review 137–89.
Freyfogle, Eric T. 1989. ‘Context and Accommodation in Modern Property Law,’ 41 Stanford Law Review
1529–56.
Friedman, David. 1998. ‘In Defense of Private Orderings: Comments on Julie Cohen’s “Copyright and the
Jurisprudence of Self-Help,”’ 13 Berkeley Technology Law Journal 1151–72.
Frischmann, Brett M. 2012. Infrastructure: The Value of Shared Resources. New York: Oxford University Press.
Frischmann, Brett M., Michael J. Madison, and Katherine J. Strandburg. 2014. Governing Knowledge Commons.
New York: Oxford University Press.
Fromer, Jeanne C. 2009. ‘Claiming Intellectual Property,’ 76 University of Chicago Law Review 719–96.
Gergen, Mark P., John M. Golden, and Henry E. Smith. 2012. ‘The Supreme Court’s Accidental Revolution?
The Test for Permanent Injunctions,’ 112 Columbia Law Review 203–49.
Gervais, Daniel. 2011. ‘The Landscape of Collective Management,’ 24 Columbia Journal of Law & the Arts
423–49.
Golden, John M. 2010. ‘Principles for Patent Remedies,’ 88 Texas Law Review 505–92.
Gordon, Wendy J. 1982. ‘Fair Use as Market Failure: A Structural and Economic Analysis of the Betamax Case
and its Predecessors,’ 82 Columbia Law Review 1600–657.
Haddock, David D., and Lynne Kiesling. 2002. ‘The Black Death and Property Rights,’ 31 Journal of Legal
Studies S545–87.
Heald, Paul J. 2005. ‘A Transaction Costs Theory of Patent Law,’ 66 Ohio State Law Journal 473–509.
Heller, Michael. 1998. ‘The Tragedy of the Anticommons: Property in the Transition from Marx to Markets,’
Heller, Michael. 2008. The Gridlock Economy. New York: Basic Books.
Heller, Michael, and Rebecca S. Eisenberg. 1998. ‘Can Patents Stifle Innovation? The Anticommons in
Heverly, Robert A. 2003. ‘The Information Semicommons,’ 18 Berkeley Technology Law Journal 1127–89.
Holbrook, Timothy R. 2006. ‘Possession in Patent Law,’ 59 SMU Law Review 123–76.
Kelly, Daniel B. 2011. ‘Strategic Spillovers,’ 111 Columbia Law Review 1641–721.
Kieff, F. Scott. 2001. ‘Property Rights and Property Rules for Commercializing Inventions,’ 85 Minnesota Law
Review 697–754.
Kitch, Edmund W. 1977. ‘The Nature and Function of the Patent System,’ 20 Journal of Law & Economics
265–90.
Langlois, Richard N. 2002. ‘Modularity in Technology and Organization,’ 49 Journal of Economic Behavior &
Organization 19–37.
Laycock, Douglas. 2011. ‘The Neglected Defense of Undue Hardship (and the Doctrinal Train Wreck in Boomer
v. Atlantic Cement),’ 4 Journal of Tort Law 1–35.
Lee, Brian Angelo, and Henry E. Smith. 2012. ‘The Nature of Coasean Property,’ 59 International Review of
Economics 145–55.
Lee, Peter. 2011. ‘The Accession Insight and Pent Infringement Remedies,’ 110 Michigan Law Review 175–241.
Lee, William F., and A. Douglas Melamed. 2016. ‘Breaking the Vicious Cycle of Patent Damages,’ 101 Cornell
Lemley, Mark A. 2002. ‘Intellectual Property Rights and Standard-Setting Organizations,’ 90 California Law
Review 1889–980.
Lemley, Mark A. 2014. ‘Taking the Regulatory Nature of IP Seriously,’ 92 Texas Law Review See Also 107–19,
accessed February 26, 2019 at http://texaslawreview.org/wp-content/uploads/2015/08/Lemley-92-SeeAlso.pdf.
Lemley, Mark A., and Carl Shapiro. 2007. ‘Patent Holdup and Royalty Stacking,’ 85 Texas Law Review 1991–2050.
Lemley, Mark A., and Carl Shapiro. 2013. ‘A Simple Approach to Setting Reasonable Royalties for Standard-
Essential Patents,’ 28 Berkeley Technology Law Journal 1135–66.
Leval, Pierre N. 1990. ‘Toward a Fair Use Standard,’ 103 Harvard Law Review 1105–36.
Libecap, Gary D. 1989. Contracting for Property Rights. Cambridge: Cambridge University Press.
Libecap. Gary D., and Dean Lueck. 2011. ‘The Demarcation of Land and the Role of Coordinating Property
Institutions,’ 119 Journal of Political Economy 426–67.
Liivak, Oskar. 2013. ‘Establishing an Island of Patent Sanity,’ 78 Brooklyn Law Review 1335–89.
Litman, Jessica D. 1987. ‘Copyright, Compromise, and Legislative History,’ 72 Cornell Law Review 857–904.

Loren, Lydia Pallas. 2007. ‘Building a Reliable Semicommons of Creative Works: Enforcement of Creative
Commons Licenses and Limited Abandonment of Copyright,’ 14 George Mason Law Review 271–328.
McManus, John C. 1972. ‘An Economic Analysis of Indian Behavior in the North American Fur Trade,’ 32
Journal of Economic History 36–53.
Madison, Michael J. 2005. ‘Law as Design: Objects, Concepts, and Digital Things,’ 56 Case Western Reserve
Menell, Peter S. 2011. ‘Governance of Intellectual Resources and Disintegration of Intellectual Property in the
Digital Age,’ 26 Berkeley Technology Law Journal 1523–59.
Menell, Peter S., and Michael J. Meurer. 2013. ‘Notice Failure and Notice Externalities,’ 5 Journal of Legal
Analysis 1–59.
Merges, Robert P. 1996a. ‘Contracting into Liability Rules: Intellectual Property Rights and Collective Rights
Organizations,’ California Law Review 1293–393.
Merges, Robert P. 1996b. ‘Property Rights Theory and the Commons: The Case of Scientific Research,’ 13
(Summer) Social Philosophy and Policy 145–67.
Merges, Robert P. 2005. ‘A Transactional View of Property Rights,’ 20 Berkeley Technology Law Journal
1477–520.
Merges, Robert P. 2011. Justifying Intellectual Property. Cambridge, Mass: Harvard University Press.
Merrill, Thomas W. 2009. ‘Accession and Original Ownership,’ 1 Journal of Legal Analysis 459–510.
Merrill, Thomas W., and Henry E. Smith. 2000. ‘Optimal Standardization in the Law of Property: The Numerus
Clausus Principle,’ 110 Yale Law Journal 1–70.
Merrill, Thomas W., and Henry E. Smith. 2010. The Oxford Introductions to U.S. Law: Property. New York:
Oxford University Press.
Merrill, Thomas W., and Henry E. Smith. 2017. Property: Principles and Policies. New York: Foundation Press,
3rd ed.
Miller, Joseph Scott. 2007. ‘Standard Setting, Patents, and Access Lock-In: RAND Licensing and the Theory
of the Firm,’ 40 Indiana Law Review 351–96.
Mossoff, Adam. 2013. ‘Introduction,’ in Adam Mossoff, ed., Intellectual Property and Property Rights.
Cheltenham, UK and Northampton, MA, USA: Edward Elgar Publishing.
Newman, Christopher M. 2011. ‘Transformation in Property and Copyright,’ 56 Villanova Law Review 251–325.
Newman, Christopher M. 2013. ‘A License is Not a “Contract Not to Sue”: Disentangling Property and
Contract in the Law of Copyright Licenses,’ 98 Iowa law Review 1101–62.
O’Rourke, Maureen A. 2000. ‘Toward a Doctrine of Fair Use in Patent Law,’ 100 Columbia Law Review 1177–250.
Ostrom, Elinor. 1990. Governing the Commons. Cambridge: Cambridge University Press.
Penner, J.E. 1997. The Idea of Property in Law. Oxford: Clarendon Press.
Penner, James. 2013. ‘On the Very Idea of Transmissible Rights,’ in James Penner and Henry E. Smith, eds.,
Philosophical Foundations of Property Law. Oxford: Oxford University Press.
Rich, Giles S. 1942. ‘The Relation between Patent Practices and the Anti-Monopoly Laws,’ 24 Journal of the
Patent Office Society 159–81.
Robinson, Glen O. 2004. ‘Personal Property Servitudes,’ 71 University of Chicago Law Review 1449–523.
Rose, Carol M. 1990. ‘Energy and Efficiency in the Realignment of Common-Law Water Rights,’ 19 Journal of
Legal Studies 261–96.
Rose, Carol M. 2003. ‘Romans, Roads, and Romantic Creators: Traditions of Public Property in the Information
Age,’ 66 (Winter/Spring) Law and Contemporary Problems 89–110.
Rose, Carol M. 2011. ‘Ostrom and the Lawyers: The Impact of Governing the Commons on the American
Legal Academy,’ 5 International Journal of the Commons 28–49, accessed on February 26, 2019 at https://
www.thecommonsjournal.org/articles/10.18352/ijc.254/.
Rothman, Jennifer E. 2013. ‘Copyright, Custom and Lessons from the Common Law,’ in Shyamkrishna
Balganesh, ed., Intellectual Property and the Common Law. Cambridge: Cambridge University Press.
Schwartz, Herbert F. 1964. ‘Injunctive Relief in Patent Infringement Suits,’ 112 University of Pennsylvania Law
Review 1025–48.
Sichelman, Ted. 2010. ‘Commercializing Patents,’ 62 Stanford Law Review 341–412.
Sichelman, Ted. 2014. ‘Purging Patent Law of “Private Law” Remedies,’ 92 Texas Law Review 516–71.
Sichelman, Ted, and Henry E. Smith. 2017. ‘Modeling Legal Modularity,’ (draft), accessed on February 26, 2019
at https://extranet.sioe.org/uploads/sioe2017/sichelman_smith.pdf.
Sidak, J. Gregory. 2013. ‘The Meaning of FRAND, Part I: Royalties,’ 9 Journal of Competition Law & Economics
931–1055.
Simon, Herbert A. 1981. The Sciences of the Artificial. Cambridge, Mass.: MIT Press, 2nd ed.
Smith, Henry E. 2000. ‘Semicommon Property Rights and Scattering in the Open Fields,’ 29 Journal of Legal
Studies 131–69.

Smith, Henry E. 2002. ‘Exclusion versus Governance: Two Strategies for Delineating Property Rights,’ 31
Journal of Legal Studies S453–87.
Smith, Henry E. 2004. ‘Exclusion and Property Rules in the Law of Nuisance,’ 90 Virginia Law Review 965–1049.
Smith, Henry E. 2007a. ‘Intellectual Property as Property: Delineating Entitlements in Information,’ 116 Yale
Smith, Henry E. 2007b. ‘Governing Water: The Semicommons of Fluid Property Rights,’ 50 Arizona Law
Review 445–78.
Smith, Henry E. 2009. ‘Institutions and Indirectness in Intellectual Property,’ 157 University of Pennsylvania
Smith, Henry E. 2011. ‘Toward an Economic Theory of Property in Information,’ in Kenneth Ayotte and Henry
E. Smith, eds., Research Handbook on the Economics of Property Law. Cheltenham, UK and Northampton,
Smith, Henry E. 2012. ‘Property as the Law of Things,’ 125 Harvard Law Review 1691–726.
Smith, Henry E. 2013. ‘Property as Platform: Coordinating Standards for Technological Innovation,’ 9 Journal
of Competition Law and Economics 1057–89.
Smith, Henry E. 2015a. ‘The Elements of Possession,’ in Yun-Chien Chang, ed., The Law and Economics of
Possession. Cambridge: Cambridge University Press.
Smith, Henry E. 2015b. ‘Equity as Second-Order Law: The Problem of Opportunism,’ (January 15, 2015), SSRN,
accessed on February 26, 2019 at http://ssrn.com/abstract=2617413 or http://dx.doi.org/10.2139/ssrn.2617413.
Smith, Henry E. 2016. ‘Semicommons in Fluid Resources,’ 20 Marquette Intellectual Property Law Review
195–212 (2015 Annual Helen Wilson Nies Lecture in Intellectual Property).
Smith, Henry E. 2017. ‘The Economics of Property Law,’ in Francesco Parisi, eds., The Oxford Handbook of Law
and Economics, Volume 2: Private and Commercial Law. Oxford: Oxford University Press, 148–77.
Smith, Henry E. 2018. ‘Complexity and the Cathedral: Making Law and Economics more Calabresian,’
European Journal of Law and Economics (2018), accessed on February 26, 2019 at https://doi.org/10.1007/
s10657-018-9591-x.
Strang, Lee J. 2008. ‘Damages as the Appropriate Remedy for “Abuse” of an Easement: Moving toward
Consistency, Efficiency, and Fairness in Property Law,’ 15 George Mason Law Review 933–74.
Symposium. 2011. ‘Property: A Bundle of Rights?,’ 8 Econ Journal Watch, accessed on February 26, 2019 at
https://econjwatch.org/issues/volume-8-issue-3-september-2011.
Taylor, David O. 2014. ‘Using Reasonable Royalties to Value Patented Technology,’ 49 Georgia Law Review
79–162.
Van Houweling, Molly S. 2008. ‘The New Servitudes,’ 96 Georgetown Law Journal 885–950.
Van Houweling, Molly S. This volume. ‘Intellectual Property as Property.’
Vetter, Greg R. 2004. ‘The Collaborative Integrity of Open-Source Software,’ 2004 Utah Law Review 563–700.
Wyman, Katrina M. 2005. ‘From Fur to Fish: Reconsidering the Evolution of Private Property,’ 80 New York
Yoo, Christopher S. 2007. ‘Copyright and Public Good Economics: A Misunderstood Relation,’ 155 University
of Pennsylvania Law Review 635–716.
Yu, Peter K. 2005. ‘Intellectual Property and the Information Ecosystem,’ 2005 Michigan State Law Review
1–20.

PART II
IP AND INCENTIVES

4. Philosophical foundations of IP law: the law and
economics paradigm
Robert P. Merges*
Contents
I. Introduction
A. Consequentialism versus Deontology: Rudiments
1. Deontology
II. Consequentialism and its Critique
A. The Macro Theory
B. Consequentialism ‘All the Way Down’: Micro-Consequentialist Roots of
Macro-Consequentialism
1. Consequentialism with nuance
C. Critiques of Consequentialism
1. The calculability critique
D. An Illustration of the Problem: New Theories, Pro and Con
E. Modest Consequentialism
III. Deontological Theories, and Their Critics
A. The Critique of Subjective Moral Intuitionism
1. A shared moral sense
B. Empirical Studies of Moral Judgment
1. Studies of children’s judgments about ownership
2. Summary: a universal ‘intellectual property instinct’?
IV. Reconciling Consequentialism and Deontological Approaches
A. The Two-Level Approach
1. But is it not inconsistent?
V. Conclusion
References
I. INTRODUCTION
This chapter describes various philosophical perspectives on the law and economic para-
digms in intellectual property (IP). It begins with a description of utilitarianism, which
is the philosophical foundation on which law and economics is built. It then describes an
alternative way that law and economics can be understood: as a highly effective set of tools
that are useful even if one is not convinced that IP law can be justified at the foundational
* Wilson Sonsini Goodrich and Rosati Professor of Law and Director of the Berkley Center
for Law and Technology, University of California, Berkeley.
72

Philosophical foundations of IP law 73
level by a utilitarian account. Various alternatives to foundational utilitarianism are

described; the chapter seeks to explain how a law and economics approach to issues in the
IP field is consistent with these alternative foundational justifications for IP law.
A. Consequentialism versus Deontology: Rudiments
Utilitarianism is one form of consequentialism, which is the label applied to philosophies

that direct us to choose from among alternative courses of action by deciding which will
produce the best consequences in the world (Smart and Williams, 1973; Darwall, 2003;
Scheffler, 1988; Kaplow and Shavell, 2006. We look at one course of action and add up all
the positive effects it would have on all the people involved, then look at all the negative
effects on all involved. Then we do the same with all other courses of action. The best
course of action is the one that produces the highest ‘net positive.’
‘Best consequences’ in this analysis can be described in a number of ways, but what they
all have in common is the assumption that we can assess the state of the world using some
sort of impersonal measuring stick. In other words, we can assess positive and negative
effects using a common denominator, some measure that allows us to take account of the
effects of the action on all persons, taken together. Put differently, this common measuring
stick applies across all people and allows us to reach an aggregated result.
Philosophers over the years have used various measuring sticks in this analysis. Utility,
or hedonic pleasure, was the original one, suggested by Jeremy Bentham (with an assist
from John Stuart Mill). A more generalized measure called ‘happiness’ is often invoked
in more recent theories; this measure abstracts somewhat from the pure physical pain and
pleasure of the Benthamite form of hedonism. The corollary measure of ‘welfare’ was
proposed in the nineteenth century, and quickly taken up by economists, who found it
more susceptible to mathematical modeling and therefore more tractable for their par-
ticular purposes (Harsanyi, 1990). This in turn gave rise to what is now known as ‘welfare
economics’ (Feldman and Serrano, 2006). Because of the dominance of this form of social
science analysis in the contemporary academy, it is safe to say that for many scholars
outside philosophy proper, consequentialism equals utilitarianism, which equals welfare
economics. This three-way approximation certainly prevails in the field of IP law. In its
most frequent usage, utilitarianism means the use of economic principles and empirical
evidence to determine what is the best course of action when confronted with a particular
policy choice in the IP field. It is of course also defined partly in opposition: utilitarianism
is the opposite of deontological theories.
Legal scholars began applying economic logic to legal doctrines in the 1960s. The
usual origin story centers on the University of Chicago (Rowley, 2005). Almost from the
beginning, there was debate over how closely ‘law and economics’ followed the tenets of
utilitarian philosophy. Some of the debates simply tracked traditional debates about the
viability of utilitarianism itself, most notably the question whether interpersonal utility
assessments can be aggregated or compared. There was also a robust early discussion
over whether law and economics was a positive methodology (merely describing, and
not evaluating, the content of legal rules) or had instead (or in addition) a normative
agenda as well. This culminated in an oft-cited exchange concerning welfare: is it a
neutral measuring stick? Is it an independent value that needs to compete with other
values such as fairness or justice, when it comes to the final calculus for deciding on legal

rules? Richard Posner, a founding figure in the movement, has expressed slightly different
answers to these questions over time (compare Posner, 1979; Posner, 1980; Posner, 1988;
Posner, 1998; Parisi, 2005). Because of this, and because consistent philosophical justifica-
tion is not that central to the law and economics discipline, there is no strong consensus
about these issues. Instead, there is general agreement on a loose set of understandings:
rationality (perfect or constrained) drives most behavior; data-driven empirical findings
(‘evidence-based reasoning’) take precedence over ‘value-laden’ analyses; and, above all,
some generalized sense of welfare maximization (perhaps leavened at times with fairness
considerations) is the proper organizing principle and operational methodology for legal
scholarship.
1. Deontology
A deontological theory is one, quite simply, based on notions of right and wrong. In
this type of theory, a moral principle determines right and wrong; violating a principle
cannot (in general) be justified by the fact that such a violation would produce good
consequences. Deontological theories speak of individuals and their rights, and they are
not in general oriented toward interpersonal aggregations or measures to figure out a
correct course of action.
These are, as the title suggests, rudimentary statements of each type of theory. The
reality is much more complex, which should not be surprising, because philosophers
have been debating these issues since at least the eighteenth century. Just to take some
examples of many that might be chosen: there are consequentialist theories that try not to
force individuals to make choices that override their own deeply held principles or values.
And there are deontologists who defend vigorously the notion that part of being a moral
person is taking account of how one’s actions affect other people – a form of interpersonal
consideration, if not outright aggregation.
But it is beyond the ken of this chapter to go too deeply into these nuances. For the
most part I will take the conventional, rudimentary views at face value as they are used in
the IP field. The goal here is simply to explain how these different foundational principles
– consequentialism versus deontology – work as justifications for IP law, and as guides to
policymaking with respect to discrete questions in IP law.
II. CONSEQUENTIALISM AND ITS CRITIQUE
The prevailing view is that US IP law has always had a utilitarian orientation.1 What has
not received as much attention is the precise relationship between utilitarian foundations,
law and economics as a scholarly orientation or methodology, and empirical evidence for
various propositions about IP law. Those are the topics I take up in this section.
1
See, for example, Fromer (2012, p. 1752):
At bottom, utilitarian theories of intellectual property rest on the premise that the benefit
to society of creators crafting valuable works offsets the costs to society of the incentives the
law offers to creators. Because this utilitarian framework establishes a cost-benefit analysis, the
leading scholarly analysis of intellectual property has used an economic lens.

In reality there are two varieties of consequentialist justification for contemporary IP

systems. They are related in a general way, but quite different once one pokes beneath the
surface of things.
The first operates at the macro level, and asserts simply that the presence of a function-
ing IP system is correlated with (i.e., very often observed in conjunction with) objective
indications of economic growth, such as rising Gross National Product (GNP). The
second provides that the best way to decide about proper IP policy is to weigh the positive
and negative effects of any proposed course of action, and then choose the action with
the greatest net positive outcome. This second theory is what can be termed micro-level
IP consequentialism. It is the unstated, but widely shared, philosophical orientation of
many (if not most) US-based IP scholars.
A. The Macro Theory
The first argument depends largely on macro-level studies showing a correlation between
patenting behavior, copyright registrations, and so on, and well-respected indicators of
economic growth and health: research and development (R&D) spending, GNP growth,
and the like. We will therefore begin with these macro-level data. As we will see, while
there is a fairly clear correlation between accepted indicators of economic growth and
the presence of an IP system, the devilish problem is to determine which way causation
runs. Put simply, it is just not clear if people file more patents because they are more
prosperous; or if prosperity flows from having filed more patents. As two highly skilled,
and neutral, observers put it: ‘[T]he availability of intellectual property for innovation
creates incentives for investment as well as potential impediments to diffusion and cumu-
lative innovation. The net effects are quite complex to sort out from both theoretical and
empirical perspectives’ (Menell and Scotchmer, 2005).
Let us put causation aside for a moment, however, and begin with correlation. Many
scholars over the years have shown that there is a close relationship between productivity
and economic growth. The basic idea is that economies grow when people learn to get
more output from a fixed supply of inputs. Economic growth, in turn, is closely related to
all sorts of positive trends such as improved health, longevity, greater personal autonomy,
and so on.
The next link in the chain centers around innovation. The key to productivity increase
is to get new technologies that multiply the output from a given stock of inputs. The
classic studies by Robert Solow showed that most of the increase in per capita GNP
over time stems from increased productivity, as opposed to increases in the stocks of
tangible resources (Solow, 1957). Technology is the key to this process. And, given that
old technologies are quickly incorporated into the production function of an economy,
new technologies are required if productivity is to grow. Thus the classic triad: economic
growth; productivity; and innovation.
Studies of innovation are based on empirical survey data. Over time, innovation
research has expanded from an original focus on new products and processes; it now
includes organizational innovations (new ways of arranging work flows, for example) and
new marketing methods (OECD, 2005).
Within this formulation, the typical proxy for investment in innovation is spend-
ing on R&D. R&D spending has been studied exhaustively, primarily because it is a

ell-accepted statistical measure, and it is represented in data gathered by many national

w
governments. The connection between IP protection and R&D spending is robust, and
has been established in a variety of economies over an extended period of time. As usual
for this sort of study, the results are stronger for patents than copyrights. But overall, IP
can be said to be correlated with standard indicators of economic growth.
If correlation were causation, this would be enough to show that IP is a beneficial
social institution. It would answer what has been termed the foundational question in IP:
whether society ought to establish an IP system at all (beyond foundations – the details
of such a system – are another matter.) The problem is that causation is a tricky issue in
this case. Perhaps stronger IP protection causes – or helps to cause – economic growth.
But perhaps it is the other way around: as wealth increases, people seek to protect their
intellectual creations more vigorously. The first description has IP causing growth. The
second has growth leading to a demand for more IP protection. Disentangling the two
has proven quite difficult. While some claim the macro evidence shows IP helps cause
growth, it is also clear that IP alone does no such thing. Very poor countries that pass
stronger IP laws have no positive net gain at all. The best that can be said, then, is that
IP protection may help cause some part of a nation’s growth, when combined with
other policies. Development economist Keith Maskus emphasizes in particular that a
positive impact for IP rights depends on institutions that otherwise support economic
competition:
The evidence supports the view that product innovation is sensitive to IPRs in developing
countries, while [foreign direct investment] and technology transfer go up when patent rights
are strengthened. Overall, there is a positive impact on growth, but this impact depends on the
competitive nature of the economy. (Maskus, 2000, p. 472)
Other scholars stress that IP is helpful only when it is strengthened at the ‘right’ moment
in a country’s development trajectory (Kim et al., 2012). One key aspect of this is that a
country must have evolved IP enforcement mechanisms, as well as having passed IP laws,
to foster actual development (Eicher and Newiak, 2013).
To summarize the macro case, we might say this. Although the data is incomplete, it is
possible to put together a plausible case that IP law has proven to be an effective part of
modern economic systems. IP law might be said then to have made a positive contribution
to economic development. At a minimum, IP law has certainly not slowed economic
development, at least not convincingly, in a way that empirics have been able to capture.
Either argument might be enough to establish that the costs of having a patent system
have not been proven to outweigh the benefits,2 thus making out a basic macro-level case
for IP.
2
It is obvious from the formulation that this way of stating the issue favors the existence of
a patent system. If the burden of proof was on someone trying to establish that the patent system
actually confers net benefits, this would favor those who oppose patents. On this general point,
David McGowan (2004) discusses the same point with respect to the copyright system, and argues
that this points toward nonconsequentialist justifications for copyright.

B. Consequentialism ‘All the Way Down’: Micro-Consequentialist Roots of

Macro-Consequentialism
The methodology of macro-consequentialism leads directly to the application of conse-

quentialist analysis to the details of an IP system. Indeed, the macro case can be seen as
simply the aggregate result of the application of all the discrete rules and details of the IP
system. This is not to say that every particular rule in an IP system is optimal, or even net
positive; just that the aggregate result of all the detailed rules exerts a net positive effect
on growth. The macro picture, in other words, in no way obviates the need to search for
detailed data that might help determine correct choices concerning specific policy issues.
This combination of macro-level faith in the IP system, together with a commitment
to a consistent consequentialist approach to all detailed issues might be described as the
working theoretical orientation of many US IP scholars. They take the macro case as
more or less given, and then proceed to look for the best available data to help answer
questions about the details of the IP system. Most scholars in effect combine macro
consequentialist foundations with an aspirational micro-consequentialist orientation to
specific policy issues.
Despite this working consensus, micro consequentialism is actually compatible with
two different views of the macro issue just described. As mentioned, one who believes
that macro data prove IP is a net positive institution simply applies the empirical method
‘all the way down’: each micro-level decision within the system is made with reference
to the best data available. Because one believes, for example, that copyright law makes
overall economic sense, one applies consequentialist reasoning to the question of whether
copyright law ought to include a ‘fair use’ privilege; and if so, how far that privilege should
extend. Or one applies economic/consequentialist analysis to the question of how long the
term of copyright protection ought to be. Indeed, a theoretically consistent consequen-
tialist would say that the macro case is simply the outcome or sum of all the individual
micro cases; the aggregate positive social welfare contribution of IP law is a product of
the fact that each micro-level policy is maximized in terms of net positive social welfare,
so that the net positive macro-level outcome is foreordained. It simply follows from the
consequentialist approach to each micro-level issue.
1. Consequentialism with nuance

There are modifications and variations on consequentialism too numerous to detail here.
Even so, a few of the more nuanced accounts bear directly on justifications for IP, so I
will provide a brief sense of several relevant variants.
One variant worth noting is the class of utilitarian theories that incorporates gener-
alization. Utilitarian generalization is a form of utilitarianism under which an action
is judged according to the effects that would follow if everyone acted similarly. The
best-known form is rule utilitarianism; according to this approach, utilitarianism, ‘the
rightness or wrongness of particular acts can (or must) be determined by reference to
a set of rules having some utilitarian defence, justification, or derivation’ (Lyons, 1965,
p. 11). (So-called ‘act utilitarianism’ looks instead at the non-generalized consequences
only of one’s own individual action.) Utilitarian generalization is relevant in that it bears
some similarity to Kantian ethics, in particular the categorical imperative. Both involve
a universalizing step: actions are determined with reference to whether they would be

desirable if pursued by all people. The difference, of course, is that for Kant the effects of
actions generally do not matter, only their intrinsic rightness or wrongness. In any event
Kantian style analysis has been applied both to a basic justification for IP rights (Merges,
2011), as well as to a number of specific problems within IP (see, e.g., Haemmerli, 1999;
Treiger-Bar-Am, 2008). And utilitarian generalization is also related to contractarian
models of government and social justice, though the contours of the fit are debated
(Mueller et al., 1974; Mueller, 2002).
Yet even though there is some family resemblance between generalized utilitarianism,
Kant’s ethics, and contractarianism, in the end even this form of consequentialism
encounters problems. The main one is that in some situations there is no single rule that
both produces good consequences and complies with a basic sense of morality. The
typical example involves cooperative interactions where multiple people must interact to
produce a benefit, but adjustments are in order regarding situation-specific contributions
by individuals. As Lyons (1965) shows, in such cases simple distributional rules (such as
taking turns accessing a shared resource) are capable of both producing greater total ben-
efits and complying with a basic sense of fairness. No single utilitarian rule does as well
(if everyone acts first, there is congestion and overuse; if everyone refrains from acting,
the benefits of access and use are lost) (Lyons, 1965, pp. 167–70). The general point is that
under a number of plausible conditions rule utilitarianism, just as with act utilitarianism,
leads to outcomes that seem unfair or unreasonable or both.
The second variant worth mentioning concerns the special problem of how the indi-
vidual treats him or herself. Utilitarianism requires that one’s personal preferences not be
weighted in any way that distorts the consequentialist calculus. This seems implausible, in
the sense that humans faced with a decision would seem to just naturally overweight the
decision’s impact on themselves. In addition, it strikes some observers as unfair, in that it
will often deprive the individual of the ability to shape his or her own life prospects; the
relentless need to account for total consequences (i.e., the impact on everyone, not just
one’s self) would seem a real burden.
The relevance to IP law is that this is a branch of law that has often been understood as
one that has special regard for the individual. Artists, writers, inventors, and all creators
are involved in acts that not only benefit others but that also (and sometimes primarily)
lead to self-fulfillment. In this context, the imperative to always include others (i.e.,
everyone else) in one’s decisional calculus seems especially onerous.
Fortunately for the individual, at least one philosopher has addressed the issue.
Samuel Scheffler (1994) provides a modified form of consequentialism that gives far
greater leeway to the particularism of individuals. With what he calls an ‘agent-centered
prerogative,’ Scheffler permits individuals in a consequentialist framework to weight the
outcomes of their own projects more heavily than they weight general consequences, that
is, outcomes with respect to everyone else. Indeed, Scheffler even provides that individuals
can maximize value as calculated from their personal point of view. There is an upper
bound, however; individuals cannot exceed some threshold level of cost imposed on
others. Individual value calculations are permitted, in other words, but not beyond some
level of ‘social’ cost. Agent-centered valuation is permitted, up to some level of socially
determined (i.e., ‘impersonally valued’) cost. In addition, Scheffler says that it is always
acceptable for people to maximize total benefits; they may, if they choose, be thoroughgo-
ing consequentialists, with no special regard to their own projects.

Here is a form of consequentialism with real promise for the IP field. Under it, one
may seek IP protection for a work that appeals to many people, and hence for which
the restricted access that follows from IP protection might produce net negative social
consequences. (Up to some threshold, anyway; at some point under Scheffler’s individual-
adjusted consequentialism the costs become too high and protection must be eschewed.)
Yet this protection will be bounded, as mentioned, so that the individual IP right cannot
harm others too much.
It should be mentioned, however, that the self-regarding values Scheffler describes have
a good deal of overlap with deontological considerations. ‘I want to protect that work
despite (some) negative effects on others; I place a very high personal value on it’ comes
awfully close to saying ‘That work has intrinsic value to me and it deserves protection;
seeking protection for it is the right thing to do for me.’ Idiosyncratic valuation bumps up
here against basic issues of fairness, or right and wrong. In that sense, Scheffler’s views
can be seen as quite consistent with some justifications for IP rights that are not normally
thought of as consequentialist in nature. Locke, for example, by virtue of his three pro-
visos, explicitly recognizes the rights and needs of others when discussing the right of an
individual to obtain property rights (Merges, 2011, ch. 2). Kant too, through his Universal
Principle of Right, seeks to reconcile strong individual property claims with the need to
account for the freedom of others (Merges, 2011, ch. 3). In this sense, Scheffler’s brand
of consequentialism comes very close to philosophical views that treat property claims
as more of an intrinsic right. If Scheffler’s notion of personal or idiosyncratic value can
be expanded to include a sense of intrinsic right and wrong; or if the ‘third party effects’
recognized by Locke and Kant can be conceptualized as taking into account negative
consequences for all others, the two families of theories might even be reconciled.
C. Critiques of Consequentialism
There are two basic objections to consequentialism. One is that it can lead to outcomes
that are morally reprehensible. The other is that it is impractical. I address them both
briefly.
In its pure form, utilitarian thinking can lead to things like this: given a certain demand
for literature in a society, it could be plausible that the best way to obtain a steady stream
of consumable literature would be to enslave a small group of high-output writers. Under
threat of death or severe punishment, put them on a strict quota of words per day or
chapters per month of novels, plays, short stories, and so on. Feed them and house them
but otherwise pay them nothing. If the literature they produce is even mediocre, many
dedicated readers might be tolerably satisfied. The cost of each book or story would be
minimal; just the cost of confining and supporting the writers, spread over all the readers
of their literature.
How could this be defended? In a large enough society, the aggregated satisfaction of
the readers might well outweigh the intense and concentrated disutility (i.e., agony) of the
writers. If net total satisfaction is all that matters, the small pleasure of the many could
dominate the deep pain of the few.
Of course, it might bother some people that their literature is generated in such a
fashion. That might reduce their utility in consuming it. But maybe not enough to make
the practice overall net negative in terms of total utility. (One also wonders whether being

‘bothered’ in this way contains at least a hint of feelings of intrinsic right and wrong –
shades of a deontological pull?) The point is not to predict what the net utility numbers
would be. It is to point out the simple fact that if the numbers came out net positive,
a consistent utilitarian would have to approve of the practice. The motto ‘live by the
numbers, die by the numbers’ might be appropriate.
In short: all sorts of things most of us find morally repulsive might be possible under
a strictly utilitarian setup. It is that sense of moral repulsion that in effect constitutes the
counterargument to pure-form utilitarianism.
1. The calculability critique

The second critique of consequentialist theories such as utilitarianism is that it is imprac-
tical. It is too difficult to determine the net consequences of almost any simple action,
which means it is hopeless to attempt to make consequentialism the comprehensive basis
for all social policy.
This objection is about the limits of calculability. There are several dimensions to this.
First, it is very difficult to predict all the consequences of a given action. This is of course
well known in the literature on causation. Even deciding on what is an ‘important’ or
‘proximate’ cause of a downstream event is fraught with complexity. Matters become
even more difficult when we take into account interaction effects – when we try to predict
how one decision may affect others, and how related actions may interact with the first
decision to produce final outcomes.
Here is an example of this dimension of calculability. Let us say we are deciding
whether the owners of internet platforms have to take responsibility for the copyright
status of material that users post on the platforms. One proposal is to generally shield the
platform owner from liability, subject to a duty to shut down the online posting activity
of specifically identified, high-volume copyright violators. The other proposal is to raise
the platform owners’ level of responsibility – to make them liable in more circumstances
for online postings that infringe copyrights, whether the person posting the material can
be identified or not.
What are the consequences of this decision? The high-shield proposal will certainly
benefit the platform owner; it need not worry much about ruinous copyright infringement
liability. But perhaps the creators of copyrighted works will suffer if this shielding cuts
into their ability to make money? Perhaps. But then again the relative openness of the
platform may provide a forum for more creators. Perhaps amateur creators will gain a
larger audience. Perhaps users of the platforms will benefit from a greater diversity of
creative works in the low-copyright-enforcement milieu? On the other hand, perhaps the
low-shield/high-liability option will lead to the creation of sophisticated filtering software
that helps to identify copyrighted works. Maybe this software will abet censorship. Maybe
it will lead to automated compensation mechanisms that help identify specific instances
where copyrighted works are used, and provide direct small payments to creators. And
perhaps this will lead to more creators entering the field in hopes of making a living. But
then again if platform companies have to pay copyright infringement claims on a regular
basis they may innovate less. They may not have the money to invest in creating their own
studios for making videos and music. This may cause artists who would have benefitted
from these investments to be worse off than they otherwise would be. Perhaps some of
these artists who would have benefitted will leave creative industries and the world will

lose out on a masterpiece or two. One or more may even die in despair, fail to have children
they otherwise would have had. Maybe one of these unborn children would have been the
scientist to cure cancer or discover cheap, safe fusion energy. And so on.
When we spin out the potential consequences in this way, what jumps out is the hope-
lessness of comprehensive consequentialism. As one philosopher put it: ‘We may not be
strictly without a clue, but we are virtually without a clue. The trouble for consequentialism
then is that the foreseeable consequences of an action are so often a drop in the ocean of
its actual consequences’ (Lenman, 2000, p. 350) (emphasis in original). Even an attempt to
mitigate the problem by assigning probabilities to various outcomes will not save the day.
This just moves the necessary complexity to a different task (assigning weights to events),
it does not eliminate it. Put simply, ‘if utilitarians are worried about the impracticality
objection, they should not turn to expected utility utilitarianism. That theory does not
provide the basis for a cogent reply to the objection’ (Feldman, 2006, p. 49).
It is easy to see the force of this critique when applied to IP law. To begin, there is
almost no area of IP law that has been studied extensively enough to warrant a ‘net grand
total’ conclusion. The many empirical studies in this field often consider only one isolated
doctrine or practice; very few are anywhere near comprehensive (Buccafusco and Heald,
2013; Budish et al., 2015; and generally Merges, 2016). There is, however, one possible
exception. It seems safe to say that removing patent protection from the pharmaceutical
and chemical industries would work a serious hardship on those industries as currently
constituted (see Graham et al., 2010). Even here though, conclusions can only be tentative,
because the studies concentrate on the industries in their current form. It is quite conceiv-
able some fundamental restructuring (e.g., government subsidies; or strict regulation
of consumer advertising expenditures relative to R&D investments) might work such
changes that patent protection is no longer essential. So even in this canonical case we
cannot conclude that we have sufficient data on which to base a permanent and compre-
hensive decision about the consequences of eliminating IP protection for these industries.
D. An Illustration of the Problem: New Theories, Pro and Con
To show why there is uncertainty about the net benefits of IP law, it may help to review two
waves of recent IP scholarship. One develops ideas about the benefits of patents that are
a modern variant of the classic ‘incentive to invent’ theory. The other examines carefully
industries where IP has largely not been available – and argues that thriving innovation
in these industries demonstrates that IP rights are not necessary in them, and possibly in
other industries as well.
Both theories are interesting and innovative. But from the perspective of determining
whether IP is a net positive or negative social institution, they are of no help. They point
in precisely opposite directions.
One new branch of literature emphasizes patent-related benefits in the form of
enhanced incentives to disclose, exchange, and license information. Briefly put, this
research emphasizes the transactional role IP rights play in economic activity. Traditional
theory simplified research activity as taking place inside a large firm; the role of patents
was to allow a firm to recoup expensive R&D costs. This newer literature emphasizes
patents as facilitators of exchange. Thus, patents are said to make small, independent
‘idea shops’ more viable, because patents make possible a specialized market in pure ideas.

Patents affect not only the overall volume of innovation, but, critically, they affect where
innovation takes place. Because small companies have some recognizable benefits, patents
may indirectly affect the overall volume and quality of innovative ideas. But they do so
indirectly: by making it easier for more innovative small firms to earn income through the
licensing of innovative ideas to other firms that engage in actual production.
In some related research, two economists have shown that the availability of published
patent applications (which came into US law in 2000) led to (1) a higher volume of patent
licensing, and (2) more rapid licensing of inventions than in earlier periods (Hegde and
Luo, 2016). These results show that patents tend to speed the transfer of ideas, and pro-
mote more rapid diffusion of innovations throughout the economy. Again, the effect of
patents on innovation is positive but in a way different in kind from the classic treatment
of patents as incentive to aggregate R&D spending.
The other branch of theory I want to mention looks at industries where IP rights are not
available, but which seem to foster significant innovation nevertheless. From French chefs
(Fauchart and von Hippel, 2008) to stand-up comics (Sprigman and Oliar, 2008) and from
fashion designers (Sprigman and Raustiala, 2006) to tattoo artists (Perzanowski, 2013),
creative people working in various industries develop norms and practices that provide
an adequate, and perhaps often superior, alternative to formal IP protection. Some have
taken to calling these areas, collectively, IP’s ‘negative spaces’ (Rosenblatt, 2011). From
these studies, a consensus theory has begun to emerge, which holds that IP law is far less
necessary than many have traditionally supposed. The absence of IP, they say, not only
fails to impede creative contributions in these areas; in some cases at least there is more
activity. From a descriptive or positive beginning, in other words, negative space theory
often moves to a more normative point: the essential wrongheadedness of the traditional
story that IP rights are always and everywhere necessary to call forth creative works.
One note of caution is in order here. Though the negative space literature looks impres-
sive, the truth is that taken together, the ‘negative space’ industries studied to date (other
than fashion) total roughly $12.75 billion in revenue per year (tattoos, French cooking,
and stand-up comedy). The fashion industry is much larger – but the problem is that
many believe the ‘negative space’ story (thriving innovation without IP rights) does not
fit the industry at all (see, e.g., Hemphill and Suk, 2009). And if we remove fashion from
the list of negative space industries, the fields studied in this research represent only a
small percentage of the economic activity that appears closely tied to IP protection. The
International Intellectual Property Protection Alliance (IIPPA), for example, estimates
that the ‘copyright industries’ alone add $1 trillion to the US economy each year.3 It is
very difficult to arrive at similar estimates for the ‘patent industries’ because companies
in so many industries obtain large numbers of patents every year. But we can say that
the pharmaceutical industry is worth nearly $300 billion per year alone (Statista, 2015)
and that the chemical industry totals roughly $800 billion (Seijo, 2017). Add in medical
devices, not to mention information technology, and it is apparent that the accounts of
negative space fields show we have a long way to go before deciding that many other
industries would benefit from the elimination of IP rights.
3
Figures such as these should always be taken with a grain of salt as they are (1) based on large
aggregate data sources, and (2) prepared by interest groups with a string agenda (Verrier, 2013).

Yet another area where new data is available is in the area of small companies and
patents. Here again a patent-positive story competes with one that is far less sanguine
(Reidenberg et al., 2015). On the positive side, Reidenberg et al. constructed a random
sample of small companies in the smartphone industry to assess how patents have affected
them. The authors found that holding a small number of patents was associated with firm
survival, financing, and other positive outcomes, and in addition that these firms were not
subject to excessive patent litigation, as some had feared in this patent-intensive industry
where many patents are held by large companies. On the other hand, Colleen Chien found
that non-practicing entities or ‘patent trolls’ disproportionately harmed small companies
with their activities (Chien, 2014). She reported that:
Most unique defendants to [patent] troll suits are small. Companies with $100M of annual
revenue or less represent at least 66% of unique defendants to troll suits and at least 55% of
unique defendants in troll suits make $10M or less per year. . ..To the extent patent demands
‘tax’ innovation, then, they appear to do so regressively, with small companies targeted more as
unique defendants, and paying more in time, money and operational impact, relative to their
size, than large firms. (Chien, 2014, pp. 461, 462)
Chien even found that in at least one case, a formerly innovative startup, OpenWave,
Inc., had shifted its own business model to become a patent troll itself (Chien, p. 479,
n. 71).
The point here is not to scrutinize the new research topics just mentioned. Instead, the
object is to simply notice the enormous complexity of the economic landscape within
which IP rights operate. And to observe that sometimes it seems in the IP literature
as though one consequentialist/empirical step in one direction (either defending IP, or
showing its inefficiency) is met immediately by a step in the opposite direction.
E. Modest Consequentialism
So it is no wonder that many scholars are agnostic about the macro level; perhaps IP
makes economic sense, in total, perhaps not. For these skeptics, the best approach in
either case is to take each particular issue and maximize net social welfare based on the
best available data about that particular issue. This view is closely compatible with the
famous admonition by the Austrian economist Fritz Machlup, to the effect that it would
be unwise based on current data to establish a patent system, based on what we know;
but that it would also be unwise to do away with the patent system as we know it, now
that it has been around for so long (Machlup, 1958). The idea is to maximize the net
social welfare stemming from any particular decision about IP policy, notwithstanding
an admission of ignorance regarding (and perhaps even the impossibility of knowing) the
ultimate systemic net social welfare determination.
When it comes to the macro question, Machlup himself falls right in between a ringing
endorsement and a harsh rebuke. Yet this tepid assessment is still – many years after it
was made – the starting point for most discussions of the field’s Big Question: are patents
justified on the economic evidence, or are they not?
The same conundrum has plagued copyright law (Breyer, 1971). And in copyright,
even the micro-consequentialist theory has been called into doubt. As David McGowan
states it:

[O]ne cannot justify one or another copyright policy solely with consequentialist arguments.
We do not even know such basic facts as how many people use the DeCSS program to make
unauthorized copies of movies, or how much people who use file ‘sharing’ software to copy
music would pay for that music if they had to. Worse yet, without some consensus on normative
principles we could not make sense of the data even if we had them. If people who would not
pay the lowest price a record company would accept for music copy the music on their own, is
this bad or good? Should policy encourage or discourage it? (McGowan, 2004, p. 2) (footnotes
omitted)
Other scholars agree (Zimmerman, 2011, p. 32; Boyle, 2008, pp. 207–8).

A slightly different flavor of skepticism was described by Posner in 1998. He wrote:
What the economist can say, which is a lot but not everything, is that if a society values prosper-
ity (or freedom or equality), these are the various policies that will conduce to that goal, and
these are the costs associated with each. The economist cannot take the final step and say that
a society’s ultimate goal should be growth, equality, happiness, survival, conquest, stasis, social
justice, or what have you. An economist discussing a ‘hot’ topic, such as whether human cloning
should be permitted, might estimate the private benefits and social costs of human cloning, and
even advise on the consequences of ignoring costs and benefits in fashioning public policy. But
he could not tell the policymaker how much weight to give costs and benefits as a matter of social
justice. (Posner, 1998, p. 1670)
Either Posner is saying that empirical data are ultimately inputs into a more value-laden
welfare calculus, or that welfare itself must be subsumed into a more comprehensive deci-
sion-making process (based, perhaps, on deontological considerations). Either way, this is
a more modest form of welfarism than strict micro-summing-to-macro-consequentialism
as described earlier.
III. DEONTOLOGICAL THEORIES, AND THEIR CRITICS
Deontological theory is based on the simple idea of moral duties. We base our decisions
on moral rules, on ideas of right and wrong, as opposed to basing them strictly on the
consequences of our actions. Decisions about what to do are tested against compliance
with rules of morality. Those rules inform whether an act is required, permitted, or
prohibited. So for example, a deontological approach to murder would say it is wrong
intrinsically. Thus even if the net consequences of a murder would be demonstrably
positive (because the murdered person is very bad, for example, or because the murdered
person is a billionaire who has pledged to leave his money to hundreds of people who are
in great need), in general a deontological theory would prohibit it.
Kantian deontological ethics are for the most part absolutist: rules of behavior do not
change with context or circumstances. Lying, stealing, murder, and so on, are simply
wrong – and that is it. Other forms of deontology are less absolute. They permit the
pursuit of net positive outcomes, a la consequentialism, but subject to constraints. So
for example, the pursuit of net positive welfare might be subject to the constraint that
people should not be harmed in the process. Even these constraints may be overridden in
some cases if enough good or bad is at stake; it may be permissible, for example, to kill a
person in order to save the lives of hundreds or thousands of other people (Kagan, 1998;
Zamir and Medina, 2008). As Rawls puts it ‘[D]eontological theories are defined . . . not

as views that characterize the rightness of institutions and acts independently from their
consequences. All ethical doctrines worth our attention take consequences into account
in judging rightness’ (Rawls, 1999, p. 26).
The obvious question, however, is this: where do these deontological duties come from;
how do we know what they are? Critics of deontological theories are quick to point out
that these are inherently subjective judgments, that they will often differ among different
people, and that they are not a trustworthy basis for making decisions, either individually
or collectively. This is both a critique of deontology per se, and, indirectly, a defense
of even a limited form of consequentialism. The implicit point is a version of the old
academic adage, ‘it takes a theory to beat a theory.’ If consequentialism – whatever its
deficiencies – is a real theory, while deontological theory is purely mush, then consequen-
tialism wins by default. This is true regardless of its faults and limitations.
Various forms of this argument can be found throughout the literature on law and
economics. In the IP field, the strongest version calls one non-consequentialist theory a
‘retreat from evidence,’ and labels it with the pejorative of ‘faith-based IP’ (Lemley, 2015,
p. 1337). It goes on to attack the theory by saying its ‘adherents are taking the validity
of the IP system on faith and . . . the rationale for doing so is a form of religious belief’
(Lemley, 2015, p. 1337).
Lemley in no way originated the claim that consequentialism has a unique and privi-
leged status as a normative theory of law. In fact, this is a very old claim, in philosophy
as well as law. As one commentator puts it:
If normative law and economics is only one possible position in normative legal philosophy,
then any suggestion that one should cede terrain in favor of the other seems out of place.
However, economic lawyers might want normative law and philosophy to be the final position in
normative legal theorizing, the ‘end of public ethics,’ so to speak. Interestingly, this tenet could
not be grounded on any number of theoretical or empirical advances made in positive law and
economics. Whatever its degree of explanatory or predictive success, positive law and economics
cannot yield normative consequences. This is just a way of rephrasing Hume’s famous motto
that ‘is’ does not imply ‘ought.’ A similar idea was defended in twentieth-century metaethics
by claiming that—without assuming the proper naturalist definition—it is fallacious to draw
moral judgments from a set of premises only containing factual statements. Normative law
and economics, like any other consequentialist moral theory, is not ‘naturally’ true. It is absurd
to attack nonconsequentialism as implausible or irrational just by noting that its results are at
odds with consequentialism. In the absence of further argumentation, it is question-begging to
dismiss nonconsequentialism by taking a stand on consequentialism. By the same token, it is
question-begging to attack fairness-based theories of law by relying on the welfare maximization
thesis of normative law and economics. (Spector, 2004, pp. 349–50)
The argument goes back much further however – to the debate over whether ‘wealth is
a value’ that took place in the early days of law and economics ascendance in US law
schools.
From time to time, lawyers and judges seeking to promote economic values such as efficiency
and wealth-maximization have taken the further step of arguing that competing paradigms
of jurisprudence based on commonsense notions of justice are theoretically inadequate bases
on which to ground a system of legal rights and obligations. [Oliver Wendell] Holmes himself
is a leading example of this tendency. [Richard] Posner’s book (Posner, 1999) is perhaps the
most ambitious attempt yet by a judge to attack the foundations of natural justice in order to
promote the law and economics movement in this way. It contains many telling criticisms of

recent professional moral philosophy, delivered with a force and directness only a secure outsider
can provide. Nonetheless, the book’s central arguments are remarkably unsuccessful. In his zeal
to criticize philosophers, Posner commits surprising mistakes that suggest he misunderstands
basic questions, not only of moral philosophy, but of jurisprudence as well. (Mikhail, 2002,
pp. 1057–8; for Holmes’s views, see Alschuler, 2000)
A. The Critique of Subjective Moral Intuitionism
The critique of deontology I have been describing boils down to the idea that notions of
right and wrong are rooted in subjective judgments. They are the product of intuition,
whim, or taste. And so they cannot form the firm foundation of any kind of rigorous,
consistent theory about how people should act.
Defenders of deontological theories have made various responses to this objection over
the years. The most thorough perhaps comes from Kant, who wrote at great length about
separating mere inclinations or intuitions from considered moral judgments (Kant, 1785).
The key for Kant is the notion that every person has not just inclinations and prefer-
ences, but also an elaborate internal construct that seeks the good. This construct can be
cultivated and consulted; it is a rational part of the self, distinct from pleasure-seeking
aspects of the self. Kant said that we can use this rational apparatus to generalize any
proposed action – to envision what the world would be like if, when we choose to act, that
choice was embodied in a universal principle applicable to all. According to Kant, our
internal reasoning mechanism can be combined with this universalizing step to produce
the famous categorical imperative. This in turn provides a rational and consistent set of
rules to govern human action (O’Neil, 1990).
Another answer to the subjectivity critique of deontological systems takes a similar
form. The philosopher John Rawls asks us to imagine making a list of moral rules, and
then testing and refining them against moral judgments about specific acts or decisions.
The back-and-forth between rules and applications produces what Rawls calls ‘reflective
equilibrium,’ which is a much more refined and rational set of judgments as compared
with initial, simple moral intuitions (Rawls, 1971, pp. 44–5). Others have extended this
idea to more explicitly consider generalizing the equilibrium across a number of people
(Scanlon, 2000). And there are of course defenders of a shared sense of morality who
argue that this sense emanates from outside people and can be recognized and developed
using spiritual practices (Kwall, 2009, pp. 139–40).
1. A shared moral sense

The elaborate procedures just described are meant to show that deontology need not be
based on a primitive form of moral intuitionism. In fact the word ‘deontology’ derives
from ‘duty’ (deon) and ‘the study or science of’ (logia). It is meant to be systematic, not
ad hoc.
Even so, without taking anything away from the systematic approaches described, a
separate defense of deontology proceeds from the fact that people make moral judgments
all the time. Despite the critics’ point that intuition is inherently subjective, whimsical, and
unreliable, proponents of an innate deontology show that there is more overlap among
people regarding moral intuitions than is sometimes supposed. Though deontology
makes great efforts to distinguish between ‘raw’ moral intuitions and a refined system of

well-considered moral rules, there is solid evidence that some moral intuitions are in fact
very widely shared.
B. Empirical Studies of Moral Judgment
Evidence has mounted in the past 25 years or so that there are very extensive overlaps in
people’s basic moral intuitions. Indeed, it has become common in some circles to speak
confidently of ‘moral universals.’ And in general, recent research suggests that quick first
impressions – based on intuition – are more reliable bases of judgment than many would
suppose (Gladwell, 2007).
Consider, for example, a well-known hypothetical situation designed to explore moral
judgments. In the basic version of this scenario, called ‘the trolley problem,’ a person
observes that an out-of-control trolley is heading for five people and will kill them unless
it is diverted (Thomson, 1985; Kamm, 1989). The only way to divert it is to pull a switch
that turns the train onto a different track – where it will kill one person. In another basic
variation, the only way to save the five people is to throw a large person, standing on a
walkway above the train track, in front of the trolley to stop it.
Students of moral theory use these (and many other) variations to explore moral
judgments. But an empirical branch of this field uses the trolley problem to test moral
judgment across cultures, age groups, and other demographic categories. In one study,
researchers asked people from many countries about two simple variations on the famous
‘trolley problem,’ and found widespread agreement about which actions were right and
wrong across many ethnicities, religions, ages, and backgrounds:
[A]cross a variety of nationalities, ethnicities, religions, ages, educational backgrounds (including

exposure to moral philosophy), and both genders, shared principles exist. That is, across every
subpopulation tested, scenario 1 (turning the train) elicited a significantly higher proportion of
permissibility judgments than scenario 2 (shoving the man), suggesting that one of . . . three
[common moral] principles . . ., or their combination, guided the moral judgments made by each
group. (Hauser et al., 2007, p. 16)4
A number of influential scholars have also documented what they call ‘human universals’
– consistent moral judgments across cultures about important subjects such as murder,
rape, and theft. Anthropologists, in particular, have revised the widespread view of
‘cultural relativism’ that prevailed in an earlier era, in part by revisiting some famous case
4
The three principles are explained by the authors:
[T]he observed pattern of judgments was consistent with at least three possible moral distinc-
tions: (1) Foreseen versus intended harm (Principle of the double effect): it is less permissible
to cause harm as an intended means to an end than as a foreseen consequence of an end; (2)
Redirection versus introduction of threat: it is less permissible to cause harm by introducing a new
threat (e.g. pushing a man) than by redirecting an existing threat (e.g. turning an out-of-control
train onto a man); and (3) Personal versus impersonal: it is less permissible to cause harm by
direct physical contact than by an indirect means. The first two distinctions have been discussed
in the philosophical literature as the content of plausible moral principles, while the third has
emerged from considerations of both behavioral and neurophysiological evidence (Hauser et al.,
2007, p. 15).

studies purporting to show aberrant practices, and revising those findings in light of more
detailed research (see, e.g., Pinker, 2003; Brown, 1991). But some philosophers are less
than convinced (Alfano and Loeb, 2014).
Part of the misunderstanding about moral intuition has to do with the nature of intui-
tion itself. Many casual observers equate ‘intuition’ with a ‘retreat from evidence,’ meaning
a substitute for empirical observation or even scientific fact (Lemley, 2015, p. 1337).
But this view seriously misunderstands the nature of moral reasoning. One prominent
explanation of moral judgments is that they are based not on ‘irrational’ intuitions, but
instead on a ‘universal moral grammar’ that is inherent to all human beings. This concept
is based on an analogy to Noam Chomsky’s famous finding that all humans are born with
a common human template for grammar and language (Mikhail, 2013). The key point
is how researchers learned of this template. It was not itself invented speculatively, but
instead was induced from extensive empirical observations. In the world of language, the
universal grammar notion came to prominence because researchers constantly observed
that native speakers without sophisticated language training ‘just knew’ proper word
usage and sentence structure, even with regard to words and sentence-types they had never
encountered before (Cook and Newson, 1996; Jackendoff, 1994; Pinker, 1994). This led
linguists to look for further evidence of an ingrained, ‘hard-wired’ capacity for language
and language structure – evidence which has mounted ever since.
Moral philosophers familiar with empirical studies of moral judgment across cultures
posited a similar type of template, only in the moral sphere. Observations, such as those
centered on the trolley problem, confirm that people from all cultures and with all educa-
tion backgrounds share similar judgments concerning right and wrong actions in several
basic human situations (see also Pinker, 2008).
These studies are not, unfortunately, widely known in law and economics circles – or
perhaps among consequentialist thinkers generally. Leading law and economics pioneer
and theorist Richard Posner, for example, has stated that ‘there are no general moral
principles, just particular moral intuitions’ (Posner, 1999, p. 11). Moral philosopher John
Mikhail critiques the Posner view this way:
Posner takes for granted the standard ‘particularist’ assumption that moral judgments are made
on a case-by-case basis, without the support of moral principles. This assumption is incompat-
ible with the best explanation of the properties of moral judgment. To explain how the normal
individual is able to make stable and systematic moral judgments about an indefinite number of
novel cases, we must assume she is guided, implicitly, by a system of principles or rules. Without
this assumption, her ability to make these judgments—and our ability to predict them—would
be inexplicable. Posner’s insistence that the capacity for moral judgment consists of nothing more
than ‘theoretically ungrounded and ungroundable preferences and aversions’ is therefore suspect
from the start. (Mikhail, 2002, p. 1092)
To paraphrase: even if I could not explain an ‘objective’ set of moral principles to skeptics
of deontology, that does not mean they do not exist. Mikhail once again explains:
There is no justification for Posner to insist that the normal individual must be fully aware of
the operative principles which constitute her moral knowledge, or that she can become aware
of them through introspection, or that her statements about them are necessarily accurate.
On the contrary, he should recognize that just as normal people are typically unaware of
the principles guiding their linguistic or visual intuitions, so too might they be unaware of the
principles guiding their moral intuitions. In any event, the important point is that, as with the

theory of language or of vision, the theory of moral cognition attempts to describe the operative
principles of moral competence, not what experimental subjects may or may not report about
them. (Mikhail, 2002, p. 1094)
1. Studies of children’s judgments about ownership

A different strain of empirical research has some important things to say about moral
judgments concerning IP. Psychologists who study children’s responses to moral issues are
interested in a number of issues, including the way moral sensibility develops as children
grow. But one subfield in the area of children’s moral judgment is of particular relevance
to issues of right and wrong as they pertain to IP. A group of studies with children
subjects centers on questions of ownership. A brief review of these studies will be helpful
in exploring further the issue of shared moral intuitions.
One set of studies deals with judgments about ownership. In a typical experiment,
pre-school age children are exposed to a person handling an object. The object is then
left, and picked up by another person. A robust finding is that children infer ownership
from first possession:
[T]he first possession heuristic guides children’s ownership inferences. The findings provide the
first evidence that preschoolers can infer who owns what, when not explicitly told, and when
not reasoning about objects with which they are personally acquainted. (Friedman and Neary,
2008, p. 829)
Research shows that the first possession heuristic extends to ideas as well as physical
objects. As one article title in this field says, ‘Children Apply Principles of Physical
Ownership to Ideas’ (Shaw et al., 2012). At the same time, children are also able to
distinguish ownership from possession. In studies where an object moves among a group
of people, children track the person that others look to for permission to take possession.
This person they infer is the owner. In this and other studies, it has been demonstrated
that even very young children make sophisticated judgments about issues of ownership.
The authors of one study put it this way:
[C]hildren (6–8 years old) determine ownership of both objects and ideas based on who first
establishes possession of the object or idea. Study 2 shows that children use another principle
of object ownership, control of permission—an ability to restrict others’ access to the entity in
question—to determine idea ownership. In Study 3, we replicate these findings with different
idea types. In Study 4, we determine that children will not apply ownership to every entity,
demonstrating that they do not apply ownership to a common word. Taken together, these
results suggest that, like adults, children as young as 6 years old apply rules from ownership not
only to objects but to ideas as well. (Shaw et al., 2012)
Other studies, concentrated on creative labor, are also highly pertinent to moral judgments
about IP. For example, when presented with a conflict between someone who abandons an
object and someone who finds it, adults are more likely to endorse the first owner when he
or she has invested creative labor in the object (Beggan and Brown, 1993). In one study,
children were shown an object, such as modeling clay, owned by person A; then observed
as person B expended labor in making something creative (like a small figure) with the
clay. In this and similar studies, ‘creative labor increases the likelihood that 3-year-old and
4-year-old children will endorse the creator and not the original owner of materials as the

owner of a final product’ (Kanngiesser et al., 2010, p. 1238). And the instinct to recognize
creative effort extends both to children’s own creations and those of others:
[W]e found that children applied the same rules to their own property and to the property
of the person they were directly interacting with. Although this behavior may have arisen as
a by-product of the cooperative social setting of the experiment, our results do indicate that
young children are able to overcome a previously established bias to maximize their own gain
. . . (Kanngiesser et al., 2010, p. 1240)
Beyond the general assimilation of ideas to objects when it comes to notions of ownership,
children also have a distinct reaction to plagiarism. In one article, the authors show that
adults, older children (9–11 years old), and children from age 5 and up, all respond with
negative moral judgments about plagiarism (Olson and Shaw, 2011). Original creativity is
more highly valued, and one who copies and claims credit is assessed negatively even by
these very young children. The study authors found that only 3- and 4-year-old children
fail to distinguish between original creators and plagiarists. As the authors conclude, ‘by
age 5 years old, children understand that others have ideas and dislike the copying of these
ideas’ (Olson and Shaw, 2011, p. 431).
2. Summary: a universal ‘intellectual property instinct’?

What these studies show is that there are strong regularities in people’s thinking about
ownership, fairness, and the importance of creative labor. And because the studies are
cross-cultural, involve children, or both, they support the idea that moral judgments
about these issues may be less due to socialization in a particular culture and more due to
a basic shared moral sense.
What these studies tend to prove is that moral judgments about ownership and creative
labor are highly idiosyncratic and unstable. Critics of deontological theories, who argue
against it as being ‘unscientific,’ should take note. As one commentator put it, in travers-
ing the arguments of the critic Richard Posner:
Posner fails to come to terms in any serious way with the hypothesis that human beings
share a sense of justice rich enough to support a universal system of rights and obligations,
including the right not to be murdered. This hypothesis is plausible and supported by a
considerable body of empirical evidence. Throughout [his book] Problematics, Posner adopts
the mantle of science and pokes fun at philosophers for being unscientific. But in truth, it is
his relativism, not their universalism, that seems out of touch with modern science. (Mikhail,
2002, p. 1062)
The same may well be said of those who criticize deontological theories of IP law. It is
they who are out of touch with modern science. Hence, even if one rejects the systematic
deontological systems of Kant or Rawls, one must still confront the fact that, empirically
speaking, when it comes to issues of property rights many people revert to a common,
innate template for making moral judgments. Empirically speaking, when it comes to
moral issues, people are not primarily empiricists. They rely on a set of moral judgments
so common and pervasive they are close to being universal.

IV. RECONCILING CONSEQUENTIALISM AND

DEONTOLOGICAL APPROACHES
For many IP scholars the choice between deontology and consequentialism is difficult;
each seems to capture something important about the field. Their instinct is to fight the
binary choice, the compromise, to reconcile, to integrate.
There are three ways to do that. The first two rely primarily on one methodology, using
the other as a constraint. The other divides the methodological problem into two sharply
delineated halves or levels.
The threshold approach works like this: you begin by saying ‘I am consequentialist – but
not all the way down the line. In the right circumstance, I might yield to a moral judgment,
sacrificing some positive consequences because not to do so feels wrong.’ Or you say, ‘I
am deontological in my approach – but not all the way down the line. In the right circum-
stances I might permit something that feels wrong in general, if there are overwhelming
positive consequences for doing so’ (Kagan, 1998; Zamir and Medina, 2008).
The other approach is to say, ‘I am not sure if there is an airtight consequentialist case
for the existence of IP rights; but if I think IP is a morally just institution, something
that would be included in any fair society, then I think the rules of IP should in general
be designed to maximize positive consequences.’ This divides the problem of justifying
IP into two distinct levels: the foundational level and a more policy-oriented, more
operational level.
A. The Two-Level Approach
The two-level approach is appealing because it recognizes the computational limits that
plague the consequentialist argument for the existence of IP, yet preserves plenty of room
for efficiency considerations. But is this too neat, or does the two-level approach suffer
from a sort of fatal incompatibility between the two levels?
To see how the two-level approach might hold together, consider the everyday problem
of choosing whether to rent a home or buy one. Many people will say there is a solid,
rational case for buying over renting. Given reasonable assumptions about interest rates,
tax rates, the future trajectory for apartment rents, and real estate appreciation, buying
just makes sense. There is, in other words, a solid foundation for this decision based on
hard-headed, empirically-verified ‘dollars and cents’ considerations.
One who is a bit skeptical, or who does not easily accept certain assumptions, might
not be so sure. Maybe rental costs will be lower than projected; maybe the home mortgage
deduction will be changed; maybe when one factors in the risk of uninsurable or expensive
catastrophes, the case for ownership is not so solid after all.
Within this latter group, even if one were to decide that the financial/economic case
for home ownership is inconclusive based on available data, one might still decide to
buy. This decision would simply be based on criteria other than strict dollars and cents
calculations. Maybe ownership would bring a sense of accomplishment, or send a signal
of commitment, that renting would not do. Whatever the alternate criteria, they would
constitute the basic justification for the decision to buy.
However, once the buy decision was made, cost minimization would provide a very
important operational principle that should exert significant influence on the way the

decision is implemented – that is, on exactly which house to rent or buy. The key distinc-
tion is that dollars and cents considerations – efficiency – is to be seen here as a tool in the
service of some other, ultimate criterion. Efficiency is an operational principle meant to
best implement a decision made for other (non-efficiency) reasons.
If you are an IP scholar who is skeptical of the thorough consequentialist case for the
existence of IP, then for you law and economics (or the efficiency principle) serves
the same role in IP law as efficiency does in the case of a home buyer/renter who rejects
the adequacy of the purely economic case for home ownership. It is a highly useful
operational principle that can be deployed to help shape the precise contours of the IP
system. It is not so powerful as to justify the system as a whole, but it is powerful enough
to exert a strong influence on the way the system should be organized and administered.
1. But is it not inconsistent?

There are several references above to ‘consequentialism all the way down.’ It is also
possible of course to be a deontologist ‘all the way down.’ Neither may seem ideal; but
aren’t they both preferable to being nothing all the way down? Put differently, how can
one defend switching approaches ‘mid-stream,’ as suggested by the two-level solution just
described?
This is of course a reasonable view. It has the obvious appeal of consistency. Yet it also
makes sense to employ different theoretical tools at the two levels of analysis. To see why,
consider those two levels again.
The foundational question is whether the existence of any IP system can be justified
or defended. Should a fair and well-organized society have IP protection? The book
Justifying Intellectual Property (‘JIP’) (Merges, 2011) sets out a detailed discussion of this
question, together with some attempts at an answer. JIP makes two basic points about the
foundation level. First, reasonable people disagree. Some believe a good consequentialist
defense of IP can be made. Others believe in foundational consequentialism in general,
but are skeptical or agnostic about whether the consequentialist case has been made
adequately. Others identify strong deontological roots for the IP field, and justify IP
systems on this basis. Still others explain IP as consistent with conventional religious
morality.
JIP makes one observation on this topic. Although this level of disagreement may
seem troubling, it is not a very big issue in the day-to-day debates in the field. This is
because the debates take place on a different plane from that of foundational justifica-
tion. Policy arguments employ standard themes that are distinct from the rigorous
requirements needed to ground foundational justifications for the field. These themes
are general principles. They are common policy tenets that are embodied in many diverse
doctrines and rules. A good example is the principle of proportionality, which dictates
that an IP right must be roughly commensurate in scope with the merit or value of the
IP-protected work covered by the right. This principle can be found throughout patent
law (nonobviousness, enablement, injunctions), copyright law (substantial similarity,
thin copyright, fair use), and even trademark law (strength of marks, remedies). This
principle is a general theme of IP law, and it does not depend for its force on any particu-
lar foundational justification. It is compatible with consequentialist and deontological
theories alike. And it both transcends individual doctrines and shapes and guides the
application of those doctrines.

Other general or ‘midlevel’ principles (so-called because they mediate between founda-
tional justifications and specific doctrines) are efficiency, nonremoval (or solicitude for
the public domain), and dignity (Merges, 2011). This is a nonexhaustive list; there are
undoubtedly other primary themes in IP law that span specific doctrines and rules.
One way to think about general or midlevel principles is that they are like a shared
public space within which people holding divergent foundational belief can come together
to discuss policy and reach consensus. They have been analogized to John Rawls’ idea
of an ‘overlapping consensus’ among people holding different beliefs regarding religion
and other fundamental life issues. Just as Rawls observed that the ‘public sphere’ can
bring together people with divergent basic beliefs, so too JIP theorizes that the general
or midlevel principles constitute a shared space that makes possible policy debate and
consensus among people with differing opinions about the basic justifications for IP
law (Merges, 2011; Rawls, 1971, p. 340). Another analogy is to Cass Sunstein’s idea of
‘incompletely theorized agreements,’ which in a similar way promotes consensus-building
(Sunstein, 1995). For Sunstein, the solution comes by avoiding generalities and seeking
consensus on specific, particular cases. But the result is the same: a procedure, vocabulary,
or norm that permits consensus-building by avoiding the need to resolve issues at the most
fundamental (and therefore most contentious) level. So the observation in JIP is that
general policy principles allow us to avoid contentious arguments over the foundations
of IP rights.
The argument in JIP is something different. That book argues that we cannot justify
IP rights based on current knowledge of their costs and benefits. The data is inadequate
to the task. But, given the decision to have IP rights, important features of our current
IP regime can be explained by recourse to the principle of efficiency. Using the best data
and analytic techniques at our disposal, we as a society try to maximize the net benefits
of IP rights. Put simply, the argument is that we do not have IP rights because we are sure
they are efficient, but given that we have them – because they are justified by Lockean,
Kantian, or other deontological normative precepts – we strive to make our IP system as
efficient as possible. So, efficiency explains and ties together important aspects of the IP
system as practiced, even though we cannot be sure of the utilitarian case for IP. To state
this simply, one might say that efficiency is a good quality or feature for an IP system, but
not an acceptable rationale for it (Merges, 2011, 2013).
This has important implications for the way IP research is conducted. If one is a
utilitarian or consequentialist, striving to rationalize each aspect of IP on the grounds of
total social welfare or utility, a concern with efficiency is of course essential. If one does
not agree that the consequentialist case has been made adequately to date, one can still
take a great interest in efficiency. It is an important tool in the operation of an effective IP
system. Even if the foundations of that system lie outside the realm of net consequences,
efficiency considerations may still be highly relevant.
V. CONCLUSION
Some scholars see little need for a deep discussion of whether IP law is necessary. It is
obviously here, part of our world; why not just take it as it is, try to make the best of it
given that it is here, and save the mental anguish of trying to justify it.

For others, it only seems reasonable to sound the depths of the field. In this camp there
are two primary options. The inquiry can center on whether the IP system as a whole
leaves the world better or worse off. This measures the desirability of IP by looking at the
consequences of having these rights. The good effects are weighed out with the bad ones;
a net figure or assessment is made; and that is that. There are pluses and minuses to this.
The other main avenue of inquiry looks to notions of right and wrong, to some set of
moral or deontological rules. Different facets or qualities of the IP system are considered,
and a general judgment is reached: IP either is, or is not, an institution that a good society
would foster and support.
Each of these two formulations has its attractions and problems. Toting up conse-
quences gets very complicated, and it may prove difficult to reach a net determination.
Deontology faces the criticism that it is too subjective, too unstable to support a viable
justification.
Each of these criticisms has answers. Calculability may be a problem, but at least
consequentialism strives for precision and reproducibility; it has the aura of science and
empirical fact on its side. Deontological judgments are more reliable than they appear,
because they can be made systematically, and because even to the extent they are based
on intuition, there is more consensus about moral intuitions than most critics believe.
One way to bring these opposing justifications into greater harmony is to cordon off
the foundational question – whether society should have an IP system – from the level of
basic principles that are to operate if such a system is established. For a thoroughgoing
consequentialist, this entails simply a two-step process; the net merits of each aspect of the
IP system are plumbed, and then, in a separate step, a grand (net) total is arrived at. For
one who relies on deontological foundations, such a procedure works especially well. He
or she can arrive at an answer to the foundational question; if it is positive, then efficiency
is one important tool that can be brought to bear on the running of the IP system in
practice. Though consequentialist theory might seem lacking as a way to justify the IP
system, consequentialist tools can be deployed once an IP system is established.
This two-level solution to the problem of justifying IP has one attractive feature. It
permits people to disagree about foundational justifications, while cooperating fully in the
development of IP policies inside an actual IP system. It has the potential for us to hold to
our deepest intellectual commitments, while debating policy on a more operational plane.
So we can disagree, without being disagreeable.
REFERENCES
Alfano, Mark, and Don Loeb. 2014. ‘Experimental Moral Psychology,’ Stanford Encyclopedia of Philosophy,
accessed March 31, 2019 at https://plato.stanford.edu/entries/experimental-moral/.
Alschuler, Albert W. 2000. Law without Values: The Life, Work, and Legacy of Justice Holmes. Chicago, IL:
University of Chicago Press.
Beggan, James K., and Ellen M. Brown. 1993. ‘Association as a Psychological Justification of Ownership,’ 128
The Journal of Psychology 365–80.
Boyle, James. 2008. The Public Domain: Enclosing the Commons of the Mind. New Haven, CT: Yale University
Press, 205–207.
Brown, Donald. 1991. Human Universals. New York, NY: McGraw-Hill Education.

Buccafusco, Christopher, and Paul J. Heald. 2013. ‘Do Bad Things Happen When Works Enter the Public
Domain? Empirical Tests of Copyright Term Extension,’ 28 Berkeley Technology Law Journal 1–43.
Budish, Eric, Benjamin N. Roin, and Heidi Williams. 2015. ‘Do Firms Underinvest in Long-Term Research?
Evidence from Cancer Clinical Trials,’ 105 American Economic Review 2044–85.
Chien, Colleen. 2014. ‘Startups and Patent Trolls,’ 17 Stanford Technology Law Review 461–506.
Cook, Vivian J., and Mark Newson. 1996. Chomsky’s Universal Grammar: An Introduction. Maiden, MA:
Blackwell Publishing.
Darwall, Stephen. 2003. Deontology. Hoboken, NJ: John Wiley & Sons.
Eicher, Theo S., and Monique Newiak. 2013. ‘Intellectual Property Rights as Development Determinants,’ 46
Canadian Journal of Economics 4–22.
Fauchart, Emmanuelle, and Eric von Hippel. 2008. ‘Norms-Based Intellectual Property Systems: The Case of
French Chefs,’ 19 Organization Science 187–201.
Feldman, Alan M., and Roberto Serrano. 2006. Welfare Economics and Social Choice Theory, 2d ed. New York,
NY: Spring Science + Business Media, Inc.
Feldman, Fred. 2006. ‘Actual Utility, the Objection from Impracticality, and the Move to Expected Utility,’ 129
Philosophical Studies 49–62.
Friedman, Ori, and Karen R. Neary. 2008. ‘Do Children Infer Ownership from First Possession,’ 107 Cognition
829–49.
Fromer, Jeanne C. 2012. ‘Expressive Incentives in Intellectual Property,’ 98 Virginia Law Review 1745–824.
Gladwell, Malcolm. 2007. Blink: The Power of Thinking Without Thinking. New York, NY: Back Bay Books.
Graham, Stuart J.H., Robert P. Merges, Pamela Samuelson, and Ted Sichelman. 2010. ‘High Technology
Entrepreneurs and the Patent System: Results of the 2008 Berkeley Patent Survey,’ 24 Berkeley Technology
Haemmerli, Alice. 1999. ‘Whose Who? The Case for a Kantian Right of Publicity,’ 49 Duke Law Journal
383–491.
Harsanyi, John C. 1990. ‘Interpersonal Utility Comparisons,’ in John Eatwell, Murray Milgate, and Peter
Newman, eds., Utility and Probability. The New Palgrave. London: Palgrave Macmillan.
Hauser, Marc, Fiery Cushman, Liane Young, R. Kang-Xing Jin, and John Mikhail. 2007. ‘A Dissociation
between Moral Judgments and Justifications,’ 22 Mind & Language 1–21.
Hegde, Deepak, and Hong Luo. 2016. ‘Imperfect Information, Patent Publication, and the Market for Ideas,’
14-019 Harvard Business School working Paper (Feb. 8, 2016), accessed April 1, 2019 at http://www.hbs.edu/
faculty/Publication%20Files/14-019_726e14eb-8ebb-49d9-ae75-f368916cdf60.pdf.
Hemphill, C. Scott, and Jeannie Suk. 2009. ‘The Law, Culture, and Economics of Fashion,’ 61 Stanford Law
Review 1147–99.
Jackendoff, Ray. 1994. Patterns in the Mind: Language and Human Nature. New York, NY: BasicBooks.
Kagan, Shelly. 1998. Normative Ethics. Boulder, CO: Westview Press.
Kamm, Francis M. 1989. ‘Harming Some to Save Others,’ 57 Philosophical Studies 227–60.
Kanngiesser, Patricia, Nathalia Gjersoe, and Bruce M. Hood. 2010. ‘The Effect of Creative Labor on Property-
Ownership Transfer by Preschool Children and Adults,’ 21 Psychological Science 1236–41.
Kant, Immanuel. 1785. Groundwork for the Metaphysics of Morals. New Haven, CT: Yale University Press.
Kaplow, Louis, and Steven Shavell. 2006. Fairness Versus Welfare. Cambridge, MA: Harvard University Press.
Kim, Yee Kyoung, Keun Lee, Walter G. Park, and Kineung Choo. 2012. ‘Appropriate Intellectual Property
Protection and Economic Growth in Countries at Different Levels of Development,’ 41 Research Policy
358–75.
Kwall, Roberta. 2009. The Soul of Creativity: Forging a Moral Rights Law for the U.S. Stanford, CA: Stanford
University Press.
Lemley, Mark A. 2015. ‘Faith-Based Intellectual Property,’ 62 UCLA Law Review 1328–46.
Lenman, James. 2000. ‘Consequentialism and Cluelessness,’ 29 Philosophy & Public Affairs 342–70.
Lyons, David. 1965. Forms and Limits of Utilitarianism. Oxford: Oxford University Press.
Machlup, Fritz. 1958. ‘An Economic Review of the Patent System, Study of the Subcommittee on Patent,
Trademarks and Copyrights,’ Committee on the Judiciary, U.S. Senate, 85th Congress, 2nd Session.
Maskus, Keith E. 2000. ‘Intellectual Property Rights and Economic Development,’ 32 Case Western Reserve
Journal of International Law 471–505.
McGowan, David. 2004. ‘Copyright Nonconsequentialism,’ 69 Missouri Law Review 1–72.
Menell, Peter S., and Suzanne Scotchmer. 2005. ‘Intellectual Property,’ in A. Mitchell Polinsky and Steven
Shavell, eds., 2 Handbook of Law and Economics. Oxford: North-Holland.
Merges, Robert P. 2013. ‘Foundations and Principles Redux: A Reply to Professor Blankfein-Tabachnick,’ 101
Merges, Robert P. 2016. ‘Economics of Intellectual Property Law,’ in Francesco Parisi, ed., The Oxford
Handbook of Law and Economics. Oxford: Oxford University Press.

Mikhail, John. 2002. ‘Law, Science, and Morality: A Review of Richard Posner’s the Problematics of Moral and
Legal Theory,’ 54 Stanford Law Review 1057–127.
Mikhail, John. 2013. Elements of Moral Cognition: Rawls’ Linguistic Analogy and the Cognitive Science of Moral
and Legal Judgment. Cambridge: Cambridge University Press.
Mueller, Christian. 2002. ‘The Methodology of Contractarianism in Economics,’ 113 Public Choice 465–83.
Mueller, Dennis C., Robert D. Tollison, and Robert D. Willett. 1974. ‘The Utilitarian Contract: A Generalization
of Rawls’ Theory of Justice,’ 4 Theory and Decision 345–67.
OECD (United Nations Organisation for Economic Co-operation and Development). 2005. Oslo Manual:
Guidelines for Collecting and Interpreting Innovation Data. Paris, France: OECD Publishing.
O’Neil, Onora. 1990. Constructions of Reason: Explorations of Kant’s Practical Philosophy. Cambridge:
Olson, Kristina R., and Alex Shaw. 2011. ‘“No Fair, Copycat!”: What Children’s Response to Plagiarism Tells
Us about Their Understanding of Ideas,’ 14 Developmental Science 431–9.
Parisi, Francesco. 2005. ‘Methodological Debates in Law and Economics: The Changing Contours of a
Discipline,’ in Francesco Parisi and Charles K. Rowley, eds., The Origins of Law and Economics. Cheltenham:
Edward Elgar Publishing.
Perzanowski, Aaron. 2013. ‘Tattoos and IP Norms,’ 98 Minnesota Law Review 511–91.
Pinker, Steven. 1994. The Language Instinct: How the Mind Creates Language. New York, NY: Harper Perennial
Modern Classics.
Pinker, Steven. 2003. The Blank Slate: The Modern Denial of Human Nature. New York, NY: Penguin Books.
Pinker, Steven. 2008. ‘The Moral Instinct,’ The New York Times Magazine (Jan. 13, 2008), accessed April 1, 2019
at http://www.nytimes.com/2008/01/13/magazine/13Psychology-t.html.
Posner, Richard A. 1979. ‘Utilitarianism, Economics, and Legal Theory,’ 8 Journal of Legal Studies 103–40.
Posner, Richard A. 1980. ‘The Ethical and Political Basis of the Efficiency Norm in Common Law Adjudication,’
8 Hofstra Law Review 487–507.
Posner, Richard A. 1988. ‘The Ethics of Wealth Maximization,’ 36 University of Kansas Law Review 261–5.
Posner, Richard. 1998. ‘The Problematics of Moral and Legal Theory,’ 111 Harvard Law Review 1637–717.
Posner, Richard A. 1999. The Problematics of Moral and Legal Theory. Cambridge, MA: Harvard University
Press.
Rawls, John. 1971. A Theory of Justice. Cambridge, MA: Harvard University Press.
Rawls, John. 1999. A Theory of Justice. Cambridge, MA: Harvard University Press.
Reidenberg, Joel R., N. Cameron Russell, Maxim Price, and Anand Mohand. 2015. ‘Patents and Small
Participants in the Smartphone Industry,’ 18 Stanford Technology Law Review 375–429.
Rosenblatt, Elizabeth L. 2011. ‘A Theory of IP’s Negative Space,’ 34 Columbia Journal of Law & the Arts 317–65.
Rowley, Charles K. 2005. ‘An Intellectual History of Law and Economics: 1739–2003,’ in Franceso Parisi and
Charles K. Rowley, eds., The Origins of Law and Economics. Cheltenham, UK and Northampton, MA, USA:
Scanlon, Thomas. 2000. What We Owe to Each Other. Cambridge, MA: Belknap Press.
Scheffler, Samuel. 1988. Consequentialism and Its Critics (Oxford Readings in Philosophy). Oxford: Oxford
University Press.
Scheffler, Samuel. 1994. The Rejection of Consequentialism: A Philosophical Investigation of the Considerations
Underlying Rival Moral Conceptions. Oxford: Clarendon Press.
Shaw, Alex, Vivian Li, and Kristina R. Olson. 2012. ‘Children Apply Principles of Physical Ownership to Ideas,’
36 Cognitive Science 1383–403.
Seijo, Bibiana Campos. 2017. ‘The business of chemistry,’ 95 Chemical & Engineering News (No. 19) 2 (May 8,
2017), accessed April 1, 2019 at https://cen.acs.org/articles/95/i19/business-chemistry.html.
Smart, J.J.C., and Bernard Williams. 1973. Utilitarianism: For and Against. Cambridge: Cambridge University
Press.
Solow, Robert M. 1957. ‘Technical Change and the Aggregate Production Function,’ 39 Review of Economics
and Statistics 312–20.
Spector, Horacio. 2004. ‘Foreword: Law and Economics and Legal Scholarship,’ 79 Chicago-Kent Law Review
345–54.
Sprigman, Christopher, and Dotan Oliar. 2008. ‘There’s No Free Laugh (Anymore): The Emergence of
Intellectual Property Norms and the Transformation of Stand-Up Comedy,’ 94 Virginia Law Review
1787–867.
Sprigman, Christopher, and Cal Raustiala. 2006. ‘The Piracy Paradox: Innovation and Intellectual Property in
Fashion Design,’ 92 Virginia Law Review 1687–777.
Statista. 2015. ‘Domestic and International Revenue of the U.S. Pharmaceutical Industry between 1975 and
2015 (in Million U.S. Dollars),’ accessed April 1, 2019 at https://www.statista.com/statistics/275560/domestic
-and-international-revenue-of-the-us-pharmaceutical-industry/.
Sunstein, Cass. 1995. ‘Commentary: Incompletely Theorized Agreements,’ 108 Harvard Law Review 1733–72.

Thomson, Judith J. 1985. ‘The Trolley Problem,’ 94 Yale Law Journal 1395–415.
Treiger-Bar-Am, Kim. 2008. ‘Kant on Copyright: Rights of Transformative Authorship’, 25 Cardozo Arts &
Entertainment Law Journal 1059–103.
Verrier, Richard. 2013. ‘U.S. Copyright Industries Add $1 Trillion to GDP,’ Los Angeles Times (Nov. 19,
2013), accessed April 1, 2019 at http://articles.latimes.com/2013/nov/19/business/la-fi-ct-intellectual-prop
erty-20131119.
Zamir, Eyal, and Barak Medina. 2008. ‘Law, Morality, and Economics: Integrating Moral Constraints with
Economic Analysis of Law,’ 96 California Law Review 323–91.
Zimmerman, Diane L. 2011. ‘Copyrights as Incentives: Did We Just Imagine That?,’ 12 Theoretical Inquiries
in Law 29–58.

5. Intellectual property law and the promotion of
welfare
Christopher Buccafusco* and Jonathan S. Masur**
5 6
Contents
I. Introduction
II. IP Law and Consequentialism
III. Versions of IP Consequentialism—What Should Congress Promote?
A. Promoting Science and Useful Arts
B. Promoting Social Welfare
1. Preferences/wealth
2. Objective approaches
3. Happiness
IV. Implications for IP of Varying Approaches to Welfare
A. Consumer-Side Welfare Effects
1. Patent law
2. Copyright
B. Producer-Side IP Effects
1. Creativity as benefit
2. Employment effects
V. Conclusion
References
I. INTRODUCTION
The US Constitution grants Congress the power ‘to Promote the Progress of Science
and the Useful Arts’ by granting copyrights and patents to authors and inventors (U.S.
Const. art. I, § 8, cl. 8). This language is understood by most courts and scholars to entail
a utilitarian or consequentialist approach to intellectual property (IP) law. The possibility
of obtaining copyrights and patents encourages creators to produce new innovations that
ultimately redound to the public good. By providing incentives to creators that are balanced
with the interests of the public and with subsequent creators, IP laws can optimize creative
** Professor of Law and Director of the Intellectual Property and Information Law Program,
Cardozo School of Law, Yeshiva University.
** John P. Wilson Professor of Law and David and Celia Hilliard Research Scholar at the
University of Chicago Law School. Masur thanks the Wachtell, Lipton, Rosen & Katz Program
in Behavioral Law, Finance and Economics and the David and Celia Hilliard Fund for financial
support.
98

Intellectual property law and the promotion of welfare 99
and innovative production. Unlike IP systems in other parts of the world, US IP law gener-
ally eschews so-called ‘moral’ or deontological considerations such as justice and fairness.
Although there is considerable consensus regarding IP law’s philosophical orientation,
there has been little discussion of its deeper normative goals. Most courts and scholars
agree with the idea that IP law should provide incentives to creators, but there has been
almost no analysis of why creativity and innovation are good. This is simply taken as
given. But what, exactly, are the interests that IP law should promote? How should we
understand what constitutes ‘the Progress of Science and the Useful Arts?’ Various
answers to these questions exist. One possibility would be to interpret the constitutional
language literally and narrowly. On this view, IP laws should encourage developments in
knowledge and technology irrespective of broader interests. Another option would be
to interpret the constitutional language broadly to encompass a general social welfare
calculus: IP laws should be subjected to some form of cost-benefit analysis to determine
whether they ultimately make people better off.
In this chapter, we discuss a variety of ways of understanding the normative goals of a
consequentialist IP regime. We argue that the best approach derives from recent work in
the field of hedonic psychology. The principal consequentialist goal of IP laws should be
to maximize social welfare, where welfare is understood as subjective well-being. We do
not argue that IP laws cannot have other interests beyond consequentialism; there may
be room for concerns about fairness and justice, as well. But where IP law is motivated by
welfare considerations, it should be evaluated in light of how laws and policies will affect
people’s happiness.
We begin by reviewing the commitments of US courts and scholars to a consequen-
tialist orientation for IP law. Although there is a broad consensus that the law should
promote good outcomes, there has been less discussion of the kinds of outcomes that the
law should be promoting. In Section III, we address different ways in which IP law could
promote good outcomes. For example, IP law could be narrowly focused on promoting
creativity and innovation or it could be more broadly directed toward general social
welfare. We argue that the latter, broader focus is more appropriate. We then consider
three different accounts of human welfare and how the law can promote it: preferentist,
objectivist, and hedonic. We support the hedonic account of human welfare and describe
its strengths over the other options. Then, in Section IV, we discuss what a hedonically
oriented IP policy would look like.
II. IP LAW AND CONSEQUENTIALISM
The constitutional text provides the foundation for IP law in the US, but, like most provi-
sions of that document, it leaves considerable room for debate and interpretation. Most
important for our purposes is the language ‘To Promote the Progress of Science and the
Useful Arts.’ Each of the principal terms has been discussed by scholars at considerable
length (Solum, 2002; O’Connor, 2015; Oliar, 2006). For example, scholars have ques-
tioned whether this text imposes a meaningful limitation on Congress’s power or whether,
instead, it is simply preambular and non-limiting. They have also attempted to define
‘Science’ and ‘useful Arts’ and to understand their relations to copyright and patent law.
One way or another, however, most courts and scholars have agreed that IP law in the

US should be governed by some version of consequentialism. That is, IP laws exist to

promote certain goals rather than, for example, to vindicate natural rights. This stands
in stark contrast to the philosophical foundations of most European IP systems, where
the moral rights of authors and inventors provide justification for the legal system. In the
US, IP rights are merely the means to a desired end. As the Supreme Court has explained:
The monopoly privileges that Congress may authorize are neither unlimited nor primarily
designed to provide a special private benefit. Rather, the limited grant is a means by which an
important public purpose may be achieved. It is intended to motivate the creative activity of
authors and inventors by the provision of a special reward, and to allow the public access to the
products of their genius after the limited period of exclusive control has expired. . ..The copy-
right law, like the patent statute, makes reward to the owner a secondary consideration. (Sony
Corp. v. Universal Studios, 464 U.S. 417 (1984) (internal citations and quotations omitted))
In the patent context, the Court has offered similar reasoning:
The patent laws promote this progress by offering a right of exclusion for a limited period
as an incentive to inventors to risk the often enormous costs in terms of time, research, and
development. The productive effort thereby fostered will have a positive effect on society through
the introduction of new products and processes of manufacture into the economy, and the
emanations by way of increased employment and better lives for our citizens. In return for the
right of exclusion—this ‘reward for inventions,’—the patent laws impose upon the inventor a
requirement of disclosure. (Kewanee Oil Co. v. Bicron Corp., 416 U.S. 470 (1974))
On this view, the ‘rights’ that authors and inventors receive do not devolve from the
spiritual connection between creators and their works, nor are they representative of the
Lockean intellectual labor that creators have mixed with the common. Instead, copyrights
and patents are an administrative solution to an economic problem (Aronson v. Quick
Point Pencil Co., 440 U.S. 257 (1979)).
The difference between these approaches is evident in a variety of legal doctrines. In US
copyright law, for example, authors generally do not receive the right to attribution for
their works or the right to prevent their destruction or defacement. These are important
components of European systems that protect droit d’auteur, or moral rights. In the US,
they are thought to excessively impinge upon speech interests, sequential creation, and
creativity markets (Adler, 2009). In patent law, inventors must be named in patent filings,
but US law imposes no further requirements that inventors be named in products commer-
cializing inventions. To the extent that US IP law does incorporate non-consequentialist
concerns, such as in the Visual Artists Rights Act (17 U.S.C. § 106A (2018)) or in the
ability of authors to terminate transfers of their copyrights, these injections of justice and
fairness concerns are the exceptions that prove the consequentialist rule.
Academics in the US have generally followed the Supreme Court, and they have fleshed
out the economic theory behind the rights (Landes and Posner, 2003; Cass and Hylton,
2013). Both in copyright and patent law, rights are understood to flow from the state’s
desire to promote certain valuable outcomes by providing incentives to individuals and
corporations (Lemley, 2005). Because creating new works and inventions is costly, and
because the works and inventions can be cheaply copied, creators would not have suf-
ficient incentives to invest in innovative activity. Copyrights and patents provide periods
of exclusivity during which rights holders can charge supracompetitive prices to offset

the costs of their creative investments. Providing rights is costly, however, because they
create deadweight losses and can inhibit further innovation. Accordingly, IP regimes seek
to balance the rights given to creators with those available to subsequent creators and to
the public.
Although scholars often differ sharply with respect to how the law should maintain this
balance, they generally concur in the belief that this is IP law’s overriding goal. And while
important and forceful arguments in favor of natural rights or deontic approaches to IP
law have been made (Hughes, 1988; Merges, 2011; Kwall, 2009), these approaches have
yet to find consistent support.
III. VERSIONS OF IP CONSEQUENTIALISM—WHAT SHOULD

CONGRESS PROMOTE?
We join the belief that IP is best understood in consequentialist terms rather than
deontic ones. The IP Clause seems to speak in consequentialist terms, and most courts
and commentators to have considered the question have described the purposes behind
IP in consequentialist terms as well. But the notion that IP is consequentialist raises
as many questions as it answers. What, precisely, is IP meant to promote? What is the
maximand, or across what factors should IP be maximizing? There is a wide variety of
different options, and those options have significantly divergent ramifications for the
shape of intellectual property law. In this section, we consider the various possibilities
and highlight which of them have gained the greatest currency within the judicial and
scholarly communities.
A. Promoting Science and Useful Arts
First, as the Constitution suggests, IP could be meant to promote science and the useful
arts, full stop. In other words, the Constitution could be dictating that IP be structured
to maximize the production of useful innovations (patents) and creative works (copy-
right), with those innovations and works viewed as ends in and of themselves. This type
of approach might be justified on a number of different theoretical grounds. First, it
may be that the Constitution does not mean to instantiate one vision of the good, be
it welfarist (in any of its forms) or non-welfarist. Accordingly, to avoid taking sides,
the Constitution—and the law that springs from it—should perhaps be interpreted to
maximize some other worthwhile quantities without striving for any greater theoretical
ambitions. Innovation and creativity might be those quantities. Alternatively, one might
imagine that it is too difficult for any social planner—court, Congress, or otherwise—to
fashion rules that are effective at increasing welfare. Every rule will create too many
unforeseen effects—too many ripples in the pond. Any attempt to manage welfare
directly is doomed to failure through unintended consequences. Accordingly, it might be
that social planners are advised to concentrate on more easily identifiable or obtainable
quantities, such as innovative and creative works.
We believe that these are defensible but ultimately unpersuasive arguments. Innovation
and creativity are not ends in and of themselves under any view of the good. While we
do not wish to minimize the difficulty involved in promoting welfare-enhancing law and

policy, this is a task that policymakers regularly undertake. Indeed, the entire enterprise of
cost-benefit analysis, which now pervades the regulatory state, is geared toward enhancing
welfare. Moreover, if the goal were simply to maximize the production of innovation
and creative works, patent and copyright law would likely assume very different shapes.
Generally speaking, the longer the intellectual property term, the greater the incentive to
innovate because the greater the rewards. We say ‘generally speaking’ because in some
cases extensive IP protections might retard follow-on innovation, as when blocking
patents or patent thickets increase the costs to subsequent inventors trying to follow in the
footsteps of an initial inventor. In addition, there is reason to doubt the incentive effects of
IP beyond a certain time horizon (Balganesh, 2009). Nonetheless, if the goal were only to
maximize innovation and creativity, the patent term would likely be substantially longer,
at least for certain types of inventions such as pharmaceuticals. (The copyright term is
already extremely long and ever-increasing.) Accordingly, we are not surprised that this
view of the IP Clause has acquired few adherents within either the courts or the scholarly
literature.
B. Promoting Social Welfare
If IP does not exist to promote science and the useful arts alone, then it must be geared
toward promoting some view of social welfare. However, that very basic conclusion
immediately raises a more vexing question: what is the proper measure of welfare? What,
precisely, should IP law be maximizing?
1. Preferences/wealth
The most widely adopted view of welfare is that it consists of the satisfaction of indi-
vidual preferences—typically fully informed, rational preferences, but occasionally simple
expressed preferences (Sumner, 1996). This view prevails among the vast majority of
economists. It relies on a straightforward logic: if a person obtains what she desires, that
must make her better off—otherwise, why would she desire it? Many economists would
impose the additional limitation that the preference must be self-interested in order to
cope with the issue of individuals with preferences for helping others at their own expense
(Adler and Posner, 2006). Economists believe that an individual’s preferences are best
measured by using market transactions: if an individual chooses to purchase a good for
some amount of money, the transaction reveals that the individual values the good more
than she values the money used to obtain it (Boyle, 2003). Accordingly, this theory of
welfare comes with a mechanism for obtaining some understanding of how much welfare
a society possesses, and whether that welfare is increasing or decreasing.
This conception of welfare as preferences is the dominant view within intellectual
property as well (Demsetz, 1969; Menell and Scotchmer, 2007; Boyle, 2007; Kapczynski
and Syed, 2013; Tur-Sinai, 2016). Essentially all economists who study IP subscribe to it,
and many legal scholars do as well. Indeed, the entire structure of IP in the United States
seems to favor a preferentist account of social welfare. The market is the mechanism
through which IP is meant to incentivize innovation. By rewarding inventors and creators
with monopoly rights over their innovations and creative works, IP allows them to reap
supracompetitive profits. But the owners of IP can only realize these profits if individu-
als are actually willing to purchase their products and services. If there are no market

participants with preferences for the patented or copyrighted goods, their creators earn
nothing and IP is irrelevant. The value of IP is linked to the market value of the underlying
good, and thus to individual preferences (Beebe, 2017).
Contrast this system of intellectual property with alternative mechanisms for encourag-
ing creativity and innovation, such as government-sponsored grants or prizes (Hemel and
Ouellette, 2013). Mechanisms of promoting innovation that do not rely on the market
similarly do not rely on preference satisfaction as the measure of whether an innovation
should be produced or will be funded. The policymakers choosing which projects to
fund could rely upon any other type of criteria, including deontic criteria. While there
is a substantial amount of public grant-based funding for science and the arts, the
vast majority of funding incentives for innovation and creativity are provided through
intellectual property. Accordingly, while it is not the case that every type of innovation
policy rests upon preferentist foundations, intellectual property—at least as currently
conceived—generally does.
Nevertheless, preferentist approaches to welfare are subject to trenchant critiques. Some
opponents, typically from a philosophical tradition, argue that there is more to the good
life—more to welfare—than realizing the things that an individual wants. These scholars
locate human welfare in a more objective notion of human flourishing: either possessing
a set of objective capabilities and entitlements (Sen, 1993) or doing well the things that
humans should do (Foot, 2001). We discuss flourishing and objective approaches at
greater length in the following section.
A second type of objection, arising primarily from the disciplines of psychology and
behavioral economics, observes that individuals often desire things that turn out not to
make their lives better off by any observable measure (Bronsteen, Buccafusco, and Masur,
2014). A typical example involves a family that trades an apartment in the city, close to
the parents’ offices, for a large house and a yard in the suburbs. In many cases it turns out
that the larger house offers few hedonic benefits to the family. They do not appear happier
living in it than they were in their smaller apartment, and they quickly adapt to the size
of the house so that it soon feels as though it is not especially large. However, the move
to the suburbs is tremendously costly in psychological terms. The parents must now drive
in traffic an hour each direction, and driving in traffic makes them miserable. Not only
is the experience itself very unpleasant, it also takes them away from their children—the
entire purpose of moving to the suburbs in the first place.
While this example is hypothetical, the data behind it are very real and well documented
(Kahneman et al., 2004; Stutzer and Frey, 2008). And this is only one example of what
is variously termed an ‘affective forecasting error’ or an instance of ‘mis-wanting.’ The
upshot of this significant body of research is that individual preferences cannot be relied
upon to increase well-being according to any other measure. Unless one accepts prefer-
ences as a tautological definition of well-being, there is ample evidence that preference
satisfaction does not necessarily lead to the improvement of an individual’s life.
For the most part, these critiques have not penetrated intellectual property. Most
scholars persist in equating welfare with preference satisfaction for purposes of measuring
the effectiveness of intellectual property. And as noted above, the intellectual property
system continues to be organized predominantly around market mechanisms that rely
upon preference satisfaction. But that should not obscure the fact that cracks are showing.

2. Objective approaches

The leading alternative to preference-based conceptions of welfare is a set of loosely
related philosophical approaches that are commonly grouped under the heading of
‘objective accounts.’ Some of these approaches, following Sen (2005) and Nussbaum
(2001), describe welfare as the possession of a set of goods enumerated in a list.
These ‘objective list’ theories of welfare do not always perfectly correspond, but they
typically involve a related and overlapping list of objective goods as their measure
of welfare (Lewinsohn-Zamir, 2003). Other objectivist theories borrow more heavily
from the Aristotelian concept of flourishing, typically understood as ‘doing well that
which it is characteristically human to do’ (Foot, 2001). These ‘flourishing’ or ‘virtue
ethics’ conceptions of welfare tie individual well-being to participation in a set of
positive or ‘virtuous’ activities. Here, too, there is substantial overlap with objective
list approaches. While virtue ethics does not typically specify in advance the particular
activities that lead to flourishing and a good life, the goods and activities that form
the basis of objective list theories are commonly viewed as constitutive of flourishing
by Aristotelians.
What these two flavors of ethical theory have in common is that they locate welfare
in a set of objective considerations outside of the subjective preferences or desires of
any given individual. For the objectivist theorist, the goods on the objective list, or the
activities constitutive of flourishing, are valuable regardless of whether (or how much)
the individual desires them. The individual’s welfare is determined by the existence or
non-existence of those quantities and activities, without reference to the individual’s own
views. This approach has sown the seeds of criticism by preferentists (and hedonists, as we
will describe below), who point out the incoherence of relating individual welfare—which
would seem to be the most subjective and individual of quantities—to external objective
considerations (Adler and Posner, 2006). Other scholars have observed that the selection
of the particular items found on objective lists will typically reflect the preferences of the
philosophers who put them there (Sumner, 1996). In addition, most objective theories
do not offer mechanisms for weighing the various welfare goods against one another or
for making inter- or intra-personal welfare comparisons. Regardless, objectivist theories
are the principal competitor to preferentist theories of welfare, and they have gained
significant currency within the philosophical and legal communities.
Objectivist theories have recently begun to make their way into intellectual property as
well. Some scholars have begun with objective approaches to welfare and argued that they
should be imported directly into IP, equally with any other area of law (Frischmann, 2014;
Derclaye, 2013; Derclaye, 2012; Fisher and Syed, 2007). Others have argued in reverse,
claiming that IP cannot be justified or defended on preferentist or hedonic grounds
(Tur-Sinai, 2016). These scholars argue instead that IP is best conceived as furthering and
representing an objectivist conception of welfare and should be reformulated to better
accomplish those ends.
The connection between IP and objective theories of welfare is not difficult to under-
stand. Creativity and innovation, the foundations of copyright and patent law, are canoni-
cally ‘virtuous’ activities that are commonly thought of as partially constitutive of a good
life according to essentially every objective conception of welfare. If one imagines IP
from the producer side—IP should exist to encourage and enable individuals to produce
creative and innovative works—then it becomes natural to envision IP in objectivist terms.

In addition, numerous scholars have pointed to failures and shortcomings in the market
for innovation and the ways in which ability to pay does not always map onto welfare
(Kapczynski, 2012). For instance, pharmaceutical drugs that produce substantial benefits
for poorer individuals—such as citizens of developing nations—might increase welfare
dramatically by improving the lives of many people. But the market for such drugs might
be quite meager because of the limited purchasing power of the individuals who could
consume them. By contrast, pharmaceuticals that cater to wealthy individuals might be
highly valuable in the marketplace even if they improve few lives (and improve those lives
only marginally). In light of these pathologies, the natural step for many scholars is to
argue that IP should be reformulated to advance a set of ends chosen objectively, without
reference to the particular preferences of the affected individuals as expressed through
their wealth.
The problem with these objective welfare theories of intellectual property is that they
are typically tied very closely to the outcomes of individual lives (Tur-Sinai, 2016). To
proponents of objective conceptions of welfare and IP, what makes a particular innova-
tion worth pursuing is that it improves the lives of some group of individuals—witness
the emphasis on medicines, particularly medicines for the poor. Advocates for objectivist
conceptions of IP have not offered independent justifications for these normative goals
beyond the fact that they promise to improve the lives of the individuals concerned. In
other words, these objective considerations are actually founded in traditional subjective
conceptions of welfare—preferences or hedonics (Griffin, 1986). Proponents of objective
conceptions of IP welfare have succeeded in demonstrating that the market, as it currently
exists, is not the ideal mechanism for generating welfare-enhancing outcomes. They have
not succeeded in demonstrating that objective conceptions of welfare are superior to
hedonics-based formulations. Indeed, the fact that arguments regarding objective concep-
tions of welfare for IP inevitably resort to reliance upon subjective improvement of life is
suggestive of the objective view’s inherent weakness.
3. Happiness
In recent years, a third conception of welfare as based in subjective hedonic experience has
returned to prominence within philosophy and legal academia (Bronsteen, Buccafusco,
and Masur, 2014). This conception of welfare as subjective well-being (or ‘happiness’)
dates back to Jeremy Bentham but fell largely out of favor over the course of nearly two
centuries. It has lately been resurrected in part due to work by psychologists, who have
innovated new mechanisms for measuring individual happiness (Kahneman et al., 1999).
Their work has been followed by normative arguments from philosophers and legal
scholars who press the point that happiness is the best way to understand what it means
for an individual to live a good life.
As these philosophers and legal scholars have pointed out, the subjective well-being
conception of welfare holds several substantial advantages over preferentist and objectiv-
ist conceptions of welfare (Bronsteen, Buccafusco, and Masur, 2014). First, it accounts
for the fact that individuals often desire goods that do not appear to improve their lives
and that they cannot predict the types of things that will improve their well-being. Most
notably, while individuals typically strive for ever-greater levels of wealth and material
possessions, these goods do not appear to impact individuals’ subjective well-being sig-
nificantly (Bronsteen, Buccafusco, and Masur, 2013). This is different from a claim, which

objectivists might make, that there is nothing worthwhile or valuable in material wealth
and possessions for theoretical normative reasons. Rather, it is an empirical observation
that wealthier people do not lead lives that are substantially happier or better even by their
own lights than people of more modest means.
Second, unlike objectivist approaches the subjective well-being approach treats welfare
as intimately related to an individual’s own well-being and desires, not an outsider’s. That
is, whether a good or experience is welfare-enhancing for a given individual depends upon
the feeling it produces in that individual and how that individual relates to it, not upon
whether a third party judges the good or experience as valuable. Third, it possesses a
close, intuitive tie to standard notions of welfare: if an experience makes a person happy
or produces good feelings, it makes intuitive sense to say that it has increased her welfare.
Fourth, it is just as easy to make inter- or intra-personal hedonic comparisons as it is to
make such comparisons from a preferentist standpoint, and generally easier than such
comparisons are within any type of objectivist account of welfare. Fifth and finally, the
hedonic account of welfare does not rely upon any theoretically difficult constructs, such as
‘fully informed preferences,’ which are impossible to determine in practice. A policymaker
or legal decision maker who wishes to assess whether some legal rule or project will increase
subjective well-being need not engage in complicated and normatively fraught laundering
of individual preferences (Adler and Posner, 2006). She need only measure the happiness of
an affected population before and after a policy intervention to know whether the interven-
tion produced a positive effect on well-being (Bronsteen, Buccafusco, and Masur, 2014).
A hedonic approach to IP would in many respects resemble the objectivist approach
described above. One of the fundamental precepts of a hedonic conception of welfare is
that individuals often (mistakenly) desire things that do not make them happy and do not
improve their own welfare, and they frequently fail to desire things that will make them
happy and improve their welfare. This is true even when people’s desires are motivated by
the goal of increasing their own happiness (Adler, Dolan, and Kavestos, 2015). Thus, mar-
kets cannot be relied upon to produce welfare-maximizing outcomes when participants
are not making fully informed, rational, and self-interested decisions about which goods
to consume. Accordingly, a happiness-based IP policy could be used to help determine the
types of goods that are most likely to increase well-being, rather than simply relying upon
market participants to select the goods they prefer. Policymakers could then structure IP
law to help facilitate the production of those goods.
The crucial difference between a hedonic approach to IP and an objectivist approach
lies in how one determines the types of goods whose production IP law should be designed
to promote. An objective approach would rely upon the intuition and reasoning of the
policymaker, informed by philosophical theory. A hedonic approach would instead rely
on empirical data: policymakers would study which goods and activities improve overall
cumulative individual subjective well-being over time and then promote or advance the
development of those goods and activities. The hedonic approach would be tied to lived
individual experiences; the objectivist approach would be divorced from the experiences
or feelings of the people who will be affected by IP. Given that welfare is intuitively
understood as a subjective concept, tethered to the individual’s own life, we view this
distinction as a decisive advantage for the hedonic conception.
Partly because the study of law’s effects on subjective well-being is in its infancy
(Bronsteen, Buccafusco, and Masur, 2014), relatively few scholars have adopted a hedonic

approach to IP. There is, however, a growing cohort who have at least explored the possi-
bility that IP law might be reoriented in a hedonic direction (Manta, 2013; Derclaye, 2013;
Rai, 2007). As we noted above, IP law in its current instantiation adopts a fundamentally
preference-based approach to welfare. In the following section, we will sketch the outlines
of how a hedonic-centered view of IP law might be constructed.
IV. IMPLICATIONS FOR IP OF VARYING APPROACHES TO

WELFARE
We have already alluded to the various ways in which different conceptions of welfare will
affect IP policy differentially, and in this part of the chapter we sketch such competing
conceptions more directly. Due to space limitations we focus on patent and copyright law,
the two major areas of IP, although similar analyses of trademark, trade secret, design
law, and others would undoubtedly be illuminating.
IP law has typically focused on creating consumer welfare by giving producers eco-
nomic incentives to do things they would not do without those incentives. The law thus
affects welfare in two distinct ways: on the consumer side and on the producer side. This
part differentiates between the two ways in which IP law affects welfare, turning first to
consumer welfare and then to producer welfare.
A. Consumer-Side Welfare Effects
1. Patent law

One of the central teachings of the voluminous literature on hedonic psychology is
that individuals are often mistaken as to what will improve their lives. People often
believe that certain goods or experiences will increase their happiness, only to find their
happiness unchanged or even diminished. The implication of this research is that the
sorts of choices that individuals make in the marketplace—the goods or experiences
they purchase—may not be a perfectly reliable indicator of what will improve these
individuals’ welfare. On the contrary, there might be particular types of goods or services
that will increase individual well-being much more than others, and much more than
market-based decisions would indicate. This suggests that intellectual property policy
should self-consciously favor innovation in welfare-enhancing technological sectors,
even at the expense of other areas of technology that do not have the same effect on
individual welfare.
Which areas of technology are most likely to lead to gains in well-being? The lead-
ing candidates are technologies that improve health and prolong life. There is ample
evidence that health is one of the primary drivers of well-being (Powdthavee and van
den Berg, 2011). An individual’s health status is highly correlated with her well-being,
and because most people’s lives are happy even through old age, extending their lives
will typically provide significant boosts in their lifetime welfare. Improvements in phar-
maceutical and medical device technology should thus lead to improvements in human
welfare, perhaps more so than advances in any technological sector. A policymaker
who has adopted a hedonic view of welfare would thus favor legal rules that optimize
innovation in these technological fields. Because pharmaceutical drugs and medical

devices have large up-front research costs and low copying costs, scholars tend to believe
that progress in those fields is best served by longer patent terms and stronger IP rights
(Burk and Lemley, 2009). There is also evidence to indicate that strong patent protec-
tion for pharmaceuticals does not significantly inhibit follow-on innovation (Sampat
and Williams, 2015). And because most Americans are covered by health insurance,
patented drugs and medical devices remain widely available despite monopoly prices.
There is thus a plausible case for strong patent protection for medical technologies as a
matter of hedonic welfare.
The problem with attempting to favor health-related technologies is that patent law is
generally technology-neutral. The same patent law that governs pharmaceutical drugs
and medical devices also governs semiconductors, consumer electronics, and every other
type of technology. Burk and Lemley (2009) have observed that in some cases judges can
tailor patent law to fit the particular needs of a particular industry or technological field,
but this is a marginal effect. To first approximation, patent law functions similarly across
technologies. And while strong patents might increase innovation in pharmaceuticals and
medical devices, the same might not be true in computers, software, or electronics. These
fields of technology suffer more from problems of overlapping patents—thickets and
anticommons problems—than medical fields (Burk and Lemley, 2009). Excessively strong
and numerous patents might eventually begin to retard innovation. If courts or legislators
were to make patents more powerful, this would likely inhibit progress in computers and
consumer electronics at the same time that it promoted innovation in pharmaceuticals
and medical devices.
Under typical assumptions, scholars and policymakers would face significant uncer-
tainty as to the proper course of action in the face of such competing considerations.
There is no a priori reason to favor either electronics or pharmaceuticals over the other.
The market heavily rewards innovation in both sectors, as evidenced by the volume and
value of sales. On a preferentist account of welfare, a policymaker would be forced to
conduct a complicated calculation to determine which area of technology yields the
greatest returns, and thus whether increasing or decreasing patent strength would increase
overall welfare. This type of cost-benefit analysis is extraordinarily difficult if not entirely
impossible in the patent context (Masur, 2016).
Yet a hedonic approach to welfare may offer an answer to this difficult problem. In con-
trast to health-related fields, some technologies—including computers and e lectronics—
likely have relatively minor impact (if any) on overall welfare. A great deal of innovation
in the computer, software, and consumer electronics sectors is directed at expensive
consumer products (smartphones, tablets, plasma televisions, and so on) available only to
individuals above a certain level of income. At the same time, there is mounting evidence
that increases in wealth have only extremely minor impacts on happiness (Kahneman
and Deaton, 2010; Oswald and Powdthavee, 2008). Even individuals who dramatically
increase their income realize only marginal gains in well-being (Bronsteen, Buccafusco,
and Masur, 2013). This implies that these consumer electronic devices are contributing
very little to well-being, particularly given how much they are being used. The average
American watches several hours of television per day, and smartphone users average more
than an hour of daily use (Nielsen, 2014). If high-tech televisions and smartphones were
contributing positively to our well-being, it should be reflected in higher hedonic levels for
wealthier people. Yet the data reveal no such trend. In fact, at least one study has linked

smartphone use to depression, although the finding is correlative and not necessarily
causal (Saeb et al., 2015).
Accordingly, from a hedonic perspective it could be beneficial to pursue stronger patent
protection in the interest of increasing innovation in pharmaceuticals and medical devices,
even if this resulted in harm to other industries such as computers and electronics. If all
technologies are not created equal for purposes of improving welfare, then policymakers
need not refrain from taking steps that maximize innovation in some at the expense of
others. Macro-level hedonic understandings of welfare provide a means for answering
difficult policy questions without a full-blown cost-benefit analysis. A patent policy
formulated on hedonic terms might thus be structured very differently than the status
quo, and this could be accomplished while holding patents facially neutral with respect
to technology.
In addition to drugs and medical devices, there are other types of technology that
likely have a meaningful impact on human welfare, measured hedonically. In particu-
lar, any technology that reduces environmental pollution—and thus morbidity and
mortality from harmful exposure—should significantly improve welfare (Bronsteen,
Buccafusco, and Masur, 2013). This is especially true in light of the catastrophic welfare
effects predicted from global climate change (Masur and Posner, 2011). Likewise for
technologies that improve automobile safety or reduce other types of everyday risks.
And there are even more commonly overlooked types of technology with the potential
for meaningful welfare impacts. For instance, hedonic studies have demonstrated
that the manner, time, and distance of individuals’ commutes to and from work can
dramatically affect individual welfare (Gilbert, 2006). Driving in traffic is one of the
least pleasurable activities in which an individual can engage (Kahneman et al., 2004).
If an individual moves an hour further away from her workplace, she substitutes two
hours of leisure per day—during which she might engage in a pleasant activity—for
two hours of unhappy driving in traffic. The diminution in overall well-being could be
very substantial. Thus, a hedonic approach to welfare would support whatever intellec-
tual property rules would maximize innovation in technologies to improve or shorten
commutes, including driverless cars, more intelligent traffic management, railroad
technology, telecommuting equipment, and the like. These industries have not been the
subject of sufficient study to allow us to determine whether stronger or weaker patents
would be more productive of innovation and development. Accordingly, it is possible
that the same patent rules that are helpful for pharmaceuticals might be harmful for
environmental technologies and other welfare-enhancing improvements. Regardless,
the broader point is that a hedonic conception of welfare offers a set of prescriptions
for patent law that differ substantially from the status quo. Once policymakers come
to understand welfare in hedonic terms, a similar rethinking of patent law should not
be far behind.
2. Copyright
Understanding copyright law’s effects on human welfare is considerably more complex
than understanding those of patent law. Few would doubt that contemporary science
and technology are much better than they were in the past. Almost no one would choose
nineteenth-century surgical practice over twenty-first century medicine. But there is
not nearly the same uniformity of judgment about the superiority of Taylor Swift over

Beethoven.1 Although many people accept that science and technology are capable of
‘progressing,’ it is difficult in the modern world to argue the same for the arts (Beebe,
2017). Assuming that one of the goals of creative production is to improve people’s lives,
how could we know if it does so?
The difficulty of answering questions like this one has led to nearly total dominance
of copyright jurisprudence by preferentist thinking. If consumers are willing to spend
money to purchase the latest books, albums, and movies, they must think that doing so
will make them happier. Why should the law judge for people which works will most add
to their happiness? De gustibus non est disputandum—in matters of taste, there can be no
dispute. As Oliver Wendell Holmes, Jr. warned in his famous Bleistein opinion, ‘It would
be a dangerous undertaking for persons trained only to the law to constitute themselves
final judges of the worth of pictorial illustrations, outside the narrowest and most obvious
limits’ (Bleistein v. Donaldson Lithographic Co., 188 U.S. 239 (1903)). Creative works
have value to the extent that they ‘command the interests of any public.’
As Barton Beebe (2017) has recently argued, copyright law’s approach to aesthetic
progress has largely been accumulative—more works and more sales equal more value and
more welfare. From the perspective of consumer welfare, this approach seems difficult to
debate. How could we measure whether television programs contribute more to happiness
than sculpture and dance? Some people prefer television, and some dance. According
to this approach, copyright law should encourage all of these products equally and let
consumers choose.
But do people make accurate judgments about their own welfare when it comes to
consuming content? And do more options lead to more happiness? At least some evidence
of television watching suggests otherwise. Heavy television viewing seems to be associated
with lower levels of subjective well-being, although testing causality is often difficult when
measuring people’s behaviors in the real world (Frey et al., 2007; Bruni and Stanca, 2008).
There is some evidence that people who watch a lot of television do so at the expense
of ‘relational’ activities with others that tend to make them happier (Bruni and Stanca,
2008). And some people, especially heavy television watchers, seem to do far worse when
they have more channels to watch (Benesch et al., 2010). While we do not think that these
studies are strong enough to recommend removing or reducing copyright protection for
television programs, they are important because they illustrate potential limitations with
the preferentist approach that dominates the field.
The ‘more is better’ approach to creative production assumes that allowing people to
find products that precisely match their tastes is valuable. Because consumer tastes are
heterogeneous, people are more likely to find things that maximize their welfare if they
have more options to choose from. But empirical social science research has called this view
into question. Studies have shown that having more choices of products may not create
greater happiness with the product chosen and, in fact, more choices may diminish happi-
ness (Schwartz, 2003; Hsee and Hastie, 2006; Scheibehenne et al., 2010). When confronted
with an enormous variety of goods, people struggle to decide which to choose, and they
focus on irrelevant aspects of goods. In some cases, the people choosing from a wider set
of options are less happy with their choices than those choosing from a narrower set.
1
Even the authors of this chapter disagree.

In addition, the value that people obtain from creative works arises in part from the
fact that they are often embedded within a larger shared cultural experience. Watching
television is fun, but talking about a new episode with friends is often just as much fun.
Copyright law’s accumulationist approach to creativity risks creating smaller and smaller
pockets of cultural and social interaction over new works. Although people may be able to
find works that satisfy their unique preferences, they may find that they have no one with
whom to discuss them. If cultural heterogeneity becomes cultural atomization, people
may lose out on the sorts of relational experiences that are most conducive to happiness.
These concerns raise challenging questions for IP law’s accumulationist approach.
More creative production will not necessarily produce more welfare. Unfortunately, given
the current state of our knowledge, it is difficult to suggest how copyright law can strike
the right balance between heterogeneous tastes and social cohesion. Further research is
needed to more precisely understand the relationship between choice and welfare.
Consumer preferences and consumer welfare may also diverge with respect to the
pace of creative production in certain fields. In many areas, consumers seem to demand
rapid replacement of existing cultural goods with new ones. Fashion design is the leading
example. But consumers’ preferences for novelty may not be consistent with their subjective
well-being. People buy clothes in order to stay on trend even though their old clothes are
still functional (Hemphill and Suk, 2009). As William Shakespeare explained ‘fashion wears
out more apparel than the man’ (Shakespeare, 2004). In these situations, works become
‘obsolete’ before they lose their function and before people stop getting pleasure from them.
In fact, the very thing that makes them obsolete is the existence of new products to consume.
To the extent that this is true (and it may be in many other creative fields as well),
copyright law should not promote rapid cultural ‘progress.’ Unlike for cancer treatments,
where we typically want the pace of innovation to be incredibly high,2 in these areas,
slower rates of creation and replacement may be optimal. If the rate of innovation is
slower, existing works will not lose their hedonic value as quickly, and consumers will not
need to purchase new works to obtain the same amounts of pleasure. From a demand-
side perspective, this is preferable. In the context of fashion design, extending copyrights
might, counterintuitively, improve consumer well-being by slowing down the pace of
innovation. Raustiala and Sprigman (2006) argue that the current lack of protection
leads to shorter fashion cycles as new designs get rapidly copied and become obsolescent.
Giving designers protection for their creations could mean that they are not copied as
rapidly and stay on trend longer, thereby allowing consumers to slow down their purchas-
ing of new clothes. In other contexts, as well, copyright law might pay closer attention to
how its doctrines affect the pace of creative production and obsolescence. For example,
book publishers have little interest in maintaining consumer demand for older fictional
works that are available in cheap used editions if that demand cuts into consumers’ desire
for expensive editions of new works (Maurer, 2015).
In fact, a similar dynamic might exist in the context of some patented goods such
as consumer electronics. A piece of technology such as an iPod or iPhone becomes
2
In fact, we want the pace of innovation to be such that the benefits of a given pace exceed
its costs. Innovation could arise too rapidly even for cancer treatments if the costs associated with
a faster pace exceed its benefits.

obsolete not just when it breaks or cannot handle modern song formats, but also when it
is surpassed by newer and more exciting versions of itself. Yet, as we note above, there is
scant evidence that these newer types of technology are contributing much, if anything,
to happiness. If stronger patent rights slowed the pace of innovation in the computer
and consumer electronics sectors and reduced the rate at which technologies became
obsolete, this might impose little or no welfare cost while preserving resources that could
be employed for other purposes. Such is the paradoxical conclusion we draw from the
fact that improvements in these types of technology do not appear to be making our lives
better.
B. Producer-Side IP Effects
Thus far we have largely concentrated on demand- or consumer-side welfare effects of IP

law and policy. That is, most of the discussion has centered on the consequences of various
policy choices for the well-being of the individuals who consume the fruits of innovation
and creativity. Yet there is an important second half to this question: the effects of IP on
the well-being of the many individuals who produce innovative and creative works. Many
millions of Americans work in industries that are shaped by patent and copyright law,
and their welfare is in some ways dependent upon the form and substance of those laws.
In the sections that follow, we survey how different conceptions of welfare might
compel different types of IP law and policy. To be sure, there are limits on the strength of
the effects that IP policy changes can generate. The work and responsibilities of a software
engineer or semiconductor designer are not likely to change substantially in response to
changes in patent law. The life of a visual artist may or may not undergo drastic revision if
courts amend copyright law at the margin. But IP can impact well-being on the producer
side as well, as explained below.
Here, we believe that patent and copyright law can in many ways produce similar types
of effects. Accordingly, we discuss patent and copyright law together and focus on the
different types of welfare benefits that creative and innovative production can generate
for those individuals on the production side.
1. Creativity as benefit

In the standard law and economics account of IP, the process of creation or invention is
generally viewed as costly. Creating new works or inventions takes time and resources that
people and firms would be unwilling to expend unless they were given the opportunity to
obtain supracompetitive prices for their creations. On this view of creators’ cost-benefit
calculations, the activities associated with creation enter exclusively on the costs side of
the equation. Yet even without the benefit of happiness research it seems evident that this
picture is too narrow. Many people create and invent for the joy that they experience while
doing so (Devlin and Sukhatme, 2009). From the perspective of subjective well-being,
however, we can more fully account for the positive value that creators experience while
engaged in new challenges.
Researchers studying well-being have observed that one of the most pleasurable experi-
ences is ‘flow,’ the sense of being totally engulfed in a challenging task (Csikszentmihalyi,
1990). One of the most reliable ways of experiencing flow is to be involved in the process
of creating something new. Writers, musicians, computer programmers, and scientists

often experience enormous positive affect when they are engaged in activities that prove
difficult but susceptible to completion and mastery. To many of them, the hours spent
writing, coding, and thinking are not costly; they are not simply the means to some other
valuable ends. Instead, the activities associated with creating and inventing are often ends
in their own right. The works and goods generated by these activities often seem to be
by-products of the truly valuable goal.
Once we see the processes of creativity and innovation not simply as costs that must
be borne on the way to generating valuable goods but rather as goods in their own right,
our perspective on IP law and policy changes. Most obviously, the law and economics
account of costs and benefits is likely to go awry if it neglects the substantial benefits
that creators and innovators experience. The addition of exclusive rights as an incentive
to create will often be supernumerary when creators have derived sufficient benefits from
the act of creating in the first place. This is not to suggest that monetary incentives are
never necessary for creative work. There is bound to be substantial variation in the extent
to which authors and inventors experience positive welfare from creating. Love poems
rarely need monetary incentives to be produced, but databases might always need them.
And even though creators might generate substantial amounts of joy from creating, they
still need to live and eat. IP law would do well to understand and adapt to the variations
in producers’ need for external stimuli.
In addition, although creativity itself may need little incentive in many cases, other
costly resources may be necessary to actually produce and distribute the goods associated
with creative production. Although some number of screenplays would likely be written
in the absence of copyright protection, movie studios will not produce movies unless they
have a chance to recoup their investments. If we think that certain kinds of creativity
and innovation demand the expenditure of significant capital resources to emerge, then
we need to make sure that those supplying the capital have sufficient reason to do so.
Thus, when analysing the kinds of IP rights that should be provided, the law should pay
attention at least as much to production and distribution costs as it does to creation costs
(Cohen, 2011).
Finally, IP rights impose restrictions on creators’ opportunities to engage in the kinds
of activities that produce flow. Drastic reductions in the costs of creative production,
particularly in the form of digital technologies, have opened up opportunities for creative
engagement to a larger proportion of the population. One of the principal ways in which
creators engage with one another is through each other’s works. Creative engagement
may take the form of individuals producing separate, independent works that they then
share with one another. Increasingly, however, creators engage with their cultures by
explicitly manipulating works within their cultures (Lee, 2008; von Hippel, 2005). In
many contexts, creativity involves reworking and remixing existing ideas and inventions
(Banks and Deuze, 2009). But IP laws, and copyright law’s derivative works doctrine in
particular, can put creators at risk when they interact with protected works. Although
copyright and patent laws might not affect artists who draw for private consumption and
pleasure or inventors who enjoy tinkering but have no interest in making and selling their
inventions, the law may limit their abilities to share their efforts with others. For example,
an aspiring musician may derive considerable happiness from manipulating and editing
existing works and sharing them on the Internet as a way of demonstrating mastery and
connecting with a community (Silbey, 2014). But when she posts her work online, she

may have infringed one or more copyrights. By limiting the uses that others can make
of creative and innovative works, IP laws impede the potential benefits that follow-on
creators experience from interacting with existing works and ideas. Just as the law needs to
account for the pleasure that creators experience when generating new works, it needs to
account for how IP laws potentially hinder the happiness of others who want to creatively
engage with those works.
2. Employment effects

Patent and copyright law’s primary function is to redistribute surplus from the production
of creative and innovative goods from consumers to producers. Instead of the goods
being sold at the competitive price—which maximizes consumer surplus—patents and
copyrights permit producers to sell at monopoly prices, which maximizes producer
surplus at the expense of significant consumer surplus. This surplus is typically thought
of in monetary terms, which reinforces the connection to a preference-based conception
of welfare. For preference-satisfaction theories of welfare, patent and copyright law thus
have significant first-order effects on welfare. If stronger IP laws enable producers—the
owners (and employee-owners) of technology companies—to get rich from their inven-
tions, this will increase their welfare significantly. As weaker IP rules diminish the returns
from innovation and creativity, so too would producer-side welfare gains diminish. For this
reason, IP scholars often describe stronger IP rules as benefitting large corporations (by
which they mean the owners of those corporations) at the expense of individual consumers.
From a hedonic perspective, however, the picture is not so clear. As we explained above,
even significant increases in wealth have only minor effects on subjective well-being. Thus,
marginally greater returns on investment will do little to improve the happiness of those
who are already well off. It might seem, then, that IP law will not substantially affect the
welfare of innovators.
There is, however, one respect in which IP law could meaningfully affect producer-side
welfare from a hedonic perspective. The mechanism is the role of IP rules, and innovation
and creativity more generally, in promoting employment. Unemployment has dramatic
negative effects on individual well-being, above and beyond its effects on income and
wealth (Bronsteen, Buccafusco, and Masur, 2013). That is, not only does involuntary
unemployment reduce an individual’s wealth and income, often to the point at which those
reductions begin to affect well-being, it can also exact a psychological toll, lead to depres-
sion, and generally produce unhappiness. While patent and copyright laws that increase
wealth will not have significant hedonic effects on people who are already relatively well
off, laws that increase employment very well might.
Patent law can, of course, affect overall employment if it affects economic growth
and development. Policymakers should endeavor to select patent rules that maximize
economic growth, though of course this will be extremely difficult (United States Patent
and Trademark Office, 2013). At a more administrable level, patent law can also be used
to favor certain industries over others, as we described in the previous section. There, we
explained that certain types of technologies might produce greater welfare gains than
others among the consumer population. Similarly, some technological industries might be
more labor-intensive, and others might be more capital-intensive. A shift of R&D invest-
ment from a capital-intensive to a labor-intensive technology sector will increase overall
technology-based employment, all other things being equal. There is now ample evidence

that these types of investments can affect overall employment and that labor markets do
not function perfectly (Masur and Posner, 2012). Consequently, shifts in patent law that
reroute R&D investment could conceivably increase employment at the margin and thus
increase welfare, particularly when viewed from a hedonic perspective. Of course, it is
difficult to know the magnitude of this effect. In addition, policymakers and courts would
need detailed data on the relative labor intensity of various areas of technology. But the
potential for beneficial policy intervention exists.
Similar effects are likely present with respect to production of creative works. Stronger
copyright rules will likely increase the production of some types of works—blockbuster
movies, or mass-market novels, for instance—but decrease the production of others, such
as fan fiction or other types of derivative works. Some of these types of works might be
more or less labor-intensive than others; some might also generate more full-time employ-
ment than others.3 We do not have enough data to offer any definitive prescriptions yet.
But a hedonically oriented policymaker would be well-advised to study the employment
effects of various types of creative production and incorporate the welfare effects of
employment and unemployment into the formulation of IP law and policy.
V. CONCLUSION
IP laws exist to improve the quality of people’s lives by providing creations and innova-
tions that make them better off. Although legal scholars have spent lots of time debating
whether particular doctrines succeed in this goal, they have spent comparatively little time
analysing what it means to improve the quality of people’s lives. IP law is committed to
welfarism, but it has not grappled with what human welfare is. In this chapter, we have
explained three competing conceptions of welfare and defended one of them—welfare as
subjective well-being. Moreover, we have begun to suggest ways in which these different
conceptions of welfare might affect IP law and policy. Definitions of welfare are not
abstract philosophical questions; they have real-world significance that scholars should
care about. Although sufficient data is still lacking to answer most IP law questions, new
advances in data gathering and analysis will soon provide opportunities for rethinking
many core doctrines.
REFERENCES
Adler, Amy. 2009. ‘Against Moral Rights,’ 97 California Law Review 263–300.
Adler, Matthew D., and Eric A. Posner. 2006. New Foundations of Cost-Benefits Analysis. Cambridge, MA:
Harvard University Press.
Adler, Matthew D., Paul Dolan, and Georgios Kavestos. 2015. ‘Would You Choose to Be Happy? Tradeoffs
between Happiness and the Other Dimensions of Life in a Large Population Survey,’ accessed at https://
papers.ssrn.com/sol3/papers.cfm?abstract_id=2640117 (February 24, 2019).
3
To some extent, the suggestion above that copyright doctrine should consider slowing the
pace of creativity to produce consumer-side benefits could conflict with the suggestion here about
employment and producer-side benefits. As with most things, the devil will be the in the details of
data that we do not yet have.

Balganesh, Shyamkrishna. 2009. ‘Foreseeability and Copyright Incentives,’ 122 Harvard Law Review 1569–633.
Banks, John, and Mark Deuze. 2009. ‘Co-Creative Labour,’ 12 International Journal of Cultural Studies 419–31.
Beebe, Barton. 2017. ‘Bleistein: The Problem of Aesthetic Progress, and the Making of American Copyright
Law,’ 117 Columbia Law Review 319–97.
Benesch, Christine, Bruno S. Frey, and Alois Stutzer. 2010. ‘TV Channels, Self-Control and Happiness,’ 10 B.E.
Journal of Economic Analysis and Policy 1–33.
Boyle, James. 2003. ‘Enclosing the Genome? What the Squabbles over Genetic Patents Could Teach Us,’ in
F. Scott Kieff, ed., Perspectives on Properties of the Human Genome Project. New York, NY: Academic Press.
Boyle, James. 2007. ‘Cultural Environmentalism and Beyond,’ 70 Law and Contemporary Problems 5–21.
Bronsteen, John, Christopher Buccafusco, and Jonathan S. Masur. 2013. ‘Well-Being Analysis vs. Cost-Benefit
Analysis,’ 63 Duke Law Journal 1603–89.
Bronsteen, John, Christopher Buccafusco, and Jonathan S. Masur. 2014. Happiness and the Law. Chicago, IL:
Bruni, Luigino, and Luca Stanca. 2008. ‘Watching Alone: Relational Goods, Television and Happiness,’ 65
Journal of Economic Behavior and Organization 506–28.
Burk, Dan L., and Mark A. Lemley. 2009. The Patent Crisis and How the Courts Can Solve It. Chicago, IL:
Cass, Ronald A., and Keith N. Hylton. 2013. Laws of Creation: Property Rights in the World of Ideas.
Cambridge, MA: Harvard University Press.
Cohen, Julie. 2011. ‘Copyright as Property in the Post-Industrial Economy: A Research Agenda,’ 2011 Wisconsin
Csikszentmihalyi, Mihaly. 1990. Flow: The Psychology of Optimal Experience. New York: Harper & Row.
Demsetz, Harold. 1969. ‘Information and Efficiency: Another Viewpoint,’ 12 Journal of Law and Economics 1–22.
Derclaye, Estelle. 2012. ‘Eudemonic Intellectual Property: Patents and Related Rights as Engines of Happiness,
Peace, and Sustainability,’ 14 Vanderbilt Journal of Entertainment and Technology Law 495–543.
Derclaye, Estelle. 2013. ‘What Can Intellectual Property Law Learn from Happiness Research?,’ in Graeme B.
Dinwoodie, ed., Methods and Perspectives in Intellectual Property. New York, NY: Edward Elgar.
Devlin, Alan, and Neel Sukhatme. 2009. ‘Self-Realizing Inventions and the Utilitarian Foundation of Patent
Law,’ 51 William and Mary Law Review 897–955.
Fisher, William W., and Talha Syed. 2007. ‘Global Justice in Healthcare: Developing Drugs for the Developing
World’, 40 U.C. Davis Law Review 581–678.
Foot, Philippa. 2001. Natural Goodness. Oxford: Clarendon Press.
Frey, Bruno S., Christine Benesch, and Alois Stutzer. 2007. ‘Does Watching TV Make Us Happy?,’ 28 Journal
of Economic Psychology 283–313.
Frischmann, Brett M. 2017. ‘Capabilities, Spillovers, and Intellectual Progress: Toward a Human Flourishing
Theory for Intellectual Property,’ 14 Review of Economic Research on Copyright Issues 1–38.
Gilbert, Dan. 2006. Stumbling on Happiness. New York, NY: Knopf.
Griffin, James. 1986. Well-Being. Oxford: Oxford University Press.
Hemel, Daniel J., and Lisa Larrimore Ouellette. 2013. ‘Beyond the Patent-Prizes Debate,’ 92 Texas Law Review
303–82.
Review 1148–99.
Hsee, Christopher K., and Reid Hastie. 2006. ‘Decision and Experience: Why Don’t We Choose What Makes
Us Happy?,’ 10 Trends in Cognitive Sciences 31–7.
Hughes, Justin. 1988. ‘The Philosophy of Intellectual Property,’ 77 Georgetown Law Journal 287–330.
Kahneman, Daniel, and Angus Deaton. 2010. ‘High Income Improves Evaluation of Life but Not Emotional
Well-Being,’ 107 PNAS 16489–93.
Kahneman, Daniel, Ed Diener, and Norbert Schwarz. 1999. ‘Preface,’ in Daniel Kahneman, Ed Diener,
and Norbert Schwarz, eds., Well-Being: The Foundations of Hedonic Psychology. New York: Russell Sage
Foundation.
Kahneman, Daniel, Alan B. Krueger, David A. Schkade, Norbert Schwartz, and Arthur A. Stone. 2004. ‘A
Survey Method for Characterizing Daily Life Experience: The Day Reconstruction Method,’ 306 Science
1776–80.
Kapczynski, Amy. 2012. ‘The Cost of Price: Why and How to Get beyond Intellectual Property Internalism,’
Kapczynski, Amy, and Talha Syed. 2013. ‘The Continuum of Excludability and the Limits of Patents,’ 122 Yale
Kwall, Roberta Rosenthal. 2009. The Soul of Creativity: Forging a Moral Rights Law for the United States. Palo
Alto, CA: Stanford University Press.
Landes, William M., and Richard A. Posner. 2003. The Economic Structure of Intellectual Property Law.
Cambridge: Belknap Press.

Lee, Edward. 2008. ‘Warming Up to User-Generated Content,’ 2008 University of Illinois Law Review 1459–548.
Lemley, Mark A. 2005. ‘Property, Intellectual Property, and Free Riding’, 83 Texas Law Review 1031–75.
Lewinsohn-Zamir, Daphna. 2003. ‘The Objectivity of Well-Being and the Objectives of Property Law’, 78 NYU
Manta, Irina. 2013. ‘Hedonic Trademarks,’ 74 Ohio State Law Journal 241–83.
Masur, Jonathan S. 2016. ‘CBA at the PTO,’ 65 Duke Law Journal 1701–35.
Masur, Jonathan S., and Eric A. Posner. 2011. ‘Climate Regulation and the Limits of Cost-Benefit Analysis,’ 99
Masur, Jonathan S., and Eric A. Posner. 2012. ‘Regulation, Unemployment, and Cost-Benefit Analysis,’ 98
Maurer, Stephen. 2015. ‘The Economics of Memory: How Copyright Decides Which Books Do (and Don’t)
Become Classics,’ 14 John Marshall Review of Intellectual Property Law 521–63.
Menell, Peter S., and Suzanne Scotchmer. 2007. ‘Intellectual Property Law,’ in A. Mitchell Polinsky and Steven
Shavell, eds., Handbook of Law and Economics, Vol. II. Amsterdam: North Holland.
Nielsen. 2014. ‘How Smartphones Are Changing Consumers’ Daily Routines around the Globe,’ accessed at
http://www.nielsen.com/us/en/insights/news/2014/how-smartphones-are-changing-consumers-daily-routines-
around-the-globe.html (February 24, 2019).
Nussbaum, Martha C. 2001. Women and Human Development: The Capabilities Approach. Cambridge:
O’Connor, Sean M. 2015. ‘The Overlooked French Influence on the Intellectual Property Clause,’ 82 University
of Chicago Law Review 733–830.
Oliar, Dotan. 2006. ‘Making Sense of the Intellectual Property Clause: Promotion of Progress as a Limitation
on Congress’s Intellectual Property Power,’ 94 Georgetown Law Journal 1771–845.
Oswald, Andrew J., and Nattavudh Powdthavee. 2008. ‘Does Happiness Adapt? A Longitudinal Study of
Disability with Implications for Economists and Judges,’ 92 Journal of Public Economics 1061–77.
Powdthavee, Nattavudh, and Bernard van den Berg. 2011. ‘Putting Different Price Tags on the Same Health
Condition: Re-Evaluating the Well-Being Valuation Approach,’ 30 Journal of Health Economics 1032–43.
Rai, Arti. 2007. ‘The Ends of Intellectual Property: Health as a Case Study,’ 70 Law and Contemporary Problems
125–30.
Raustiala, Kal, and Christopher Sprigman. 2006. ‘The Piracy Paradox: Innovation and Intellectual Property in
Saeb, Sohrab, Mi Zhang, Christopher J. Karr, Stephen M. Schueller, Marya E. Corden, Konrad P. Kording,
and David C. Mohr. 2015. ‘Mobile Phone Sensor Correlates of Depressive Symptom Severity in Daily-Life
Behavior: An Exploratory Study,’ 17 Journal of Medical Internet Research e175, 1–11.
Sampat, Bhaven N., and Heidi L. Williams. 2015. ‘How Do Patents Affect Follow-On Innovation? Evidence from
the Human Genome,’ accessed at http://papers.ssrn.com/sol3/papers.cfm?abstract_id=2679705 (February 24,
2019).
Scheibehenne, Benjamin, Rainer Greifeneder, and Peter M. Todd. 2010. ‘Can There Ever be Too Many Options?
A Meta-Analytic Review of Choice Overload,’ 37 Journal of Consumer Research 409–25.
Schwartz, Barry. 2003. The Paradox of Choice: Why More Is Less. New York, NY: Ecco.
Sen, Amartya. 1993. ‘Capability and Well-Being,’ in Martha Nussbaum and Amartya Sen, eds., The Quality of
Life. Oxford: Clarendon Press.
Sen, Amartya. 2005. ‘Human Rights and Capabilities,’ 6 Journal of Human Development 151–66.
Shakespeare, William. 2004. Much Ado about Nothing. New York: Washington Square Press.
Silbey, Jessica. 2014. The Eureka Myth: Creators, Innovators and Everyday Intellectual Property. Palo Alto, CA:
Stanford Law Books.
Solum, Lawrence B. 2002. ‘Congress’s Power to Promote the Progress of Science: Eldred v. Ashcroft,’ 36 Loyola
Los Angeles Law Review 1–83.
Stutzer, Alois, and Bruno S. Frey. 2008. ‘Stress that Doesn’t Pay: The Commuting Paradox,’ 110 Scandinavian
Sumner, L.W. 1996. Welfare, Happiness, and Ethics. Oxford: Clarendon Press.
Tur-Sinai, Ofer M. 2016. ‘Technological Progress and Well-Being,’ 48 Loyola University of Chicago Law Journal
145–204.
United States Patent and Trademark Office. 2013. ‘Regulatory Impact Analysis, Setting and Adjusting Patent
Fees in Accordance with Section 10 of the Leahy-Smith America Invents Act,’ accessed at http://www.
uspto.gov/sites/default/files/aia_implementation/AC54_Final_Regulatory_Impact_Analysis.pdf (February
24, 2019).
von Hippel, Eric. 2005. Democratizing Innovation. Cambridge, MA: MIT Press.

Cases
Aronson v. Quick Point Pencil Co., 440 U.S. 257 (1979).

Bleistein v. Donaldson Lithographic Co., 188 U.S. 239 (1903).
Kewanee Oil Co. v. Bicron Corp., 416 U.S. 470 (1974).
Sony Corp. v. Universal Studios, 464 U.S. 417 (1984).

17 U.S.C. § 106A (2018).

6. Economic models of innovation: stand-alone and
cumulative creativity
Peter S. Menell* and Suzanne Scotchmer**
4 5
Contents
I. Economic Models of Technical Change

II. Stand-Alone Innovation
A. Threshold for Protection
B. Duration
C. Breadth
D. Rights of Others (and Defenses)
1. Rights arising from independent invention
2. Rights after sale
3. Rights to share copyrighted works
E. Remedies
F. Channeling Doctrines
III. Cumulative Innovation
A. A Preliminary Model: The Virtues of Licensing
B. Duration
C. Threshold Requirements and Breadth
1. Evolving doctrines regarding subject matter
2. Adequacy of disclosure requirements
D. Rights of Others (and Limitations and Defenses)
1. Experimental use
2. Fair use
3. Reverse engineering
4. Other limitations: exemptions, compulsory licenses, voluntary
licenses, open source, and dedication to the public domain
** Koret Professor of Law and Director, Berkeley Center for Law and Technology, Berkeley
School of Law, University of California.
** Formerly Professor of Economics, Public Policy, and Law, University of California,
Berkeley, who died in 2014. Thanks to Concord Cheung, Amit Elazari, Megan McKnelly, and Reid
Whitaker for excellent research assistance. This chapter updates Parts 1.1 and 1.3 of Peter S. Menell
and Suzanne Scotchmer. 2007. ‘Intellectual Property Law,’ in A. Mitchell Polinsky and Steven
Shavell, eds., Handbook of Law and Economics, vol. II. Amsterdam: North-Holland. Aspects of
Parts 1.1 and 1.3 are reprinted with permission from Elsevier, © 2007. http://www.elsevier.com.
This chapter is dedicated to Professor Scotchmer for her seminal contributions to the economics of
intellectual property rights. Professor Menell bears responsibility for any mistakes in this updating
of the prior work.
119

E. Remedies
References
Intellectual property law seeks to promote progress in science, technology, and expressive
creativity (U.S. Const. art. I, § 8, cl. 8). The principal economic justification for intellectual
property derives from a fundamental problem: the inability of a competitive market to
support an efficient level of innovation. In a competitive economy, profits will be driven to
zero, not accounting for sunk costs such as research and development (R&D) or costs of
authorship. From an ex post point of view, this is a good outcome, as it keeps prices low
for consumers and avoids deadweight loss. But from an ex ante point of view, it produces
a sub-optimal level of investment in R&D. Most firms would not invest in developing new
technologies, and potential creators might not spend their time on creative works, if rivals
could enter the market and dissipate the profits.
Unlike tangible goods, knowledge and creative works are public goods in the sense
that their use is nonrival (Arrow, 1962; Nelson, 1959). One agent’s use does not limit
another agent’s use. Indeed, in its natural state (cartooned in the digital age as ‘bits want
to be free’), knowledge is also ‘nonexcludable.’ That is, even if someone claims to own
the knowledge, it is difficult to exclude others from using it. Intellectual property law is
an attempt to solve that problem by legal means; it grants exclusive use of the protected
knowledge or creative work to the creator. For other forms of property, exclusion is often
accomplished by physical means, such as building a fence. Intellectual property is a legal
device by which the inventor can control entry and exclude users from intangible assets.
Intellectual property results in deadweight loss to consumers, and that is its main
defect. Two other defects are that it may inhibit the use of scientific or technological
knowledge for further research, and, from an ex ante point of view, that there is no
guarantee that the research effort will be delegated efficiently to the most efficient firms
or even to the right number of firms. Commentators have been lamenting the defects of
intellectual property since the nineteenth century, in more or less the same terms as today
(Machlup and Penrose, 1950).
But intellectual property also has virtues, of which we mention three powerful ones.
Probably the greatest virtue is that every invention funded with intellectual property cre-
ates a Pareto improvement. No one is taxed more than her willingness to pay for any unit
she buys; else, she would not buy it. In contrast, funding out of general revenue runs the
risk of imposing greater burdens on individual taxpayers than the benefits they receive.
A second great virtue is decentralization. Probably the most important obstacle to
effective public procurement is in finding the ideas for invention that are widely distrib-
uted among firms and inventors. The lure of intellectual property protection does that
automatically. Decentralization is especially important if private inventors are more likely
than public sponsors to spawn good ideas for innovations.
The third virtue is that intellectual property is an effective screening device (Long, 2002
(emphasizing the role of patents as a signaling device)). Since the private value of the
invention often reflects the social value, inventors should be willing to bear higher costs
for inventions of higher value.
But these are not determinative, since other incentive mechanisms such as prizes and
government funding may share the same virtues while at the same time reducing deadweight

Economic models of innovation 121
loss. Whereas the earlier economics literature proceeded as if intellectual property protec-
tion was the self-evident solution to the R&D incentive problem, the modern literature
has sought to understand when that is true and when other incentive mechanisms might
dominate (Wright, 1983; Abramowicz, 2019; Price, 2018). The choice among incentive
mechanisms, and even the optimal design of intellectual property laws, depends on the
nature of the creative process or, in economists’ jargon, on the model of knowledge creation.
Thus, this chapter proceeds in three stages. Section I sketches the principal economic
models of technical change. Section II traces the economic modeling of stand-alone
innovation. These models provide the foundation for cumulative innovation, the most
prevalent and significant driver of economic activity, addressed in Section III. A recurrent
theme is that advances in digital and network technologies have vastly eased access to
prior works and reduced the costs of using them in new works.
I. ECONOMIC MODELS OF TECHNICAL CHANGE
Four principal models of technological change have been proposed in the economics
literature: an evolutionary model, a model of induced technical change, a production
function for knowledge, and an exogenous process of idea formation, with incentives
determining investments.
In the evolutionary model proposed by Nelson and Winter (1982) (see also Mokyr,
1990, ch. 11), technology develops through a process in which R&D investments occur
whenever profit drops below a specified level. Hence, the evolutionary model is not set up
to investigate incentives since investment is automatic.
In the model of induced technical change, innovation occurs in response to changes in
factor prices: ‘A change in the relative prices of the factors of production is itself a spur
to invention and inventions of a particular kind—directed at economizing the use of a
factor which has become relatively expensive’ (Hicks, 1932, pp. 124–5; see also Ruttan,
2001). Thus, rising energy prices can be expected to spur technological advances in energy
conservation (Newell et al., 1999).
In the production function model of discovery, which is the basis of almost all of the
patent race literature, there is an exogenously given relationship which determines, as a
function of research inputs or the number of researchers, either the quality of invention
(de Laat, 1996; Shavell and van Ypersele, 2001; Che and Gale, 2003) or the likelihood of
success in each time period (Loury, 1979; Lee and Wilde, 1980; Reinganum, 1982, 1985,
1989; Wright, 1983; Denicolò, 1996, 1997). In both the induced technical change model
and the production function model, the profit opportunities are common knowledge.
Decentralization is not important.
The ‘ideas’ model of O’Donoghue et al. (1998) (see also Scotchmer, 1999; Maurer
and Scotchmer, 2004) focuses directly on the scarcity of ideas. The basis of research is
‘imagination,’ and to achieve an innovation, a researcher must both have the idea for the
innovation and an incentive to invest in it.
Although it is the most widely used model, the production function model does not lead
naturally to intellectual property as superior to other incentive schemes. For example, the
advantages of decentralization are more important in a model where ideas are scarce than
where they are common knowledge.

II. STAND-ALONE INNOVATION
Much of the early economic modeling of the role of intellectual property in promoting
innovation posed the following question: what system of incentives or rewards would
best promote the attainment of a particular invention? Such models provide the basis for
analyzing legal protection for a distinct and relatively narrow class of inventions which
do not ultimately generate follow-on innovation. Examples from this class include the
safety razor, the ballpoint pen, and pharmaceutical innovations for which the scientific
mechanism is poorly understood (Nelson and Winter, 1982; von Hippel, 1988, p. 53). Even
where inventors depend on prior knowledge, which is almost inevitable, the lag may be
such that prior rights have expired so that the incentive system treats inventions as stand-
alone. Examples are the bicycle and the early development of the light bulb (Dyson, 2001).
Models focusing on stand-alone innovation can also be helpful in analyzing legal
protection for expressive creativity. Although such works often draw upon prior works for
inspiration or common reference points for the work’s audience, many authors, musicians,
and artists have not traditionally built so extensively upon the work of prior creators
as to require express permission. This proposition obviously turns on the underlying
right structures—copyrights tending to be relatively narrow in comparison to utility
patents—but it also reflects a fundamental difference between the fields of technological
and expressive innovation:
Science and technology are centripetal, conducing toward a single optimal result. One water
pump can be better than another water pump, and the role of patent and trade secret law is to
direct investment toward such improvements. Literature and the arts are centrifugal, aiming
at a wide variety of audiences with different tastes. We cannot say that one novel treating the
theme, say, of man’s continuing struggle with nature is in any ultimate sense ‘better’ than another
novel—or musical composition or painting—on the same subject. The aim of copyright is to
direct investment toward abundant rather than efficient expression. (Goldstein, 1986)
As copyright has expanded into functional subject matters, such as the design of useful
articles, computer software, and architectural works, however, this distinction breaks
down. Channeling doctrines aim to steer technological advances into the utility patent
regime (Menell, 1987).
The critical inquiry in seeking to promote stand-alone innovation is how much profit an
inventor or creator should receive and how it should be structured. The focus for stand-
alone innovation, therefore, is upon ex ante incentives. As we will emphasize below, all
of the results in this area depend sensitively on what is assumed about licensing (see also
Gallini and Scotchmer, 2002). Collaboration and the exchange of technological knowl-
edge across firm boundaries encounter substantial transaction costs. Arora et al. (2001)
find evidence that changes in the technology of technical change—most notably the grow-
ing use of digital information technologies—facilitate greater partitioning of innovation
tasks across traditional firm boundaries. They foresee markets for t echnology—licensing
and specialized technology transfer and innovation service firms—playing a more
significant role in the production of innovation. (When we turn to cumulative innovation
in Section III, ex post incentives enter the analysis.)
The principal policy levers affecting incentives to invent are the threshold for protection,
duration, breadth, rights of others (and defenses), remedies, and channeling doctrines.

A. Threshold for Protection
Because intellectual property protection results in deadweight loss to consumers (and

can interfere with cumulative innovation), it should only be available for significant
innovation—works that are new and would not be readily forthcoming without legal
encouragement—that is, requiring substantial cost and substantial technological risk.
Works already in the public domain should not be protectable and the threshold for
protection should be sufficiently high (or the rights sufficiently narrow) to prevent easily
achieved (‘obvious’) advances from being insulated from free market competition.
Intellectual property regimes erect several types of threshold doctrines before
protection is granted: (1) subject matter rules—categorical limitations on eligibil-
ity; (2) substantive requirements—minimum criteria for protection; and (3) formal
requirements—administrative and technical rules that must be complied with in order to
obtain and maintain protection. Utility patent protection is available for nearly all forms
of technological innovation (processes, machines, manufactures, and compositions of
matter), but applies relatively stringent standards (utility, novelty, nonobviousness (or
inventive step), and adequate disclosure) through a formal examination system.
By contrast, copyright protection applies a very low threshold for protection—a
work need only be fixed in a tangible medium of expression and reflect a modicum of
originality—and does not require examination (registration is optional). Such a low
threshold is counterbalanced by a relatively narrow scope of protection and limiting
doctrines excluding protection for technological features.
Design patent protection is similar to copyright protection for pictorial, graphic,
and sculptural copyrighted works. Its ornamentality/non-functionality doctrine guards
against protecting technological innovation as opposed to aesthetic creativity.
Trade secret law requires merely that information derives economic value from not
being generally known or readily ascertainable (by proper means) by others and from
being the subject of reasonable efforts under the circumstances to maintain its secrecy.
Trade secret protection does not, however, stand in the way of independent invention or
reverse engineering.
Utility patent law’s threshold requirements have received the most economic scrutiny.
Due to the relatively uniform nature of patent protection, some have argued that certain
classes of innovation (such as computer software and business methods) that may not
require such lengthy protection should be subject to a sui generis form of protection
(Menell, 1987 (software)) or excluded from intellectual property protection altogether
(Thomas, 1999 (business methods); Dreyfuss, 2000; Menell, 2007, 2011). The basic
contours of patent law were established during an age of mechanical innovation and
were designed with this model (and the guild system that predominated) in mind
(Menell et al., 2018, vol. I, ch. III). Mechanical innovation continued to comprise
the bulk of patent applications well into the twentieth century. During the past half
century, however, various newer fields—such as software, business methods, and
biotechnology—have increasingly come into the patent system (Allison and Lemley,
2002), calling into question the premises on which patent law was built. If specialized
protection systems are not developed to address new and distinctive fields of innovation
(as was partially done in the case of semiconductor chip designs in the Semiconductor
Chip Protection Act of 1984), the challenge remains of reshaping the relatively uniform

patent system to accommodate the growing heterogeneity of inventive activity (compare

Burk and Lemley, 2002, 2003).
Patent law’s novelty requirement—what it means to be ‘first’—turns on the location of
the ‘finish line’ in the race to invent. Most patent systems in the world apply a first-to-file
standard. Until 2013, the United States determined the winner on the basis of who was
the first to invent. In principle, the first-to-invent system rewards the first inventor to
discover new knowledge, even if they lack the specialized patent filing resources of others.
Thus, many small inventors defended the first-to-invent system as a means of leveling the
playing field relative to large companies which have more resources available and person-
nel in place to file applications expeditiously.
The first-to-file system significantly reduces the administrative costs of operating a
patent system—priority depends solely on the time and date stamped on an application.
Evidentiary disputes over the subtle nuances of who was first to grasp an invention can
be quite costly to resolve (Macedo, 1990). Empirical studies cast doubt on the notion that
small inventors tend to do better under a first-to-invent system, likely reflecting the high
costs of resolving priority disputes (Mossinghoff, 2002; Lemley and Chien, 2003).
The first-to-invent system also has incentive effects as to the choice between trade
secrecy and disclosure (Scotchmer and Green, 1990). Inventors may be inclined to delay
their applications in order to effectively extend the expiration date of a patent (20 years
from the date of filing). In order to counteract this effect and promote prompt filing,
US patent law adds an additional layer of legal complexity (and hence uncertainty and
cost): requiring that an inventor file an application within one year after the invention is
disclosed (either through patenting or publication anywhere in the world or in public use
or on sale in the United States). This reduces the delay in disclosure of new knowledge, but
does not eliminate it. The first-to-file system promotes earlier disclosure of technological
advances. Grushcow (2004) finds that the growing interest in patenting by academic
institutions since 1980 has delayed the publication of research, potentially increasing the
risk of wasteful duplication of research.
On March 16, 2013, the United States shifted from the first-to-invent system to a
first-to-file system that recognizes a one-year grace period for inventor disclosure (35
U.S.C. § 102, as amended by the America Invents Act of 2011). The grace period serves to
encourage prompt disclosure while affording inventors leeway to get their patent applica-
tions filed. Such pre-filing disclosure even by the inventor, however, will bar patenting in
most other parts of the world.
From an economic standpoint, patent law’s nonobviousness standard plays the most
important role in determining which innovations qualify for protection (and hence
what type of innovation patent law encourages). Patent law specifies that a claimed
invention must go beyond readily predictable or conventional solutions to technical,
engineering, or business problems. Articulating an objective and determinative standard
for nonobviousness, however, has proven elusive. In the 1940s, US courts interpreted
the law to require a ‘flash of creative genius’ (Cuno Engineering Corp. v. Automatic
Devices Corp., 314 U.S. 84 (1941)). Such a demanding formulation generated a backlash
within the patent community, leading Congress to frame the standard in the following
manner: a patent may not be obtained ‘if the differences between the claimed invention
and the prior art are such that the claimed invention as a whole would have been obvious
before the effective filing date of the claimed invention to a person having ordinary skill

in the art to which the claimed invention pertains’ (35 U.S.C. § 103, as amended by the
America Invents Act of 2011).What raises the nonobviousness hurdle above the novelty
standard is that the patent examiner may consider multiple references simultaneously
and use logical inferences and common sense. The examiner must also consider circum-
stantial evidence of nonobviousness (so-called ‘secondary considerations’)—long-felt
but unsolved need, commercial success of the claimed invention, failed efforts by
others, copyright by others, praise for the invention, unexpected results, and disbelief of
experts—but only to the extent that such factors are connected to the inventive aspects
of the patent.
In its actual formulation and application, the nonobviousness rule falls short of
implementing the economic gatekeeping principle. Whereas an economist would consider
paramount among relevant considerations the R&D expense in making an invention
(Merges, 1992), the US Patent Act states that patentability ‘shall not be negated by the
manner in which the invention was made,’ implying that inventions requiring minimal
effort (and hence likely to be obtained even without protection) may nonetheless qualify
for protection. In addition, R&D are not among the traditional secondary considerations,
although several court decisions on nonobviousness take note of such factors (Merges,
1992 (noting that the threshold for patentability should be lowered with regard to high-
cost research); Oddi, 1989, p. 1127 (recommending that courts expressly consider ‘qualita-
tive and quantitative investment in research and development’ among the secondary
factors)). The legal standard for nonobviousness does consider the level of uncertainty
involved in research.
In practice, the test depends on the number of parameters and the extent to which the
relevant prior art guides the experimentation process. The role of commercial success
in the nonobviousness determination has produced conflicting economic analyses and
prescriptions. Drawing upon historical and empirical research on the innovation process,
Merges (1988) finds commercial success to be a poor proxy for technical advance. What
succeeds in the market tends to reflect product strategy and marketing more than techni-
cal advances over the prior art. Hence, Merges argues for downplaying this factor and
scrutinizing the connection between market success and the technical advance. By con-
trast, Kitch (1977, p. 283) sees market success as consistent with the prospect theory. By
using subsequent economic success as a factor favoring patentability, patent law increases
‘the security of the investment process necessary to maximize the value of the patent.’
Both analyses support the idea that the consideration of market success in assessing
nonobviousness promotes commercialization, although it is not clear that a patent system
is needed to achieve this end. Where adequate incentives exist to invent, free market
forces should be adequate to promote commercialization (but compare Sichelman, 2010
(proposing a ‘commercialization’ patent); Kieff, 2001a (articulating a commercialization
theory of patent law)).
The nonobviousness standard may have some perverse collateral effects on the nature
and timing of disclosure of new knowledge. By preemptively publishing work in progress,
a firm that is ahead may induce a shake-out among rivals by raising the level of prior art
to render a rival’s subsequent invention obvious (Scotchmer and Green, 1990); a laggard
in a patent race may be able to reduce the likelihood that a leader will be able to obtain a
patent by raising the level of the prior art sufficiently to defeat patentability by a leader
(Lichtman et al., 2000; Parchomovsky, 2000; compare Bar-Gill and Parchomovsky, 2003).

This possibility might also lead competitors to collude or collaborate to maximize patent
opportunities. Under this theoretical account, a higher standard of nonobviousness
increases the viability of a preemptive patenting strategy. The likelihood that such a strat-
egy would be pursued by rivals has been questioned on doctrinal and practical grounds
(Merges, 2004a, pp. 195–6; Eisenberg, 2000; compare Hicks, 1995).
Much of the economics literature on trade secrets addresses the optimal level of
expenditures to maintain secrecy—that is, what constitutes ‘reasonable efforts’ under
the circumstances. Kitch (1980) argues that all such ‘fencing costs’ are inefficient and
would require only such expenses as are necessary to provide evidence of the existence
of a trade secret—that is, a notice or marking function. Friedman et al. (1991) make
the related point that trade secret protection should be available when it is cheaper
than the physical precautions that would be necessary to protect a particular piece of
information.
B. Duration
Nordhaus (1969) offers the first formal model of the optimal duration of intellectual
property protection. Nordhaus asks why the life of the intellectual property right should
be limited, since a longer right leads to more innovation and more innovation creates
social benefit. He argues that there is a countervailing cost. The longer right might increase
innovation, but it also increases deadweight loss on all the inframarginal innovations that
would occur even with shorter protection—that is, innovations that would be forthcoming
even in the absence of the longer right. The optimal duration of a patent or copyright
should balance the incentive effect against the deadweight loss in order to maximize social
welfare. Many economists believe that copyright duration (life of the author plus 70 years)
is much longer than justified to provide an appropriate ex ante incentive for creation of
new works (see Akerlof et al., 2003; but compare Landes and Posner, 2003, p. 218 (noting
that the deadweight loss from copyright protection is relatively small due to the narrow
scope of copyright protection)).
To see the Nordhaus argument, suppose that there is a universe of ‘ideas’ available
for investment. Let an idea be a pair (s, c) where s measures the value of the resulting
innovation and c is its cost. An idea with higher s can be interpreted as leading to a larger
market; a higher s means that the demand curve is shifted out. Let Π(s, T) be the profit
available to a right holder with an intellectual property right of length T and an idea of
quality s. Π is increasing in both T and s. Let W(s, T) be the corresponding social welfare
associated with investment in the idea. The welfare W(s, T) is the sum of consumers’
surplus for the infinite life of the innovation, sold at the competitive price, minus the
deadweight loss during the period of protection. Thus, W is increasing in s and decreasing
in T. Finally, suppose that for each R&D cost c, the distribution of ‘ideas’ is given by a
distribution function F with density f, where F(s|c) is the fraction of ideas with cost c
that have value less than s.
Then the social value of investment in ideas with cost c is Ẇ(T, c) defined below, where
s(T) is the minimum value that will elicit investment (Π(s(T), T) 5 c). That is,
W (T, c) 5 3
# `
[ W (s,T ) 2c ] f (s 0 c) ds
s (T )
W (T, c) 5 3 [ W (s,T ) 2c ] f (s 0 c) ds 2W (s (T ) ,T ) (T )
0 #
`
0T s (T )

W (T, c) 5 3
# `
[ W (s,T ) 2c ] f (s 0 c) dsEconomic models of innovation 127
s (T )
Notice that s’(T) , 0. A marginal increase in T will increase investment in amount
2W(s(T), T)s’(T). However, even though investment goes up with T, total social
welfare Ẇ(T, c) may go down. The change in social welfare is
0 #
`

0T s (T )
The last term represents the welfare due to new innovations called forth by longer protec-
tion, but the first term, which is negative, represents the loss in consumers’ surplus on all
the inframarginal innovations that would have been achieved even with shorter duration.
As T becomes large DD and s(T) becomes small, it is reasonable to think that the first term
3 5
becomes largeRrelative to the last term. Increasing the duration T beyond that point will
DC
not be in the social interest.
Of course the best length T must be established by adding up the marginal effects for all
c. Depending on the distributions of (s, c) in different product classes, the one-size-fits-all
nature of the patent system may provide excessive protection in some product classes and
deficient incentives in others.
Races for the intellectual property right introduce another inquiry as to how profitable
the intellectual property right should be, regardless of how the profit is achieved.1 Unlike
the Nordhaus argument, the inquiry leads to an argument for limited duration that applies
even if the profit is given as a prize out of general revenue and involves no deadweight
loss. The argument concerns the optimal amount of R&D effort. A more profitable right
will encourage more entry into the race (the extensive margin) or more collective effort as
each participant accelerates its effort (the intensive margin).
The potential benefits of inciting more effort by offering more profit depend on the
creative environment—the nature of the R&D process. Nordhaus implicitly addresses a
creative environment where ‘ideas are scarce’ so that duplication of costs is not the focus.
Suppose, however, that more than one potential innovator can serve the same market
niche. Then there is a second reason to limit duration. Not only will there be excessive
deadweight loss on inframarginal innovations, but the disparity between profit and cost
will also lead to duplication of R&D cost as firms vie for the very profitable rights.
Thus, part of the inquiry into the optimal strength (profitability) of an intellectual
property right concerns the extent to which additional effort is duplicative. This issue takes
us back to the question, what is the right model of the creative environment? If ‘ideas
are scarce,’ then races are not an issue. But if all investment opportunities are commonly
known, then races may or may not be efficient, depending on the ‘production function
for knowledge.’ If successes and failures in the R&D process are perfectly correlated,
then a race is duplicative. If successes and failures are independent, then a race increases
the probability of at least one success or, in another interpretation, accelerates progress
1
Innovation races are more suited to patentable subject matter than to expressive works.
Such races can only occur if several rivals are vying for a right that only one of them will receive.
Rights to expressive works are generally narrow enough in scope that several authors can obtain
protection for works that have some similarity (and hence can compete). Thus, an author may fear
a reduced market due to competition from another author but does not generally fear that he or she
will be wholly excluded from the market through a rival completing their work first.
M4754-Math Eqn.indd 1 30/05/

(Loury, 1979; Lee and Wilde, 1980; Reinganum, 1982, 1985, 1989). Further, if the creative
environment is one in which different firms have different unobservable ideas for how to
address a given need, then entrants to a race need not be the most efficient firms or those
with the best ideas (Scotchmer, 2004, ch. 2).
The number of entrants in a race may be too large or too small, as compared to the effi-
cient number, depending on the size of the private reward. Suppose, for example, that two
firms have different ideas about how to fill a market niche with value s. Suppose that each
firm’s cost is c, and that each has probability 1/2 of succeeding. Suppose that the value of
the property right will be Π(s, T), and that the firms’ prospects for success are independent.
If both firms succeed, each will receive the property right with probability 1/2.
Then a second firm will enter the patent race if Π(s, T)(3/8) . c, since its probability
of receiving the patent is 3/8. On the other hand, entry by the second firm is only efficient
if W(s, T)(3/4) 2 2c . W(s, T)/2 2 c or W(T, s)/4 . c. Thus, if Π(s, T)(3/8) . c .
W(s, T)/4, there will be excess entry to the patent race—the second firm will enter even
though that is not efficient—and if Π(s, T)(3/8) , c , W(s, T)/4 there will be too little entry.
Entry into a race may provide a private value to the entrant that is greater than the social
value of the entry and always provides a private value that is greater than the increment
to private value of both firms. The latter is because of the ‘business-stealing effect.’ The
second entrant’s expected profit is Π(s, T)(3/8) 2 c, while the increment to joint profit is
only Π(s, T)(1/4) 2 c. The second entrant’s chance of winning the race and getting the
patent comes partly at the first entrant’s expense. It is this externality that may lead to
excessive entry into a race. It also implies that if the reward were as large as the social value,
there would be too much entry. In fact, the only thing that is clear in this environment,
without imposing additional structure, is that the optimal reward is smaller than the social
value of the innovation. But this is not a very useful design principle because rewards
given as intellectual property will have that attribute almost inevitably.
Landes and Posner (2003, pp. 222–8) suggest that some works may be diminished by a
congestion externality. They illustrate their point by reference to the Disney Corporation’s
self-imposed restraint on commercialization:
To avoid overkill, Disney manages its character portfolio with care. It has hundreds of characters
on its books, many of them just waiting to be called out of retirement. . ..Disney practices good
husbandry of its characters and extends the life of its brands by not overexposing them. . . . They
avoid debasing the currency. (Britt, 1990)
Landes and Posner (2003) assert that this concern justifies perpetual protection for
some works. To balance the costs of protection, they advocate a system of indefinitely
renewable copyright protection, with the renewal fee acting as policy lever for diverting
works not subject to congestion externalities into the public domain. They note that a
similar over-saturation can arise with regard to some rights of publicity (use of persona
in advertising) and trademarks.
C. Breadth
The breadth or scope of an intellectual property right has critical bearing on its economic
value, and hence its incentive effect. A broader right preempts more substitutes than a
narrow right.

The scope of a utility patent is determined by the language of the claims, which define
the boundaries of literal infringement, and the extent to which such boundaries will be
stretched to cover similar, but not quite literal, embodiments. Under the ‘doctrine of
equivalents,’ courts will find infringement where the accused device ‘performs substan-
tially the same function in substantially the same way to obtain the same result’ (Graver
Tank & Mfg. Co. v. Linde Air Products Co., 339 U.S. 605, 608 (1950) (quoting Sanitary
Refrigerator Co. v. Winters, 280 U.S. 30, 42 (1929)); Warner-Jenkinson Co. v. Hilton Davis
Chem. Co., 520 U.S. 17 (1997)).
The scope of copyright is determined by the substantial similarity test in conjunction
with copyright’s limiting doctrines (e.g., originality, scènes à faire, non-protectability of
ideas and facts, fair use)—is the defendant’s work substantially similar to the protected
elements of the plaintiff’s work? In practice, a copyright is quite narrow. J.K. Rowling
can prevent others from copying Harry Potter novels, but she has no control over other
stories about wizards. It is unlikely that different authors operating from the same ideas
will produce substantially similar novels. Breadth issues can arise, however, with regard
to works built on copyrighted works, such as sequels and film adaptations. We deal with
these issues in the context of cumulative innovation. Breadth does not generally arise in
the context of trade secrets.
These legal tests do not map directly onto the economic concepts of breadth.
Economic models of breadth have been developed for two market contexts: where an
innovation is threatened by horizontal competition and where an innovation might be
supplanted by an improved innovation. We take up the latter question in the analysis of
cumulative innovation. For horizontal competition, breadth has been modeled in two
ways: in ‘product space,’ defining how ‘similar’ a product must be to infringe a patent,
and in ‘technology space,’ defining how costly it is to find a noninfringing substitute for
the protected market.
In the first notion of breadth, introduced by Klemperer (1990) using a spatial model,
the size of the market for the patented product depends on the closeness of noninfring-
ing substitutes. A broader patent covers more of the product space, meaning that more
substitutes infringe. The right to keep a substitute out of the market is profitable for the
patent holder in two ways: by shifting the demand for the protected good outward (where
the intellectual property owner excludes the substitute from the market) or by allowing
the intellectual property owner to charge higher prices for both the patented good and the
infringing substitute.
The second notion of breadth for horizontal substitutes, developed by Gallini (1992), is
that it determines the cost of entering the market. In this conception, the goods are exact
substitutes, and breadth implicitly refers to the technology of production (process innova-
tion) rather than closeness of substitutes in the market. Entry by a second firm does not
cause demand curves to shift, but instead causes the firms to compete in a given market.
A narrower utility patent (lower cost of entry) will lead to more entry and lower prices.
Entry stops when the cost of entry can no longer be covered by competing in the market.
In both conceptions of breadth, a narrower utility patent leads to lower per-period
profit. Thus, breadth might be conceived as a policy lever that governs profit, as described
above, in a one-size-fits-all system where the duration of protection cannot be tailored to
the cost of innovation. In the utility patent system, such tailoring is not generally done
in any systematic way by the Patent Office. Examiners focus solely on ensuring that the

application meets the threshold criteria and that the claims are clear. They do not adjust
the ‘breadth’ of claims (Menell, 2013, 2019a). The courts exercise a modest degree of
tailoring. In applying the doctrine of equivalents, courts accord ‘pioneering’ inventions
greater scope than more modest inventions. Such a rule increases the reward for major
breakthroughs. The copyright system does not systematically vary the scope of protection
with the cost or importance of the work.
Even within the one-size-fits-all system, there is a policy question as to whether, on
average, rights designed to give a pre-specified reward should be structured as long and
narrow or short and broad. The inquiry into how market rewards should be structured
has led in several papers to a ratio test: a policy reform is desirable if it increases the ratio
of profit to deadweight loss. The ratio test was devised by Kaplow (1984) in the patent/
antitrust context and was also used by Ayres and Klemperer (1999) in the enforcement
context. It reappears in the cited discussions of patent breadth. The basic notion is that
deadweight loss is the consumer cost of raising money through proprietary pricing. If the
ratio of profit to deadweight loss is higher, the money being raised through proprietary
pricing is raised more efficiently.
In a broad class of demand curves including linear ones, any price reduction from
the monopoly price will increase the profit-to-deadweight-loss ratio but will also reduce
profit, thus necessitating a compensation such as longer protection. This can be seen in
Figure 6.1, where the monopoly price is 1/2 and the lower price 1/3 is the duopoly price.
At the price 1/3, the ratio of profit to deadweight loss is the ratio of the light grey
areas (of size 2 × C) to the triangle D. At the monopoly price 1/2, the ratio of profit to
deadweight loss is the ratio of the outlined box that represents monopoly profit to the
triangle (B+M+D). One can see by inspection that the ratio of profit to deadweight loss
is smaller at the monopoly price 1/2 than at the lower price 1/3. In fact, with the linear
demand curve, this argument generalizes for any reduction in price: the lower the price,
the higher the ratio of profit to deadweight loss. This is the argument given by Tandon
(1982), arguing for compulsory licenses to lower prices, and Gilbert and Shapiro (1990),
Demand
q (P) = 1-p
1/2
A
B
1/3
C M
D
1/3 1/2 2/3 q
Figure 6.1 Monopoly pricing and deadweight loss

arguing for narrow patents, which they interpret as lower prices, although they do not say
how price reductions in a given market might flow from narrower scope.
How, though, do narrow patents lower the price in a given market? In Gallini’s (2002)
conception, breadth determines the cost that an imitator must pay to enter a proprietary
market. Entry is only tempting if the market will be protected long enough so that the
entrant, in competition with the patent holder, can still cover the cost of entry. If entry
occurs, competition between the entrant and the right holder will lower the price. Figure
6.1 can be used to compare a relatively short period of protection where entry by an
imitator is not tempting, with a longer period of protection, where entry is tempting
even though the imitator must pay a cost. With the shorter period of protection, say TM,
consumers will pay the monopoly price 1/2, but with the longer period of protection, say
TD > TM, they will pay the duopoly price 1/3. Suppose that TM and TD are chosen so that
the patent holder makes the same discounted profit in both regimes, and the cost of entry
is such that exactly one imitator will enter if the patent lasts for length TD. Then by the
above argument, consumers would be better off in the duopoly regime, despite the longer
period of protection, because of the lower price.
However, that argument does not account for the fact that the imitator must pay real
resource costs to enter the market. Gallini (1992) argues that the duplication of costs is
severe enough to overturn the above argument. Given that the price can only be reduced
by costly entry, it is better for society as a whole—including consumers, the patent holder,
and the imitator—to have a short period of monopoly pricing than a longer period that
attracts entry.
However, we have already stressed that the best design of intellectual property rights
depends importantly on what one assumes about licensing. In this case, licensing again
overturns the conclusion. Maurer and Scotchmer (2002) argue that the patent holder will
anticipate entry and offer a license instead of tolerating unlicensed entry. In this way, the
patent holder can increase his own profit without reducing the profit of the entrant, and
at the same time can eliminate the wasteful duplication. The narrow patent thus has the
effect of lowering price without imposing the social cost of duplication, and the above
welfare analysis is restored. The better policy is a narrow patent for a relatively long time.
D. Rights of Others (and Defenses)
The rights afforded others in protected works directly affect the profits from intellectual
property. Many of these rules—such as blocking rights (patent and copyright) and excep-
tions for experimental use (patent), fair use (copyright), and reverse engineering (copyright
and trade secret)—find their economic justification in the cumulativeness of innovation
and hence are covered in Section III. Doctrines relating to independent invention, prior
user rights, and ‘first sale’ (or exhaustion of rights) relate to stand-alone invention, as do
proposals about extending user rights to limited sharing of copyrighted works.
1. Rights arising from independent invention

A right of independent invention means that, provided the independent inventor was
actually an ‘inventor’ (and, in particular, did not learn the invention from any other party,
such as a prior inventor), he or she is free to practice the invention. Both copyright law and
trade secret law immunize independent inventors from liability, but patent law does not.

In the case of trade secrets, it would be impossible for an independent inventor to know
what had previously been invented. In the case of copyrights, which protect expression,
any re-expression escapes liability (broadly speaking). In the case of patent law, the right is
defined with respect to claims and (broadly speaking) not with respect to how a potential
infringer achieved the potentially infringing innovation.
Scholars have made three types of economic arguments about independent invention.
First, in the context of trade secrecy, the absence of an independent invention defense
would stifle innovation because inventors would be uncertain as to whether they could
practice the new knowledge they create.
Second, a right of independent invention can reduce the duplication of R&D costs in
patent races (La Manna et al., 1989; Blair and Cotter, 2002; Maurer and Scotchmer, 2002;
Leibovitz, 2002; Ottoz and Cugno, 2004). If the value of an exclusive right in the market is
$100, and the R&D cost is $20, five firms may enter a race. But if all five firms have rights
ex post, competition will reduce the private value of the right below $100, and fewer than
five firms will enter. The right of independent invention reduces the duplication of costs
and at the same time affords lower prices to users, all without undermining the incentive
to invent.
Landes and Posner (2003, pp. 361–2) make a similar argument for trade secrets. They
compare the American rule, under which the owner of a trade secret loses his right
to the invention if someone else patents it (W.L. Gore & Assocs. v. Garlock, Inc., 721
F.2d 1540 (Fed. Cir. 1983)), to the prior user right that prevails in some other nations.
The prior user right divides the entitlement, enabling multiple independent inventors
to share its value through an effective oligopoly structure. As in the foregoing argu-
ment, duplicative entry will only occur to the extent that all firms cover their costs.
Third, giving rights to independent inventors can induce patent holders to license ex
post on terms that reduce market price, in order to discourage ex post entry through
independent invention (Maurer and Scotchmer, 2002). Suppose that a single patent
holder is in the market. Then, whether or not the patent holder licenses, a right of
independent invention will reduce the price in the market below the proprietary price.
Without licensing, the price will fall due to entry by independent inventors. Instead,
the patent holder can license at a fee equal to the cost of independent invention. Then
independent inventors are indifferent to paying the license fee or paying the costs of
independent invention, but the patent holder prefers licensing. The price reduction in
the market (determined by the terms of license and number of licensees) must be large
enough to deter further entry.
The market price with licensing will thus depend on the cost of independent invention.
If the cost is high enough, the right of independent invention can benefit users without
undermining the incentive to invent. In fact, in plausible models, the cost of independent
invention only needs to be greater than half the cost of the original innovator (Maurer and
Scotchmer, 2002; Ottoz and Cugno, 2004). Nevertheless, Blair and Cotter (2002) rightly
point out that the economic consequences depend critically on the relative costs of first
inventors and imitators, which will differ across technologies. Giving a right of independ-
ent invention can have harmful consequences on innovation if imitation or independent
invention is too cheap.
Lichtman (1997) makes a similar argument in the context of unpatented inventions,
advocating on grounds of cost that independent inventors be allowed to copy but not

clone them. Armond (2003) proposes that independent discovery be available as a defense
to a preliminary injunction motion.
Although independent inventors are not generally exempted from liability under US
patent law, US law provides a limited prior user right where the technology was ‘commer-
cially used . . . in the United States, either in connection with an internal commercial use
or an actual arm’s length sale or other arm’s length commercial transfer of a useful end
result of such commercial use’ and ‘such commercial use occurred at least 1 year before
the earlier of either—(A) the effective filing date of the claimed invention; or (B) the date
on which the claimed invention was disclosed to the public in a manner that qualified for
the [grace period]’ (35 U.S.C. § 273(a)).
In addition, state employment law provides an employer with a royalty-free, non
exclusive, non-transferable license (‘shop right’) to use an employee’s invention where
the employee makes a patented invention using the employer’s facilities. In most research
environments today, employers require employees involved in research-related activities to
assign their inventions to the employer, although some state laws limit such agreements to
inventions developed within the scope of employment or developed using the employer’s
facilities (e.g., Cal. Labor Code § 2870). Even where no express agreement has been signed
by an employee, patents invented by the employee may nonetheless be deemed to have
been assigned where an employee has specifically been employed to invent in the field in
which the invention was made. In these circumstances, a court may imply an assignment
clause into the employment contract.
2. Rights after sale

Under what is commonly referred to as the ‘first sale’ or ‘exhaustion’ doctrine, the intel-
lectual property owner ‘exhausts’ the legal monopoly in a product by selling it to the
public, thereby enabling the purchaser to use the work and resell it without infringing.
Such a default right structure reduces transaction costs for subsequent transactions.
Similarly, purchasers of patented products are deemed to have an implied license to make
repairs, although this license does not extend to ‘reconstruction’ of the patented product.
Intellectual property owners can, subject to anti-competitive restrictions, circumvent the
first sale doctrine by imposing licensing restrictions upon the conveyance of a product.
3. Rights to share copyrighted works

Even though the purpose of copyright law is to prevent copying, a controversial idea that
keeps resurfacing is that copying or sharing is less harmful to creators than meets the eye,
at least where the sharing of each legitimate copy is limited. Where it is unlimited, such
as in peer-to-peer networks or when users make copies of copies, sharing poses a greater
threat to appropriability.
The reason that sharing might be less harmful is that the proprietor will price in a
way that anticipates sharing. Sharing allows the proprietor to charge a higher price,
since demand is determined by the willingness to pay several parties. Limitations on
sharing may arise because copies of copies degrade (Liebowitz, 1982, 1985; Liebowitz
and Margolis, 1982 (arguing in the era of analog copies)) or because it is less costly
to facilitate sharing than to produce a copy for every user, as in a video rental market
(Varian, 2000), or because the probability of detection increases with the size of the
sharing group.

The earlier set of papers in this vein relied on the fact that copying is costly. Novos and
Waldman (1984) and Johnson (1985) argue that proprietors may reduce price to avoid
copying, but the cost of copying will nevertheless preserve the proprietor’s market. The
market price will be lower than without copying, reducing the deadweight loss of exclud-
ing users, but the per-period reward to creative works will also be reduced, especially when
there is heterogeneity in tastes as well as in copying costs. The welfare effects are different
according to whether the cost of copying is per copy or per user, as when it requires the
purchase of a copying device. Scholars have also argued that copying can have an affirma-
tive benefit for right holders because it builds network effects (Conner and Rumelt, 1991;
Shy and Thisse, 1999).
A second set of papers focus on the fact that prices can be tailored to the groups
that form. Liebowitz (1985) emphasizes price discrimination according to whether the
purchaser will make the copy available to many users, as libraries do (see also Ordover
and Willig, 1978). Bakos et al. (1999) argue that, depending on the groups, sharing might
actually be more profitable than selling to individual users. This is true if, first, the willing-
nesses to pay within groups are negatively correlated or, second, if there is variance in the
sizes of groups. Thus, whether sharing enhances profit depends on what governs group
formation. However, Scotchmer (2005) argued that sharing groups will not be formed
exogenously or even randomly, and if they form in a way that is efficient for the group
members conditional on the proprietors’ prices, then group formation has no effect at all
on profit opportunities. Sharing is neither profit-reducing nor profit-enhancing.
Given that copying can have salutary effects as well as deleterious effects, these
arguments have led authors to consider an additional set of policy levers specific to the
copying context. These include taxes and subsidies on prices of legitimate copies, and
taxes and subsidies on copying devices, as well as the optimal mix of enforcement activi-
ties and other incentives (Besen and Kirby, 1989a, 1989b; Chen and Png, 2003; Netanel,
2003; Fisher, 2004).
E. Remedies
As in other bodies of law, the remedial opportunities in intellectual property law are
injunctions and damages. There are two branches of thought about the relative efficacy
of these rules, one branch focusing on whether remedies will lead to efficient use of the
property ex post and the other branch focusing on the ex ante effects.
The first set of arguments (Calabresi and Melamed, 1972; Polinsky, 1980; Kaplow and
Shavell, 1996) for the general framework, and Blair and Cotter (1998) for the intellectual
property context, considered whether property rules (injunctions) are more or less likely
than judicially imposed liability to encourage bargaining to an efficient outcome ex post.
For example, property rules (injunctions) may be preferred when transaction costs of
exchange are low and the costs of valuing violations of rights by courts are high. For
the intellectual property context, Ayres and Klemperer (1999) add the consideration
that ‘soft’ remedies, which do not actually restore the proprietary price, can be socially
beneficial because they increase consumers’ surplus without impinging much on profit,
at least for small price reductions.
The second set of arguments are not concerned with what would happen in the out
of equilibrium event of infringement but focus instead on how potential remedies affect

equilibrium profits and the ex ante incentives for R&D (Schankerman and Scotchmer,
2001; Anton and Yao, 2007). In these arguments, remedies are only important because
they do or do not deter infringement, and because they determine the terms of an ex ante
license. The terms of license that will be accepted by a potential licensee/infringer depend
on the consequences of infringement, and this threat has an effect on the ex ante division
of profit. Schankerman and Scotchmer (2001) argue that if infringement leads to profit-
eroding competition between the infringer and right holder, a wide range of remedies will
deter infringement, at least for stand-alone innovations, and are therefore equivalent from
an ex ante point of view. However, this is not necessarily true for research tools and other
potentially licensed intellectual property where infringement does not dissipate profit.
Merges and Duffy (2002) and Blair and Cotter (1998) argue, again from the ex post
perspective, that patent and copyright law are better suited to a property-rule paradigm
than a liability-rule paradigm. Since intellectual property rights are relatively well defined,
disputants or potential disputants should have little trouble resolving their differences
by negotiating licenses against the backdrop of an injunction. In contrast, if the setting
of damages (an ex post ‘compulsory license’) is left to a generalist judicial institution
under a liability rule, the court may have difficulty placing a value on the intellectual
property or on the injuries caused by infringement. Further, judicially imposed licenses
can undermine the prospect function of patent law (Kitch, 1977). Merges (1996) argues
that for complex transactions involving many players, a property rule will facilitate the
creation of private exchange institutions, such as patent pools, that can evolve in response
to changing circumstances and draw upon industry and institutional expertise.
Although infringed right holders generally have a right to enjoin unauthorized use,
injunctions are backed up by compensatory damages for past violations. These may
include enhanced (punitive) damages for patent infringement, statutory damages for
copyright infringement, and attorney fees and costs in ‘exceptional cases.’ In several
areas such as the covers of musical compositions, juke boxes, cable television broadcasts,
satellite retransmission of television signals, and webcasting, copyright law provides for
compulsory licensing. These regimes arguably economize on transaction costs, although
commentators are divided on the economic effects of such compulsory licenses (compare
Merges (1996, 2004b) with Netanel (2003)). Consistent with traditional economic analysis
of damages (harm internalization), patent and copyright law award intellectual property
owners the greater of lost profits or a reasonable royalty for the defendant’s unauthorized
use of the protected works (Blair and Cotter, 1998). Calculating these measures, however,
is quite complex, involving numerous subtle determinations of how markets would have
evolved had infringement not taken place (Graham et al., 2017).
It follows from general economic principles that enhanced damages should be awarded
where improper behavior is costly to detect and where full compensatory damages are
costly to prove (Polinsky and Shavell, 1997; Cooter, 1982; Cotter, 2004). Excessive dam-
ages (i.e., where expected damages exceed actual damages) could lead to overdeterrence
in the sense that parties may exercise caution in order to avoid a risk of liability. Several
studies indicate that courts may well be overdeterring patent infringement based on the
high rate of enhanced damages awards (Lemley and Shapiro, 2007; U.S. Federal Trade
Commission, 2003; Cotter, 2004; Moore, 2000).
Due to the very different nature of trade secret protection, the remedies available for
unauthorized use and dissemination of a trade secret are more limited. Where the secret

has not been disclosed to the public, courts will generally enjoin further use of the secret
by a misappropriating entity. But where the secret has been disclosed, the trade secret
owner will be limited to damage remedies or limited injunctive relief against the entity
that misappropriated the trade secret (such as a ‘head start’ injunction which excludes
the misappropriator from the market for a designated period). Disclosure to the public
destroys the secret and therefore it would be inappropriate (and infeasible) to enjoin use
of the information by others. Nonetheless, Sidak (2008) argues an injunction against a
misappropriating entity should be perpetual in order to encourage efficient post-litigation
bargaining over the value of continued use. Such a rule would also avoid the expense and
difficulty (error costs) of having courts adjudicate the option value of a trade secret.
The various modes of intellectual property provide overlapping protection. For example,
utility patent, copyright, and trade secret law all cover computer software. As noted
earlier, however, copyright doctrines exclude ideas, processes, and methods of operation.
Thus, software developers cannot gain copyright’s long duration of protection for the
functional aspects of computer software. Such inventions must comply with utility patent
law’s formal examination requirements and surpass utility patent law’s higher thresholds
in order to obtain legal protection. In this way, intellectual property law prevents inventors
from obtaining protection for functional features through the ‘backdoor’ of the copyright
system.
The relationship between patent and trade secret law is somewhat more complicated.
Both regimes cover technological innovation. Where an inventor obtains a patent before
a subsequent researcher invents the same technology, the patent trumps the subsequent
inventor, regardless of whether the subsequent inventor seeks to protect the invention as
a trade secret. (Such secrecy may well conceal infringement, particularly in the case of
process inventions, but that does not suggest that the trade secret would have any validity
vis-á-vis the patent.)
A somewhat more complicated issue arises where the first inventor chooses to protect a
particular technology as a trade secret. If a subsequent inventor independently discovers
the same invention and obtains a patent, two overlapping issues arise: (1) does the trade
secret invalidate the patent (on novelty grounds); and (2) if the patent is valid, can the
trade secret owner continue to practice the invention—in essence, does the first inventor
enjoy a prior user right. As suggested earlier, the trade secret will not invalidate the patent
because it does not fall within the body of prior art that may be considered in judging
novelty. Therefore, assuming the second inventor meets the other requirements of patent-
ability, she will obtain a valid patent. As regards the rights of a trade secret owner, as
noted previously, US patent law provides a limited prior user right (35 U.S.C. § 273(a)).
The prior user right in this circumstance places the technology under duopoly rather than
monopoly control. The profits available to the patentee are reduced accordingly. It can
be argued, however, that such a structure of rights might partially improve the screening
function of utility patent law—inventions that have been independently developed may
not have needed as much of an ex ante incentive in the first place. To the extent that ex ante
incentives are more than sufficient to generate the innovation, duopoly improves social
welfare by reducing the deadweight loss from exclusive exploitation.

III. CUMULATIVE INNOVATION
In the context of stand-alone inventions or creations, intellectual property rewards reflect

the social value of the contribution, since the profit is determined by demand. That is
one of the main virtues of intellectual property as an incentive system. However, when
innovation is cumulative, the most important social benefit of an innovation may be the
boost given to later innovators, and this may make the benefits harder to appropriate
(Scotchmer, 1991). Moreover, the innovation may enable rivals to enter with improved
products. In that case, social success may mean private failure—the boost given to
the rivals may cause the innovation’s own demise (Scotchmer, 1991, 1996; Green and
Scotchmer, 1995; Chang, 1995; O’Donoghue et al., 1998; O’Donoghue, 1998; Bessen and
Maskin, 2009; Hunt, 1999, 2004). Merges and Nelson (1990) give an opposite perspective
on the cumulative problem. Instead of worrying that later improvers pose a threat to
earlier innovators, they worry that earlier innovators (earlier patents) pose a threat to
later improvers. Galasso and Schankerman (2015) find that patent rights block follow-on
innovation in a few specific technology fields (computers, electronics, and medical instru-
ments), but not in drugs, chemicals, or mechanical technologies. They caution, however,
against concluding that removing patent protection would be beneficial due to adverse
incentive effects on upstream innovation (Green and Scotchmer, 1995), the benefits of
signaling (Conti et al., 2013), and a shift toward trade secrecy (and less disclosure).
The intellectual property system must resolve these contradictions. In general, the
problem of appropriating benefits has two facets: the overall level of profit and how
it is divided among the sequential innovators. The roles of the policy levers are closely
intertwined in the cumulative context, and the best design of the system will depend on
the transaction costs of licensing.
Many scholars have emphasized the importance of cumulativeness in the process of
knowledge creation, especially economic historians. As expressed by David (1985, p. 20),
‘Technologies . . . undergo . . . a gradual, evolutionary development which is intimately
bound up with the course of their diffusion.’ Secondary inventions—including essential
design improvements, refinements, and adaptations to a variety of uses—are often as
crucial to the generation of social benefits as the initial discovery (Nelson and Winter,
1982; Taylor and Silberston, 1973; Mak and Walton, 1972; Rosenberg, 1972). Cumulative
technologies tend to involve multiple components, serve as building blocks for further
incremental innovation, and often spur wide-ranging applications. Automobiles, aircraft,
electric light systems, semiconductors, and computers fall within this category. Some
chemical technologies are hybrids of discrete and cumulative models. New chemical
compounds are typically discrete in terms of the product market that they serve, but they
can suggest promising new lines of research (e.g., penicillin, Teflon) (Nelson and Winter,
1982). The biotechnology field reflects several cumulative features. The development of
research tools provides the means for decoding genomic information. Research decoding
genomes provides the input for downstream biomedical research.
Cumulativeness also extends to expressive creativity. All authors and artists draw, to
some extent, on prior works. Sequels, translations, and screenplays build directly upon
prior works. Parodies and satires comment on or employ other works. Most musical
compositions reflect rhythm and other elements of established genres. The hip-hop and
rap musical genres embody prior recordings through the use of digital sampling.

A. A Preliminary Model: The Virtues of Licensing
One of the lessons that emerges powerfully below is that the best design of the intellectual
property system depends on the fluidity of the market for licenses. Before turning to a
more detailed analysis of design issues and how licensing affects them, we illustrate the
importance of licensing by modifying the ‘reduced form’ model of Landes and Posner
(2003), where a variable z is taken as the ‘strength’ of a right. For example, the strength
of the right may be affected by breadth or exemptions such as fair use.
Let q(·) be the demand function for a protected innovation, where q(p) is decreasing
with p. Referring back to our discussion of copying, and how the threat of copying affects
the market price, let y(p, z) be the supply of illicit copies. Then the net demand faced
by the proprietor is q(p) 2 y(p, z). Assuming for convenience that the marginal cost of
copies is zero, the proprietor maximizes p[q(p) 2 y(p, z)], and sets a profit-maximizing
price p*(z) that depends on the strength of protection through the threat of copying.
Now consider how the profit-maximizing price and the proprietor’s profit depend on
the strength of protection, z. Assume that the supply y(p, z) of illicit copies increases
with p and decreases with z. Because the supply of imitations y(p*(z), z) depends on the
proprietor’s price as well as the level of protection, there is a potential indeterminacy in
the model. An increase in protection could conceivably lead to a decrease in price and an
increase in illicit copying. However, under reasonable assumptions we can assume that the
profit-maximizing price increases with the level of protection, even though the increase in
price has a feedback effect of increasing imitation or copying.
Nevertheless, the profitability of creations and hence their supply will not necessarily
increase with the level of protection z. This is because the cost of creation can also depend
on z, for example, by creating a burden to innovate outside the scope of other rights.
Suppose, in fact, that each potential innovator faces an R&D cost k plus additional costs
e(z) that reflect the burdens imposed by other intellectual property rights. The creation
is profitable if
p*(z)[q(p*(z)) 2 y(p*(z), z)] 2 k 2 e(z) . 0
If potential creations differ in their markets (e.g., if we introduce a quality variable s into
the demand function q), then the more valuable ideas will call forth investment, while the
less valuable ones will not. Without formalizing this idea, we will let N(z|k) describe the
number of profitable creations with cost k. The supply of new creations N(z|k) is not
monotonic in the strength of protection z because raising z increases the creator’s costs.
The punch line of this model is that too much protection can be bad for creators as well
as for imitators.
The point we would like to make, however, is that the punch line is largely reversed if
firms can license to avoid conflicting property rights, rather than being forced into the
costly activity of avoiding them. Following the perspective in O’Donoghue, et al. (1998),
assume that each innovating firm will initially be in the position of paying license fees on
the discoveries of its predecessors and then in the position of collecting license fees from
its followers. The effect of strong rights, z, is to increase the licensing obligations, rather
than to increase the real resource cost of avoiding prior rights. The essential point is that
licensing also creates claims against future innovators.

Suppose, in fact, that all innovators are in symmetric positions: they pay for the same
number of licenses and are paid by the same number of licensees. Then, since all the
money must go somewhere in the end, symmetry means each innovator pays as much in
licensing fees as it earns in licensing fees, say, l(z) on both sides of the ledger. The profit-
maximizing decision is then to invest in the potential creation if
p*(z)[q(p*(z)) 2 y(p*(z), z)] 1 l(z) 2 k 2 l(z) . 0
or equivalently,
p*(z)[q(p*(z)) 2 y(p*(z), z)] 2 k . 0
Hence (assuming that the revenue p*(z)[q(p*(z)) 2 y(p*(z), z)] is increasing in z),
the ambiguous effect of strong protection on innovation disappears. A strengthening of
protection leads to more creations.
Thus, with licensing, we are cast back to the same consideration as with stand-alone
inventions, namely that there is a tradeoff between deadweight loss and innovation but
no tension between protecting early innovators and protecting the later innovators who
use the knowledge they create. Licensing will largely resolve that tension, to everyone’s
benefit.
These points about the salutary effects of licensing have mostly been made in models
that distinguish between the policy levers of intellectual property, rather than lumping the
policy levers into a single variable called the ‘strength’ of the right. We now turn to that
more disaggregated discussion.
B. Duration
Although the length of protection has an obvious effect on the overall level of profit,
the statutory length may be irrelevant. When innovations are under threat of being
supplanted by improved innovations, market incumbency only lasts until that happens.
O’Donoghue et al. (1998) define a notion of ‘effective patent life’ that focuses on the rate
of market turnover and argue that the effective life of the patent may be determined by the
breadth of the right, rather than its statutory length. This is because breadth determines
how long it will take before the product is supplanted.
In fact, there is considerable evidence that the ‘effective’ lives of most patents are
shorter than their statutory lives. Mansfield (1986) reports, using survey evidence, that
in some industries 60 percent of patents are effectively terminated within four years. The
literature on patent renewals carries a similar message (Pakes and Schankerman, 1984;
Schankerman and Pakes, 1986; Pakes, 1986; Schankerman, 1998; Lanjouw, 1998). For
example, Schankerman (1998) reports that half of patents in France are not renewed
beyond the tenth year, even though renewal fees are very low.
Bessen and Maskin (2009) present a model of sequentialness where this endogeneity of
patent life is absent (because the products do not compete in the market) but argue that
statutory life should be shorter when innovators learn from previous innovators. Their
model has sequentialness in innovation (because innovators learn from each other) but
the resulting products are ‘stand-alone’ and live out their statutory lives. They argue that

the optimal statutory life should be shorter if innovators learn from each other than if
not, because the loss from impeding future innovation is greater. (This result depends
sensitively on the absence of licensing; see below for an argument that all results in this
arena turn on what is assumed about licensing.)
For the case of basic and applied research, Green and Scotchmer (1995) argue that
patent lives must last longer if the research is divided between sequential innovators rather
than concentrated in a single firm, because of the problem of dividing profit in a way that
respects the costs of both parties.
To the extent that transaction costs may impede licensing and first stage inventors do
not need large ex ante rewards to induce innovation, a shorter duration of intellectual
property protection promotes cumulative innovation. Legal protection for computer
software fits this profile (Menell, 1987). There are relatively strong non-intellectual
property incentives for developing operating system and other platform technology for
many product markets. Interoperability with widely adopted platforms is often critical
to secondary innovation, such as application programs and peripheral devices. Owners
of the intellectual property rights in widely adopted proprietary platform technologies
can exercise tremendous market power due to network effects and consumer lock-in.
Shortening the duration of protection for such technologies is one mechanism for
constraining such market power and better equilibrating the incentives of first- and
second-generation innovators.
C. Threshold Requirements and Breadth
The status of an invention that builds on other inventions can be: (1) protected and non-
infringing, (2) unprotected and noninfringing, (3) protected and infringing, or (4) unpro-
tected and infringing (Scotchmer, 1996; Denicolò, 2002a). Thus, the economic effects
of threshold (patentability) and breadth are hard to disentangle. Scenario (1) gives the
best incentive for second-generation innovators but does not force the second-generation
innovator to share the profit with the prior innovator. Scenario (2) will clearly stymie
second-generation innovation unless there is a mechanism other than intellectual property
to protect the innovator. In scenario (3)—which is possible under patent law—the works
are considered ‘blocking’: the later work infringes the prior innovation and cannot be
exploited without a license, and the later work is protected and cannot be exploited by the
pioneering inventor without a license. Such a scenario encourages the inventors to share
profit from the subsequent invention. Scenario (4), which approximates the treatment
of derivative works under copyright law, discourages improvements or adaptations by
subsequent inventors in the absence of ex ante bargaining. However, there is less differ-
ence between scenario (3) and scheme (4) than meets the eye. Even if the later product is
not protectable, it can be protected by an exclusive license on the innovation it infringes.
The literature draws widely differing conclusions about the optimal way to organize
the rights of sequential innovations, largely because authors make different assumptions
about when and whether licenses will be made and who can be party to the negotiation.
Kitch (1977) was the earliest, and perhaps most extreme, licensing optimist.
Licenses can be made either ex ante, before the follow-on innovator invests in his
project, or ex post. If licenses must be negotiated ex post, after both innovations have
been achieved, scenario (4) may stifle innovation, since the second innovator will fear that

the first innovator will simply appropriate it (Green and Scotchmer, 1995). On the other
hand, if the second innovator can approach the first innovator for an ex ante license before
investing in his idea, the second innovation is not in jeopardy under either of scenario (3)
or (4). However, the first innovator will typically collect more of the profit in scheme (4)
because the second innovator will have less bargaining power (Scotchmer, 1996).
Denicolò (2002) considers a model where ideas are common knowledge, and asks how
the various scenarios affect patent races, assuming that there will be ex post licenses but
no ex ante licenses. He finds that the choice should depend on the relative costs of the
innovators. If, for example, the cost of the first innovation is low and the cost of the
second is relatively high, it may be better not to let the first innovator share in the second
innovator’s profit. Of course, this also depends on whether the first innovation can earn
profit in the market or only through licensing. Chang (1995) also considers a context
where the firms could make ex post licenses, but not ex ante licenses, and concludes that
the choice between (1) and (3) should depend on the relative costs of the innovators.
The worst situation is when licensing may fail entirely, and when, in any case, the earlier
innovator does not need to profit from licensing in order to cover his costs. Merges and
Nelson (1990) draw on many actual examples to argue that such circumstances are quite
plausible. A defect of the one-size-fits-all intellectual property regime is that it cannot
distinguish cases where blocking rights are unnecessary for cost recovery from cases where
earlier innovators would not invest unless they can profit from licensing. In part on this
basis, Burk and Lemley (2002) argue that it would be better to make intellectual property
protection more finely attuned to industrial contexts.
But even in those cases where earlier innovators should be allowed to profit from the
later innovations they enable, blocking rights are a blunt instrument for dividing profit.
Profit shares will not necessarily reflect the cost shares. This is especially true if the licenses
will be negotiated ex post, after all innovators’ costs are sunk (Green and Scotchmer,
1995), although Chang (1995) argues that the problem can be mitigated by making the
infringement (breadth) responsive to cost. In copyright law, blocking rights are not
available as a tool at all. Follow-on creators may not prepare infringing derivative works
without permission of the owner of the copyright in the underlying work. Copyright law
therefore has less flexibility than patent law in how it balances the incentives for sequential
innovators (Lemley, 1997).
In the case of product improvements, there is a question of whether a patent’s breadth
can extend to slightly better products that have not yet been invented or only to inferior
products. O’Donoghue et al. (1998) define a notion of ‘leading breadth’ that determines
when there will be blocking rights in the cumulative context and also establishes the
‘effective life’ of the patent right. To see why leading breadth is useful as a policy lever,
suppose to the contrary that every trivial infringement is noninfringing, even if patent-
able. A potential improver may discard ideas for small improvements because they lead
to price-eroding competition between close, vertical substitutes. It is only the relatively
big ideas that will become innovations. This problem can be solved by making the small
improvements infringing. Firms may then be willing to invest in them, since control of
the improvement and its predecessor can then be consolidated through licensing in the
same firm. Instead of competing, both will be marketed together. Further, if the small
improvements are infringing, and if it takes a relatively long time for large ideas to come
along, the ‘effective life’ of each patent is prolonged. These effects cannot be achieved

by choosing the patentability standard alone; the opportunity to consolidate successive

improvements in the hands of a single firm arises because the patents are infringing.
In the ‘ideas’ model, there is not much role for a nuanced patentability standard.
However, in a ‘production function’ model, a patentability standard can make each suc-
cessive innovator more ambitious in the size of improvement he invests in. This may be
socially beneficial (O’Donoghue, 1998).
Hunt (1999, 2004) presents a production function model motivated by the Semiconductor
Chip Protection Act, in which eligibility for protection coincides with noninfringement
of previous innovations (as with copyright). He argues that the standard for protection
should be increasing in the ‘dynamicness’ of the industry. Since noninfringement coincides
with protection, it is hard to sort out their respective roles. Whereas the effective patent
life is determined by breadth in the model of O’Donoghue et al. (1998), it is determined
in the Hunt model by both.
Finally, we return to the idea of Kitch (1977), who argues that strong patent rights
should be given to pioneers so that they can coordinate the subsequent development of
the technology. These are called ‘prospect’ patents. The theory is not focused on the reward
purpose of the patent, since it would apply even if the pioneer innovation were costless
to achieve and no incentive for R&D were required. The theory thus rejects the line of
reasoning that says intellectual property is at best a necessary evil, due to deadweight loss.
The prospecting theory rests on the premise that social interests and the private interests
of the patent holder are aligned. Scotchmer (2004, sec. 5.6) shows that this may be true in
some ways, but is not true in other ways, and, in particular, that strong pioneer patents can
preempt competition policy. As in later theories of cumulativeness, the prospector’s profit
comes from getting the intellectual property into use. For this reason, the pioneer has an
incentive to encourage use at a fee. Further, the pioneer can profit from delegating research
effort to the most productive researchers and avoiding bad projects, as are socially efficient.2
These are ways in which the pioneer’s interest is aligned with the broader public interest.
However, the pioneer can preempt competition policy in two ways: by avoiding compe-
tition in the ‘innovation market’ for second-generation products and by avoiding competi-
tion among second-generation innovators once the second-generation innovations exist.
The first of these may or may not be socially harmful depending on the patent/antitrust
interplay, but the second is clearly harmful to competition, assuming that patent law is
well designed in the first place. If these harms to competition are important, it might be
better to avoid pioneer patents even if second-generation innovators must then duplicate
the cost of achieving the pioneer innovation. As with all intellectual property, the case for
pioneer patents is strongest if the pioneer innovation is costly to achieve and the patent
is actually needed as a reward.
For radical improvements that read on existing patents, Merges (1994) suggests that it
may be socially advantageous to exempt an improver from infringement under the ‘reverse
doctrine of equivalents.’ Under this doctrine, a radical improvement may be deemed
noninfringing even though it literally reads upon an existing patent. The doctrine would
allow the radical improver to avoid a holdup by the underlying patent holder and to avoid
2
In fact, this is always true if the prospector can make deals with consumers as well as follow-
on inventors, for example, if consumers can pay the prospector not to retard progress.

a potential bargaining breakdown (see also Lemley, 1997). This is again an argument that
relies on difficulties in licensing. Such a doctrine avoids over-rewarding a first-generation
inventor (who did not foresee a much greater advance) while providing strong encourage-
ment to visionary subsequent inventors. The rule has rarely been applied in actual cases,
although the possibility of its application may well have fostered licensing.
1. Evolving doctrines regarding subject matter

We turn now to how these economic arguments have been reflected in legal doctrines.
Notwithstanding the broad scope of utility patents set forth in the Patent Act (‘any new
and useful process, machine, manufacture, or composition of matter, or any new and
useful improvement thereof’), the courts have barred inventors from claiming patents
on natural physical phenomena (e.g., the properties of lightning), laws of nature (e.g.,
the theory of general relativity), mental processes, and abstract intellectual concepts
(e.g., algorithms). Courts have noted that allowing exclusive rights for such fundamental
discoveries would unduly impede future inventors—a cumulative innovation rationale
(O’Reilly v. Morse, 56 U.S. (15 How.) 62 (1854)). Implicit in this justification is the notion
that transaction costs could impede licensing. The courts have thus realized, as is more
explicit in the economic models discussed above, that licensing plays an important role in
balancing the rights of sequential innovators.
Heller and Eisenberg (1998) argue that patenting of gene sequences generates a tragedy
of the anticommons, a fragmentation of rights which vastly increases transaction costs,
thereby impeding downstream research for medical advances. This is also, at root, an
argument about the ease of licensing. Walsh et al. (2003) report survey research indicating
that university research has not been impeded by concerns about patents on research tools
as a result of licenses, inventing around patents, infringement (often informally invoking a
research exemption), developing and using public tools, and challenging patents in court.
The opening of the ‘business method’ patent floodgates raised concerns about whether
patent protection is needed at all to promote such innovation and, more troubling,
whether such protection is chilling innovation and competition. The idea of protecting
business plans runs counter to a core premise of the free market system by offering a
form of antitrust immunity for business models. As the rising tide of prior art raises
the threshold for protection, such adverse effects may abate and innovation could well
produce valuable new business methods.
The Supreme Court has sought to restrain overbroad patent protection through a
series of decisions tightening eligibility standards. The Court’s 2012 decision in Mayo
Collaborative Services v. Prometheus Laboratories, 566 U.S. 66 (2012), imposed a new
limitation on the scope of the patent system: that a useful application of a scientific
discovery is ineligible for patent protection unless the inventor also claims an ‘inventive’
application of the discovery. The following year, the Court ruled that discoveries of the
location and sequence of DNA compositions that are useful in diagnosing diseases are
ineligible for patent protection (Ass’n for Molecular Pathology v. Myriad Genetics, 569
U.S. 576 (2013)). And in its 2014 Alice Corp. v. CLS Bank International, 573 U.S. 208,
decision, the Court ruled that software-related claims are ineligible for patent protection
unless the abstract ideas or mathematical formulas disclosed are inventively applied. The
Court explained these interpretations on concern about impinging upon (‘preempting’)
cumulative innovation (Mayo, 566 U.S. at 71).

These decisions, however, conflate eligibility with other patentability requirements. An

eligibility requirement of ‘inventive application’ eligibility threshold substantially overlaps
the Patent Act’s Section 103 nonobviousness standard. In fact, it arguably is broader in
that the Court’s articulation of ‘inventive application’ excludes consideration of scientific
discoveries and algorithms, whereas Section 103 looks at the claimed invention as a whole.
The Patent Act’s Section 112 provides the tools for ensuring that patents do not extend
beyond their proper scope. That an inventor’s claim might practically preempt all use
of a discovery will ‘show more clearly the great importance of his discovery, but it will
not invalidate his patent’ (Dolbear v. Am. Bell Tel. Co., 126 U.S. 1, 535 (1888)). Thus,
the Supreme Court’s recent eligibility jurisprudence raises serious concerns for medical
diagnostic research and other innovation fields that build directly on scientific discoveries
(Lefstin et al., 2018). These decisions directly confront the tension between promoting
pioneering and follow-on innovation.
Copyright law applies a different rights structure than patent law, with significant
implication for the direction of cumulative innovation (Lemley, 1997). Unlike utility
patent law, which allows anyone to patent improvements to patented technology, copy-
right owners have the exclusive right to prepare derivative works. Therefore, a novelist can
prevent others from translating their work into another language, adapting the story for
the stage or a motion picture, selling the story in narrated form (e.g., books on tape), and
developing sequels that draw extensively on the protectable elements (including possibly
character names and attributes—e.g., Rocky IV or the next James Bond film). As we will
see below, some borrowing is tolerated under the fair use doctrine and various exemptions
and compulsory licenses, but pioneers generally have exclusive authority to pursue the
further development of their expressive work. Copyright law wholly excludes protection
for ideas and functional attributes of a work but protects creators against direct or near
exact copying of even a significant fragment of the whole for a tremendously long dura-
tion (life of the author plus 70 years), reflecting the notion that society prefers to have 100
different war novels embodying similar themes, ideas, and facts than 100 versions of War
and Peace that differ only in their final chapter. Consequently, copyright protection for an
author’s expression of ideas and the relatively long period of its duration effectuates a dif-
ferent balance than patent law. Patent law encourages cumulative innovation by allowing
follow-on to secure rights on improvements and to enable any competitor to build upon
the innovation in its entirety within a comparatively short period of time (20 years from
the time of the application). By contrast, copyright law, with its narrow scope of protec-
tion, allows subsequent creators to pursue competing works using the same ideas as the
‘pioneer,’ but copyright law also allows the pioneer exclusive rights over the development
of the expressed ideas.
Fishman (2015) identifies a creativity paradox: that limits on follow-on expression can
promote robust original pioneering creativity. He notes, for example, that George Lucas
developed the plot for Star Wars only after he failed to get a license for a remake of Flash
Gordon. Constraints of freedom to build on the works of others can promote creative
breakthroughs.
2. Adequacy of disclosure requirements

Both patent law and copyright law provide for the disclosure and dissemination of
knowledge, which promotes cumulative innovation (Menell, 2019a). Patent law requires

disclosure as the quid pro quo for patent protection. In the software field, however, disclo-
sure of source code is not required under the best mode requirement—the inventor need
only disclose the functions of the software on the grounds that a person of ordinary skill
in the art could write software to perform the functions without undue experimentation.
In practice, however, the knowledge protected by many software patents would be difficult
and costly to decipher without access to the source code, which is usually maintained as
a trade secret. This slows the process by which follow-on inventors can build upon earlier
generations.
In the case of most copyrighted works, the knowledge contained in the work may be
comprehended from direct inspection (Menell, 2016a). Therefore, the publication of a
work discloses and disseminates. Furthermore, the Copyright Act requires those who
register a work, which is optional, to deposit a copy with the Library of Congress.
Some copyrighted works, however, do not lend themselves to visual inspection and
comprehension. Computer software cannot typically be perceived unless it is available
in source code. Copyright Office regulations, however, do not require disclosure of the
entirety of source code in order to secure copyright registration. Thus, as with utility
patent law, copyright law allows protection of software without providing access to the
underlying knowledge. As a result, follow-on invention is stifled, although such rules may
deter infringement that would otherwise be difficult to detect.
D. Rights of Others (and Limitations and Defenses)
Several doctrines provide safety valves, beyond the limitations embodied in scope of pro-
tection, for promoting cumulative innovation. These include the patent law’s experimental
use doctrine, copyright law’s fair use doctrine, and the reverse engineering doctrines of
trade secret law and copyright law. In addition, copyright law provides for several exemp-
tions for educational and related purposes which can be viewed as promoting basic educa-
tion for new authors and artists. Copyright law also provides several compulsory licenses.
1. Experimental use

A subsequent inventor who wants to improve a patented technology may benefit from
experimenting with it. United States patent law has had a common law exemption for
‘philosophical experiments’ and research to ascertain ‘the sufficiency of [a] machine to
produce its described effects’ (Whittemore v. Cutter, 29 F. Cas. 1120 (C.C.D. Mass. 1813)).3
Subsequent cases, however, have declared the defense to be ‘truly narrow’ and applicable
solely to activities ‘for amusement, to satisfy idle curiosity, or for strictly philosophical
inquiry,’ (Roche Prods., Inc. v. Bolar Pharm. Co., 733 F.2d 858, 863 (Fed. Cir. 1984);
Madey v. Duke Univ., 307 F.3d 1351 (2002)), prompting several scholars to warn that
patent law unduly hinders academic and basic research and unduly supplants academic
and scientific norms promoting progress, disclosure, and cumulative innovation.
Eisenberg (1989) proposes to exempt research that could potentially lead to improve-
ments or design-arounds of patented technology, but she also points out the inherent
3
Article 27(b) of the European Patent Convention exempts ‘acts done for experimental
purposes relating to the subject matter of the patented invention.’

contradiction that arises when further research is the main use of the patented invention.
A broad exemption could entirely undermine the profitability of the patent. However,
the effect of a research exemption depends on an ancillary doctrinal question, namely,
whether the invention that is achieved by using the prior invention will infringe the prior
patent (Scotchmer, 2004, ch. 5). If so, the exemption may (counterintuitively) increase
the patent holder’s profit. Exercising the research exemption can put the improver in the
position of bargaining for a license ex post (after he has sunk his costs) rather than ex
ante. This strengthens the bargaining position of the first patent holder.
Kieff (2001b) contends that the exclusive rights provided by patents promote university
research by increasing private investment in research and improving the efficiency of
academic research environments. His analysis does not, however, directly address whether
a more permissive experimental use doctrine would adversely affect the flow of private
research investment into universities. Based on survey research, Walsh et al. (2003)
find little evidence that patents on research tools have significantly impeded university
research. There is one notable exception: the field of genetic diagnostics. Cai (2004) notes
that the chilling effect of a narrow experimental use defense may not be very significant
due to patent holders’ rational forbearance in enforcing against universities as well as legal
constraints (sovereign immunity) on suing state actors (compare Menell, 2000).
Many legal scholars in addition to Eisenberg propose alterations to existing law. Mueller
(2001) endorses a broadened experimental use defense along the lines of the European
system as well as compulsory licensing. Feit (1989) proposes a compulsory license for
patents infringed during experimentation to improve patented technology. These authors,
like Eisenberg, raise particular concerns about patents on upstream research tools, par-
ticularly in the bioscience field. O’Rourke (2000) proposes a fair use doctrine for patents
that would go beyond the similar doctrine for copyright by allowing courts to judge
permissible conduct and impose compulsory royalties. Dreyfuss (2003) proposes to allow
experimental use if the investigator’s institution promptly waives patents on subsequent
discoveries, subject to a ‘buyout’ provision. In cases where a patentee has refused to
license to a nonprofit on reasonable terms, Nelson (2004) proposes a research exemption,
provided the researcher agrees to publish his results and agrees either not to patent his
own results or to license them on nonexclusive and reasonable terms. Strandburg (2004)
proposes to exempt improvement patents and to provide for compulsory licensing of
research tools. Other authors (Epstein, 2003; Merges, 1996) counter that such proposals
would entail significant administrative costs and have complex effects upon licensing
markets and the formation of licensing institutions.
2. Fair use

The fair use doctrine in copyright law exempts a user from liability for infringement when
copyrighted works are used for criticism, comment, news reporting, teaching, scholarship,
research, and other transformative uses (17 U.S.C. § 107). In applying this doctrine, courts
balance the purpose and character of the use (including whether such use is of a com-
mercial nature or is for nonprofit educational purposes), the nature of the copyrighted
work, the amount of copying, and, most importantly, the effect of the use upon the
potential market for or value of the copyrighted work. Transformative or product uses are
more likely to be permissible, whereas uses that merely supplant the underlying work are
disfavored. In this way, the fair use doctrine promotes significant creative advances while

protecting the pioneer from direct market competition. The fair use doctrine can also be
seen as an efficient means for permitting uncompensated use of copyrighted material
where the transactions costs of licensing or other means of exchange would prevent a
transfer through the market (Gordon, 1982).
Courts have applied the fair use doctrine to enable software developers to make copies
of protected programs for purposes of learning how such software functions (Samuelson
and Scotchmer, 2002). With such code deciphered, the rivals can discern unprotectable
elements (e.g., interoperability specifications) which they can implement, using a clean
room process,4 in their own commercial products. In this way, the fair use doctrine oper-
ates as a form of ‘experimental use’ exemption.
The long duration of copyright protection can impede cumulative creativity. Liu (2002)
and Hughes (2003) suggest that courts can ameliorate this concern by expanding the scope
of fair use as copyrights age. Menell (1987, 2019b) advocates a genericide doctrine for
software interface features that become widely adopted as a means to promote positive
network externalities.
3. Reverse engineering

As with copyright law, trade secret law allows others not bound by contractual constraints
to reverse engineer technology in order to determine how it functions. To the extent that
they decipher trade secrets, they undermine the inventor’s advantage. By disclosing the
information, they destroy the trade secret. The reverse engineering limitation on trade
secret protection thus exposes the trade secret owner to free riding by others. Nonetheless,
most commentators believe that it strikes a salutary balance between protection on the
one hand and competition and the dissemination of knowledge on the other (Landes and
Posner, 2003; Samuelson and Scotchmer, 2002). The trade secret owner can ‘purchase’
greater protection against this risk by investing in higher levels of security (e.g., more
effective encryption for software-encoded technology). The inventor can also pursue
patent protection, which proscribes reverse engineering, although only for the limited
duration of the patent, and mandates disclosure of the invention to the public. By declin-
ing to pursue patent protection (or failing to satisfy the requirements thereof), however,
inventors should not be able to secure potentially perpetual rights in technologies merely
by encrypting them or otherwise obscuring how they function. To do so would undermine
the larger balance of the federal intellectual property system.
4. Other limitations: exemptions, compulsory licenses, voluntary licenses, open source,

and dedication to the public domain
Intellectual property regimes can afford cumulative creators freedom to pursue follow-on
projects through statutory exemptions and licensing options. In addition, intellectual
property owners can sometimes benefit by promoting collaboration and widespread
adoption of their technology or works.
4
Using only functional specifications that are outside of copyright protection (17 U.S.C.
§ 102(b)), clean room programmers can independently produce code that accomplishes the same
functions as a target program without infringing copyright protection in the target program’s code
(Sammi et al., 2013).

The utility patent system has strong rights with relatively few limitations. Nonetheless,
a variety of proprietors have chosen to share their technology with competitors and
downstream innovators (Rice, 2015; Merges, 2004a). These firms can benefit through
network effects (Menell 2019b) and preempting or undermining adversaries. These initia-
tives are often coordinated through open source or standard-setting organizations (SSOs)
(Contreras, 2019; Mattioli, 2012). The free and open source software (FOSS) movement
promotes ex ante dedication of technological advances to the public domain and commit-
ment to open development principles. A growing number of companies have joined these
efforts on an ex post basis by pledging their patents to the public. Technology companies
have developed the Defensive Patent License and License on Transfer pledges as a means
of reducing the costs and risks of opportunistic patent assertions (Schultz and Urban,
2012). SSOs seek to lessen the tension between employing the best technological solutions
in industry standards and ensuring widespread access to standards by requiring members
to disclose standard-essential patents (SEPs) and license them on fair, reasonable, and
non-discriminatory (FRAND) terms (Contreras, 2019).
The Copyright Act provides numerous exemptions for public interest (performance
rights for designated educational, religious, and civic uses, reproduction for people with
disabilities) and transaction cost (sound recording performance rights for small busi-
nesses, running software on computers) purposes. It also affords compulsory licenses to
promote dissemination of existing works (juke boxes, cable television, satellite retransmis-
sion, public broadcasting, and webcasting) and cumulative creativity.
The compulsory license mechanism for cover versions of musical compositions has
spurred tremendous cumulative innovation in sound recordings. Once a musical composi-
tion has been released with consent of the copyright owner as a sound recording, any
sound recording artist may record and distribute copies of that composition without
consent of any copyright owner. This privilege is made possible by limiting the scope
of the sound recording copyright to exact duplication5 and establishing a compulsory
license rate for copies made of the underlying musical composition (currently 9.1 cents
per copy) (17 U.S.C. § 115). The creative freedom associated with this privilege, however,
is constrained by the statute—the follow-on recording artist may make ‘a musical
arrangement of the work to the extent necessary to conform it to the style or manner of
interpretation of the performance involved, but the arrangement shall not change the
basic melody or fundamental character of the work’ (17 U.S.C. § 115(a)(2)). This privilege
also does not extend to the use of prior sound recordings—as, for example, in digital
sampling—without the consent of both the musical composition copyright owner and
the sound recording copyright owner. Nonetheless, this compulsory license has fostered
a wider body of interpretations of musical compositions than would have occurred if
musical composition copyright owners held exclusive (blocking) rights. Menell (2016b)
proposes bringing order to music mashup through a new compulsory license.
5
‘The exclusive right of the owner of copyright in a sound recording . . . is limited to the right
to prepare a derivative work in which the actual sounds fixed in the sound recording are rearranged,
remixed, or otherwise altered in sequence or quality. The exclusive rights of the owner of copyright
in a sound recording . . . do not extend to the making or duplication of another sound recording
that consists entirely of an independent fixation of other sounds, even though such sounds imitate
or simulate those in the copyrighted sound recording . . .’ (17 U.S.C. § 114(b)).

Building on the open source model, Creative Commons provides tools to assist authors
in promoting the reuse and remixing of their works at the time of creation by opting into
a different set of defaults than those provided by the Copyright Act (Van Houweling,
2008). Creative Commons also provides tools to assist creators in finding licensed works
that can be shared, remixed, or reused.
E. Remedies
As noted above, the main remedies for patent and copyright infringement are injunctions
and compensation for past injury, possibly compounded to treble damages in case of
willfulness. As these laws are interpreted, courts operate from a baseline of prospective
injunctive relief and compensatory damages for past injury. Hence, they do not generally
adjust the level of damages as a policy lever, except in the context of enhanced damages,
which we discuss below.
With regard to awarding damages for past infringement, the court will often be in the
position of having to decide whether, absent the infringement, the right holder would
have licensed. If not, the lost profit may be the value lost to the patent holder because
the follow-on product was preempted by the infringer. If licensing would (should) have
occurred, the lost profit is lost royalty. These are two different inquiries.
Lost royalty is even more speculative in the cumulative context than in the stand-alone
context, for a sound theoretical reason (Schankerman and Scotchmer, 2001). On one
hand, the potential damage award establishes the maximum license fee. On the other
hand, the equilibrium license fee establishes the damage award. Hence, there is an inherent
circularity that leads to multiple equilibria. Because of multiple equilibria, the profitabil-
ity of the patent is unknowable in advance to a researcher investing in it. This problem is
especially acute for research tools and will be less acute for inventions where infringement
leads to competition and dissipates total profit. Because of the dissipation, infringement
is its own punishment, and infringement is more easily deterred.
The awarding of enhanced damages in patent law (up to treble) and statutory dam-
ages in copyright law can be a policy lever, although it is restricted to penalizing willful
infringement. It is not generally seen as a way of addressing the cumulative innovation
problem. The patent law standard for awarding enhanced damages produces a deleterious
effect upon cumulative innovation. To reduce exposure for treble damages (which is based
upon a finding of willfulness), patent attorneys routinely advise their clients (including
research engineers and scientists) to avoid reading patent prior art, in effect negating a
valuable aspect of the disclosure function of the patent system (Lemley and Tangri, 2003).
This is a form of overdeterrence of socially beneficial behavior—learning from prior
discoveries. The dissuasion to consult patent prior art undoubtedly results in duplication
of research and may lead a researcher to overlook valuable potential solutions to scientific
and technical problems.
The growing use of SSOs to establish technical standards subject to FRAND licens-
ing has promoted follow-on innovation by reducing the potential for exclusionary and
extortionate remedies. Although most SSOs have not expressly barred injunctive relief
or set FRAND licensing schedules, the Institute of Electrical and Electronics Engineers
(IEEE) barred its members from holding patents covering IEEE standards from seeking
or threatening to seek injunctions or exclusion orders against potential licensees who are

willing to negotiate for licenses (IEEE, 2015). Furthermore, courts take FRAND commit-
ments into account in evaluating requests for injunctive relief under the eBay standard (see
Apple Inc. v. Motorola, Inc., 757 F.3d 1286, 1331–2 (Fed. Cir. 2014); but compare Apple
Inc. v. Samsung Elecs. Co., 809 F.3d 633 (Fed. Cir. 2015) (emphasizing right to exclude
and the importance of injunctions)). Moreover, although SSOs do not set the FRAND
license rates, courts have interpreted the principal goal of standard-setting agreements to
be widespread adoption of the standard and barring FRAND licensors from capturing the
coordination and network value of the standard (CSIRO v. Cisco Sys., Inc., 809 F.3d 1295
(Fed. Cir. 2015); Ericsson, Inc. v. D-Link Sys., Inc., 773 F.3d 1201, 1229–35 (Fed. Cir. 2014);
Microsoft Corp. v. Motorola, Inc., No. C10-1823JLR, 2013 U.S. Dist. LEXIS 60233, 2013
WL 2111217 (W.D. Wash. Apr. 25, 2013); In re Innovatio IP Ventures LLC Patent Litig.,
No. 11 C 9308, 2013 U.S. Dist. LEXIS 144061, 2013 WL 5593609 (N.D. Ill. Oct. 3, 2013)).
Channeling doctrines play a critical role in ensuring that foundational technologies are
protected unless they meet patent law’s relatively high innovation thresholds. Furthermore,
patent law has a relatively short duration. In addition, SSOs and open licensing have
promoted widespread adoption of network technologies.
This rational hierarchical structure of the intellectual property system has been
undermined by the Federal Circuit’s interpretation of copyright and design patent
channeling doctrines. In Oracle America, Inc. v. Google Inc., 750 F.3d 1339 (Fed. Cir.
2014), the Federal Circuit’s narrow interpretation of copyright law’s channeling doctrines
enabled Oracle to assert control over functional elements of the Java application program
interface elements (Menell, 2018). As a result of this decision, the safe harbor of clean
room implementation of functional specifications is no longer secure. The Oracle v.
Google precedent creates the potential for software developers to assert long-lived
copyright protection over interface specifications without meeting a substantial threshold
of technological advance. In Apple Inc. v. Samsung Elecs Co., 786 F.3d 983 (Fed. Cir.
2015), the Federal Circuit’s narrow interpretation of design patent law’s functionality
limitation enabled Apple to assert control over functional design elements of smartphone
technology. This decision affords Apple protection for the rounded, rectangular shape of
its iPhone and iPod devices and visual icons without any showing that they constituted
novel and nonobvious technological advances.
REFERENCES
Abramowicz, Michael. 2019. ‘Prize and Reward Alternatives to Intellectual Property,’ in Peter S. Menell and
Ben Depoorter, eds., Research Handbook on the Economics of Intellectual Property, vol. I. Cheltenham, UK
and Northampton, MA, USA: Edward Elgar Publishing.
Akerlof, George A., et al. 2003. ‘Amici Curiae in Support of Petitioners in Eldred v. Ashcroft, 537 U.S. 186.’
Allison, John R., and Mark A. Lemley. 2002. ‘The Growing Complexity of the United States Patent System,’
Anton, James J., and Dennis A. Yao. 2007. ‘Finding ‘Lost’ Profits: An Equilibrium Analysis of Patent
Infringement Damages,’ 23 Journal of Law, Economics, and Organization 186–207.
Armond, Michelle. 2003. ‘Introducing the Defense of Independent Invention to Motions for Preliminary
Injunctions in Patent Infringement Lawsuits,’ 91 California Law Review 117–62.

Arora, Ashish, Andrea Fosfuri, and Alfonso Gambardella. 2001. Markets for Technology: The Economics of
Innovation and Corporate Strategy. Cambridge, MA: The MIT Press.
Arrow, Kenneth. 1962. ‘Economic Welfare and the Allocation of Resources for Inventions,’ in Richard R.
Nelson, ed., The Rate and Direction of Inventive Activity. Princeton, NJ: Princeton University Press.
Ayres, Ian, and Paul Klemperer. 1999. ‘Limiting Patentee’s Market Power without Reducing Innovation Incentives:
The Perverse Benefits of Uncertainty and Non-Injunctive Remedies,’ 97 Michigan Law Review 985–1033.
Bakos, Yannis, Erik Brynjolfsson, and Douglas Lichtman. 1999. ‘Shared Information Goods,’ 42 Journal of
Law and Economics 117–55.
Bar-Gill, Oren, and Gideon Parchomovsky. 2003. ‘The Value of Giving Away Secrets,’ 89 Virginia Law Review
1857–95.
Besen, Stanley M., and Sheila Nataraj Kirby. 1989a. ‘Private Copying, Appropriability and Optimal Copyright
Royalties,’ 32 Journal of Law and Economics 255–80.
Besen, Stanley M., and Sheila Nataraj Kirby. 1989b. ‘Compensating Creators of Intellectual Property,’ Rand
Corporation, Santa Monica, CA, No. R-3751-MF.
Bessen, James, and Eric Maskin. 2009. ‘Sequential Innovation, Patents, and Imitation,’ 40 The RAND Journal
of Economics 611–35.
Blair, Roger D. and Thomas F. Cotter. 1998. ‘An Economic Analysis of Damages Rules in Intellectual Property
Law,’ 39 William and Mary Law Review 1585–694.
Blair, Roger D. and Thomas F. Cotter. 2002. ‘Strict Liability and Its Alternatives in Patent Law,’ 17 Berkeley
Britt, Bill. 1990. ‘International Marketing: Disney’s Global Goals,’ Marketing (May 17, 1990), 22–6.
Burk, Dan L., and Mark A. Lemley. 2002. ‘Is Patent Law Technology-Specific?,’ 17 Berkeley Technology Law
Journal 1155–206.
Burk, Dan L., and Mark A. Lemley. 2003. ‘Policy Levers in Patent Law,’ 89 Virginia Law Review 1575–696.
Cai, Michelle. 2004. ‘Madey v. Duke University: Shattering the Myth of Universities’ Experimental Use
Defense,’ 19 Berkeley Technology Law Journal 175–92.
Calabresi, Guido, and A. Douglas Melamed. 1972. ‘Property Rules, Liability Rules, and Inalienability: One View
of the Cathedral,’ 85 Harvard Law Review 1089–126.
Chang, Howard F. 1995. ‘Patent Scope, Antitrust Policy, and Cumulative Innovation,’ 26 The RAND Journal
Che, Yeon-Koo, and Ian Gale. 2003. ‘Optimal Design of Research Contests,’ 93 American Economic Review 646–71.
Chen, Yeh-ning, and Ivan Png. 2003. ‘Information Goods, Pricing, and Copyright Enforcement: Welfare
Analysis,’ 14 Information Systems Research 107–23.
Conti, Annamaria, Jerry Thursby, and Marie Thursby. 2013. ‘Patents as Signals for Startup Financing,’ 22
Journal of Industrial Economics 592–622.
Contreras, Jorge L. 2019. ‘Technical Standards, Standards-Setting Organizations and Intellectual Property:
A Survey of the Literature (with an Emphasis on Empirical Approaches),’ in Peter S. Menell and David L.
Schwartz, eds., Research Handbook on the Economics of Intellectual Property, vol. II. Cheltenham, UK and
Cooter, Robert D. 1982. ‘Economic Analysis of Punitive Damages,’ 56 Southern California Law Review 79–100.
Cotter, Thomas F. 2004. ‘An Economic Analysis of Enhanced Damages and Attorney’s Fees for Willful Patent
Infringement,’ 14 Federal Circuit Bar Journal 291–331.
David, Paul A. 1985. ‘New Technology, Diffusion, Public Policy, and Industrial Competitiveness,’ in Ralph
Landau and Nathan Rosenberg, eds., Harnessing Technology for Economic Growth. Washington, DC: The
National Academies Press.
de Laat, Eric A.A. 1996. ‘Patents or Prizes: Monopolistic R&D and Asymmetric Information,’ 15 International
Journal of Industrial Organization 369–90.
Denicolò, Vincenzo. 1996. ‘Patent Races and Optimal Patent Breadth and Length,’ 44 Journal of Industrial
Economics 249–65.
Denicolò, Vincenzo. 1997. ‘Patent Policy with a Finite Sequence of Patent Races,’ Department of Economics,
University of Bologna, mimeograph.
Denicolò, Vincenzo. 2002. ‘Two-Stage Patent Races and Patent Policy,’ 31 Journal of Industrial Economics 488–501.
Dreyfuss, Rochelle Cooper. 2000. ‘Are Business Method Patents Bad for Business?,’ 16 Santa Clara Computer
and High Technology Law Journal 263–80.
Dreyfuss, Rochelle Cooper. 2003. ‘Varying the Course,’ in F. Scott Kieff, ed., Perspectives on the Properties of
the Human Genome Project. New York, NY: Elsevier.
Dyson, James. 2001. A History of Great Inventions. New York, NY: Carroll & Graf.
Eisenberg, Rebecca S. 1989. ‘Patents and the Progress of Science: Exclusive Rights and Experimental Use,’ 56
Eisenberg, Rebecca S. 2000. ‘The Promise and Perils of Strategic Prior Art Creation through Publication: A
Response to Professor Parchomovsky,’ 98 Michigan Law Review 2358–70.

Epstein, Richard. A. 2003. ‘Steady the Course: Property Rights in Genetic Material,’ in F. Scott Kieff, ed.,
Perspectives on Properties of the Human Genome Project. New York, NY: Elsevier.
Feit, Irving N. 1989. ‘Biotechnology Research and the Experimental Use Exception to Patent Infringement,’ 71
Journal of the Patent and Trademark Office Society 819–40.
Fisher III, William W. 2004. Promises to Keep: Technology, Law, and the Future of Entertainment. Palo Alto,
CA: Stanford University Press.
Fishman, Joseph P. 2015. ‘Creating around Copyright,’ 128 Harvard Law Review 1333–404.
Friedman, David D., William M. Landes, and Richard A. Posner. 1991. ‘Some Economics of Trade Secret Law’,
5 Journal of Economic Perspectives 61–72.
Galasso, Alberto, and Mark Schankerman. 2015. ‘Patents and Cumulative Innovation: Causal Evidence from
the Courts’, 130 The Quarterly Journal of Economics 317–69.
Gallini, Nancy T. 1992. ‘Patent Policy and Costly Imitation’, 23 The RAND Journal of Economics 52–63.
Gallini, Nancy T. 2002. ‘The Economics of Patents: Lessons from Recent U.S. Patent Reform,’ 17 The Journal
of Economic Perspectives 131–43.
Gallini, Nancy T., and Suzanne Scotchmer. 2002. ‘Intellectual Property: When Is it the Best Incentive
Mechanism?,’ 2 Innovation Policy and the Economy 51–78.
Goldstein, Paul. 1986. ‘Infringement of Copyright in Computer Programs,’ 47 University of Pittsburgh Law
Review 1119–30.
and Its Predecessors,’ 82 Columbia Law Review 1600–655.
Graham, Stuart, Peter Menell, Carl Shapiro, and Tim Simcoe. 2017. ‘Final Report of the Berkeley Center for
Law & Technology Patent Damages Workshop,’ 25 Texas Intellectual Property Law Journal 113–40.
Green, Jerry R., and Suzanne Scotchmer. 1995. ‘On the Division of Profit in Sequential Innovation,’ 26 Journal
of Industrial Economics 20–33.
Gilbert, Richard, and Carl Shapiro. 1990. ‘Optimal Patent Length and Breadth,’ 21 Journal of Industrial
Economics 106–12.
Grushcow, Jeremy M. 2004. ‘Measuring Secrecy: A Cost of the Patent System Revealed,’ 33 Journal of Legal
Studies 59–84.
Heller, Michael A. and Rebecca S. Eisenberg. 1998. ‘Can Patents Deter Innovation? The Anticommons in
Hicks, Diana. 1995. ‘Published Papers, Tacit Competencies, and Corporate Management of the Public/Private
Character of Knowledge,’ 4 Industrial and Corporate Change 401–24.
Hicks, John. 1932. The Theory of Wages. London: Macmillan.
Hughes, Justin. 2003. ‘Fair Use Across Time,’ 50 UCLA Law Review 775–800.
Hunt, Robert M. 1999. ‘Nonobviousness and the Incentive to Innovate: An Economic Analysis of Intellectual
Property Reform,’ Federal Reserve Bank of Philadelphia Working Paper 99–3.
Hunt, Robert M. 2004. ‘Patentability, Industry Structure, and Innovation,’ 52 Journal of Industrial Economics
401–25.
Johnson, William R. 1985. ‘The Economics of Copying,’ 93 Journal of Political Economy 158–74.
Kaplow, Louis. 1984. ‘The Patent-Antitrust Intersection: A Reappraisal,’ 97 Harvard Law Review 1813–92.
Kaplow, Louis, and Steven Shavell. 1996. ‘Property Rules versus Liability Rules: An Economic Analysis,’ 109
Harvard Law Review 715–89.
Kieff, F. Scott. 2001a. ‘Property Rights and Property Rules for Commercializing Inventions,’ 85 Minnesota Law
Review 697–754.
Kieff, F. Scott. 2001b. ‘Facilitating Scientific Research: Intellectual Property Rights and the Norms of Science:
A Response to Rai and Eisenberg,’ 95 Northwestern University Law Review 691–705.
Kitch, Edmund W. 1977. ‘The Nature and Function of the Patent System,’ 20 The Journal of Law and Economics
265–90.
Kitch, Edmund W. 1980. ‘The Law and Economics of Rights in Valuable Information,’ 9 Journal of Legal
Studies 683–723.
Klemperer, Paul. 1990. ‘How Broad Should the Scope of Patent Protection Be?,’ 21 Journal of Industrial
Economics 113–30.
La Manna, Manfredi, Ross Macleod, and David De Meza. 1989. ‘The Case for Permissive Patents,’ 33 European
Economic Review 1427–43.
Lanjouw, Jean Olson. 1998. ‘Patent Protection in the Shadow of Infringement: Simulation Estimations of
Patent Value,’ 65 The Review of Economic Studies 761–810.
Lee, Tom, and Louis L. Wilde. 1980. ‘Market Structure and Innovation: A Reformulation,’ 94 Quarterly Journal
Lefstin, Jeffrey A., Peter S. Menell, and David O. Taylor. 2018. ‘Final Report of the Berkeley Center for Law &

Technology Section 101 Workshop: Addressing Patent Eligibility Challenges,’ 33 Berkeley Technology Law
Journal 551–605.
Leibovitz, John S. 2002. ‘Inventing a Nonexclusive Patent System,’ 111 Yale Law Journal 2251–87.
Lemley, Mark A. 1997. ‘The Economics of Improvement in Intellectual Property Law’, 75 Texas Law Review
989–1083.
Lemley, Mark A., and Colleen V. Chien. 2003. ‘Are the U.S. Patent Priority Rules Really Necessary?,’ 54 Hastings
Lemley, Mark A., and Ragesh K. Tangri. 2003. ‘Ending Patent Law’s Willfulness Game,’ 18 Berkeley Technology
1991–2049.
Lichtman, Douglas Gary. 1997. ‘The Economics of Innovation: Protecting Unpatentable Goods,’ 81 Minnesota
Lichtman, Douglas, Scott Baker, and Kate Kraus. 2000. ‘Strategic Disclosure in the Patent System,’ 53
Vanderbilt Law Review 2175–217.
Liebowitz, Stan J. 1982. ‘Durability, Market Structure, and New-Used Goods Models,’ 72 American Economic
Review 816–24.
Liebowitz, Stan J. 1985. ‘Copying and Indirect Appropriability: Photocopying of Journals,’ 93 Journal of
Political Economy 945–57.
Liebowitz, Stan J., and Stephen E. Margolis. 1982. ‘Journals as Shared Goods: Comment,’ 72 American
Liu, Joseph P. 2002. ‘Copyright and Time: A Proposal,’ 101 Michigan Law Review 409–81.
Long, Clarisa. 2002. ‘Patent Signals,’ 69 University of Chicago Law Review 625–79.
Loury, Glenn C. 1979. ‘Market Structure and Innovation,’ 93 Quarterly Journal of Economics 395–410.
Macedo, Charles R.B. 1990. ‘First-to-File: Is American Adoption of the International Standard in Patent Law
Worth the Price?,’ 18 AIPLA Quarterly Journal 193–234.
Machlup, Fritz, and Edith Penrose. 1950. ‘The Patent Controversy in the Nineteenth Century,’ 10 Journal of
Economic History 1–29.
Mak, James, and Gary M. Walton. 1972. ‘Steamboats and the Great Productivity Surge in River Transportation,’
32 Journal of Economic History 619–40.
Mansfield, Edwin. 1986. ‘Patents and Innovation: An Empirical Study,’ 32 Management Science 173–81.
Mattioli, Michael. 2012. ‘Communities of Innovation,’ 106 Northwestern University Law Review 103–56.
Maurer, Stephen M., and Suzanne Scotchmer. 2002. ‘The Independent Invention Defense in Intellectual
Property,’ 69 Economica 535–47.
Maurer, Stephen M., and Suzanne Scotchmer. 2004. ‘Procuring Knowledge,’ in G. Libecap, ed., Advances in the
Study of Entrepreneurship, Innovation and Growth, vol. 15. Amsterdam: JAI Press, Elsevier.
Menell, Peter S. 1987. ‘Tailoring Legal Protection for Computer Software,’ 39 Stanford Law Review 1329–72.
Menell, Peter S. 2000. ‘Economic Implications of State Sovereign Immunity from Infringement of Federal
Intellectual Property Rights,’ 33 Loyola of Los Angeles Law Review 1399–466.
Menell, Peter S. 2007. ‘A Method for Reforming the Patent System,’ 13 Michigan Telecommunications and
Technology Law Review 487–508.
Menell, Peter S. 2011. ‘Forty Years of Wondering in the Wilderness and No Closer to the Promised Land: Bilski’s
Superficial Textualism and the Missed Opportunity to Return Patent Law to its Technology Mooring,’ 63
Menell, Peter S. 2013. ‘It’s Time to Make Vague Software Patents More Clear,’ Wired (Feb. 7, 2013), accessed
March 18, 2019 at http://www.wired.com/opinion/2013/02/its-time-to-make-vague-software-patents-more-clear.
Menell, Peter S. 2016a. ‘Economic Analysis of Copyright Notice: Tracing and Scope in the Digital Age,’ 96
Boston University Law Review 967–1024.
Menell, Peter S. 2016b. ‘Adapting Copyright for the Mashup Generation,’ 164 University of Pennsylvania Law
Review 441–512.
Menell, Peter S. 2018. ‘Rise of the API Copyright Dead? An Updated Epitaph for Copyright Protection
of Network and Functional Features of Computer Software,’ 31 Harvard Journal of Law & Technology
305–490.
Menell, Peter S. 2019a. ‘Economic Analysis of Intellectual Property Notice and Disclosure,’ in Ben Depoorter
and Peter S. Menell, eds., Research Handbook on the Economics of Intellectual Property, vol. I. Cheltenham,
UK and Northampton, MA, USA: Edward Elgar Publishing.
Menell, Peter S. 2019b. ‘Economic Analysis of Network Effects and Intellectual Property,’ in Ben Depoorter and
Peter S. Menell, eds., Research Handbook on the Economics of Intellectual Property, vol. I. Cheltenham, UK
Menell, Peter S., Mark A. Lemley, and Robert P. Merges. 2018. Intellectual Property in the New Technological
Age, vol. I – Perspectives, Trade Secrets & Patents. Berkeley, CA: Clause 8 Publishing.

Merges, Robert P. 1988. ‘Commercial Success and Patent Standards: Economic Perspectives on Innovation,’ 76
Merges, Robert P. 1992. ‘Uncertainty and the Standard of Patentability,’ 7 High Technology Law Journal 1–70.
Merges, Robert P. 1994. ‘Intellectual Property Rights and Bargaining Breakdown: The Case of Blocking
Patents,’ 62 Tennessee Law Review 75–106.
Merges, Robert P. 2004a. ‘A New Dynamism in the Public Domain,’ 71 University of Chicago Law Review
183–203.
Merges, Robert P. 2004b. ‘Compulsory Licensing vs. The Three “Golden Oldies”: Property Rights, Contracts,
and Markets,’ CATO Institute Policy Analysis No 508, accessed March 18, 2019 at http://www.cato.org/pubs/
pas/pa508.pdf.
Merges, Robert P., and John F. Duffy. 2002. Patent Law and Policy: Cases and Materials. San Francisco, CA:
Matthew Bender.
Merges, Robert P., and Richard R. Nelson. 1990. ‘On the Complex Economics of Patent Scope,’ 90 Columbia
Mokyr, Joel. 1990. The Lever of Riches: Technological Creativity and Economic Progress. New York, NY: Oxford
University Press.
Moore, Kimberly A. 2000. ‘Judges, Juries, and Patent Cases: An Empirical Peek inside the Black Box,’ 99
Michigan Law Review 365–409.
Mossinghoff, Gerald J. 2002. ‘The First-to-Invent System Has Provided No Advantage to Small Entities,’ 84
Journal of the Patent and Trademark Office Society 425–31.
Mueller, Janice M. 2001. ‘No ‘Dilettante Affair’: Rethinking the Experimental Use Exception to Patent
Infringement for Biomedical Research Tools,’ 76 Washington Law Review 1–66.
Nelson, Richard R. 1959. ‘The Simple Economics of Basic Scientific Research,’ 67 Journal of Political Economy
297–306.
Nelson, Richard R. 2004. ‘The Market Economy, and the Scientific Commons,’ 33 Research Policy 455–71.
Nelson, Richard R., and Sidney G. Winter. 1982. An Evolutionary Theory of Economic Change. Cambridge,
MA: Harvard University Press.
Netanel, Neil Weinstock. 2003. ‘Impose a Noncommercial Use Levy to Allow Free Peer-to-Peer File Sharing,’
17 Harvard Journal of Law and Technology 1–84.
Newell, Richard G., Adam B. Jaffe, and Robert N. Stavins. 1999. ‘The Induced Innovation Hypothesis and
Energy-Saving Technological Change,’ 114 The Quarterly Journal of Economics 941–75.
Nordhaus, William D. 1969. Invention, Growth and Welfare: A Theoretical Treatment of Technological Change.
Cambridge, MA: The MIT Press.
Novos, Ian E., and Michael Waldman. 1984. ‘The Effects of Increasing Copyright Protection: An Analytic
Approach,’ 92 Journal of Political Economy 236–46.
O’Donoghue, Ted. 1998. ‘A Patentability Requirement for Sequential Innovation,’ 29 Journal of Industrial
Economics 654–79.
O’Donoghue, Ted, Suzanne Scotchmer, and Jacques-François Thisse. 1998. ‘Patent Breadth, Patent Life and the
Pace of Technological Progress,’ 7 Journal of Economics and Management Strategy 1–32.
O’Rourke, Maureen A. 2000. ‘Toward a Doctrine of Fair Use in Patent Law,’ 100 Columbia Law Review 1177–250.
Oddi, A. Samuel. 1989. ‘Beyond Obviousness: Invention Protection in the Twenty-First Century,’ 38 American
Ordover, Janusz A., and Robert D. Willig. 1978. ‘On the Optimal Provisions of Journals qua Sometimes Shared
Goods,’ 68 American Economic Review 324–39.
Ottoz, Elisabetta, and Franco Cugno. 2004. ‘The Independent Invention Defence in a Cournot-Duopoly
Model,’ 12 Economics Bulletin 1–7.
Pakes, Ariel. 1986. ‘Patents as Options: Some Estimates of the Value of Holding European Patent Stocks,’ 54
Econometrica 755–84.
Pakes, Ariel, and Mark Schankerman. 1984. ‘The Rate of Obsolescence of Patents, Research Gestation Lags,
and the Private Rate of Return to Research Resources,’ in Zvi Griliches, ed., R&D, Patents, and Productivity.
Chicago, IL: University of Chicago Press.
Parchomovsky, Gideon. 2000. ‘Publish or Perish,’ 98 Michigan Law Review 926–52.
Polinsky, A. Mitchell. 1980. ‘Resolving Nuisance Disputes: The Simple Economics of Injunctive and Damage
Remedies,’ 32 Stanford Law Review 1075–112.
Polinsky, A. Mitchell, and Steven Shavell. 1997. ‘Punitive Damages: An Economic Analysis,’ 111 Harvard Law
Review 869–962.
Price, W. Nicholson. 2018. ‘Grants,’ forthcoming Berkeley Technology Law Journal.
Reavis, Conner Kathleen, and Richard P. Rumelt. 1991. ‘Software Piracy: An Analysis of Protecting Strategies,’
37 Management Science 125–39.

Reinganum, Jennifer F. 1982. ‘A Dynamic Game of R&D: Patent Protection and Competitive Behavior,’ 50
Econometrica 671–88.
Reinganum, Jennifer F. 1985. ‘Innovation and Industry Evolution,’ 100 Quarterly Journal of Economics 81–99.
Reinganum, Jennifer F. 1989. ‘The Timing of Innovation: Research, Development and Diffusion,’ in Richard
Schmalensee and Robert D. Willig, eds., Handbook of Industrial Organization. Amsterdam: Elsevier.
Rice, James M. 2015. ‘The Defensive Patent Playbook,’ 30 Berkeley Technology Law Journal 725–76.
Rosenberg, Nathan. 1972. ‘Factors Affecting the Diffusion of Technology,’ 10 Explorations in Economic History
3–33.
Ruttan, Vernon W. 2001. Technology, Growth and Development: An Induced Innovation Perspective. Oxford:
Sammi, P. Anthony, Christopher A. Lisy, and Andrew Gish. 2013. ‘Good Clean Fun: Using Clean Room
Procedures in Intellectual Property Litigation,’ 25 Intellectual Property and Technology Law Journal 3–14.
Samuelson, Pamela, and Suzanne Scotchmer. 2002. ‘The Law and Economics of Reverse Engineering,’ 111 Yale
Schankerman, Mark. 1998. ‘How Valuable Is Patent Protection? Estimates by Technology Field,’ 29 The RAND
Schankerman, Mark, and Ariel Pakes. 1986. ‘Estimates of the Value of Patent Rights in European Countries
during the Post-1950 Period,’ 97 Economic Journal 1052–76.
Schankerman, Mark, and Suzanne Scotchmer. 2001. ‘Damages and Injunctions in Protecting Intellectual
Property,’ 32 The RAND Journal of Economics 199–220.
Schultz, Jason, and Jennifer M. Urban. 2012. ‘Protecting Open Innovation: The Defensive Patent License as a
New Approach to Patent Threats, Transaction Costs, and Tactical Disarmament,’ 26 Harvard Journal of Law
Scotchmer, Suzanne. 1991. ‘Standing on the Shoulders of Giants: Cumulative Research and the Patent Law,’ 5
Journal of Economic Perspectives 29–41.
Scotchmer, Suzanne. 1996. ‘Protecting Early Innovators: Should Second-Generation Products be Patentable?,’
27 Journal of Industrial Economics 322–31.
Scotchmer, Suzanne. 1999. ‘On the Optimality of the Patent System,’ 30 The RAND Journal of Economics
181–96.
Scotchmer, Suzanne. 2004. Innovation and Incentives. Cambridge, MA: The MIT Press.
Scotchmer, Suzanne. 2005. ‘Consumption Externalities, Rental Markets and Purchase Clubs,’ 25 Economic
Theory 235–53.
Scotchmer, Suzanne, and Jerry Green. 1990. ‘Novelty and Disclosure in Patent Law,’ 21 Journal of Industrial
Economics 131–46.
Shavell, Steven, and Tanguy van Ypersele. 2001. ‘Rewards versus Intellectual Property Rights,’ 44 Journal of
Shy, Oz, and Jacques-François Thisse. 1999. ‘A Strategic Approach to Software Protection,’ 8 Journal of
Economics and Management Strategy 163–90.
Sidak, J. Gregory. 2008. ‘Trade Secrets and the Option Value of Involuntary Exchange,’ accessed March 18, 2019
at http://ssrn.com/abstract=577244.
Strandburg, Katherine J. 2004. ‘What Does the Public Get? Experimental Use and the Patent Bargain,’ 2004
Wisconsin Law Review 81–153.
Tandon, Pankaj. 1982. ‘Optimal Patents with Compulsory Licensing,’ 90 Journal of Political Economy, 470–86.
Taylor, Christopher Thomas, Aubrey Silberston, and Z.A. Silberston. 1973. The Economic Impact of the Patent
System. London, England: Cambridge University Press.
Thomas, John R. 1999. ‘The Patenting of the Liberal Professions,’ 40 Boston College Law Review 1139–85.
Van Houweling, Molly Shaffer. 2008. ‘The New Servitudes,’ 96 The Georgetown Law Journal 885–950.
Varian, Hal R. 2000. ‘Buying, Renting, Sharing Information Goods,’ 48 Journal of Industrial Economics 473–88.
von Hippel, Eric. 1988. The Sources of Innovation. Oxford: Oxford University Press.
Walsh, John P., Ashish Arora, and Wesley M. Cohen. 2003. ‘Effects of Research Tool Patents and Licensing on
Biomedical Innovation’, in Wesley M. Cohen and Stephen A. Merrill, eds., Patents in the Knowledge-Based
Economy. Washington, D.C.: National Academy Press.
Wright, Brian D. 1983. ‘The Economics of Invention Incentives: Patents, Prizes and Research Contracts,’ 73
American Economic Review 691–707.
17 U.S.C. § 101, § 102(b), § 107, § 114(b), § 115, § 115(a)(2).

35 U.S.C. § 102, § 103, § 112, § 273(a).

Cal. Labor Code § 2870.

America Invents Act of 2011.
Semiconductor Chip Protection Act of 1984.
Administrative and Other Materials
Institute of Electrical and Electronics Engineers (IEEE). 2015. IEEE-SA Standards Board Bylaws, accessed
March 18, 2019 at http://standards.ieee.org/develop/policies/bylaws/approved-changes.pdf.
U.S. Federal Trade Commission. 2003. To Promote Innovation: The Proper Balance of Competition and Patent Law
and Policy (Oct. 2003), accessed March 18, 2019 at https://www.ftc.gov/sites/default/files/documents/reports/
promote-innovation-proper-balance-competition-and-patent-law-and-policy/innovationrpt.pdf.
Cases
Alice Corp. v. CLS Bank International, 573 U.S. 208 (2014).

Apple Inc. v. Motorola, Inc., 757 F.3d 1286 (Fed. Cir. 2014).
Apple Inc. v. Samsung Electronics Co., 809 F.3d 633 (Fed. Cir. 2015).
Apple Inc. v. Samsung Electronics Co., 786 F.3d 983 (Fed. Cir. 2015).
Association for Molecular Pathology v. Myriad Genetics, 569 U.S. 576 (2013).
CSIRO v. Cisco Systems, Inc., 809 F.3d 1295 (Fed. Cir. 2015).
Cuno Engineering Corp. v. Automatic Devices Corp., 314 U.S. 84 (1941).
Dolbear v. American Bell Telephone Co., 126 U.S. 1 (1888).
Ericsson, Inc. v. D-Link Systems, Inc., 773 F.3d 1201 (Fed. Cir. 2014).
Graver Tank & Manufacturing Co. v. Linde Air Products Co., 339 U.S. 605 (1950).
In re Innovatio IP Ventures LLC Patent Litigation, No. 11 C 9308, 2013 U.S. Dist. LEXIS 144061, 2013 WL
5593609 (N.D. Ill. Oct. 3, 2013).
Madey v. Duke University, 307 F.3d 1351 (2002).
Mayo Collaborative Services v. Prometheus Laboratories, 566 U.S. 66 (2012).
Microsoft Corp. v. Motorola, Inc., No. C10-1823JLR, 2013 U.S. Dist. LEXIS 60233, 2013 WL 2111217 (W.D.
Wash. Apr. 25, 2013).
O’Reilly v. Morse, 56 U.S. (15 How.) 62 (1854).
Oracle America, Inc. v. Google Inc., 750 F.3d 1339 (Fed. Cir. 2014).
Roche Products, Inc. v. Bolar Pharmaceutical Co., 733 F.2d 858 (Fed. Cir. 1984).
Sanitary Refrigerator Co. v. Winters, 280 U.S. 30 (1929).
W.L. Gore & Associates v. Garlock, Inc., 721 F.2d 1540 (Fed. Cir. 1983).
Warner-Jenkinson Co. v. Hilton Davis Chemical Co., 520 U.S. 17 (1997).
Whittemore v. Cutter, 29 F. Cas. 1120 (C.C.D. Mass. 1813).

7. Economic analysis of network effects and
intellectual property
Peter S. Menell* 6
Contents
I. Introduction
II. Functioning of Network Markets
III. Interplay of Business Strategy, Contract, Standard Setting, Intellectual Property,
and Competition Policy
IV. Ramifications for Intellectual Property and Competition Policy
A. Parsimony Principle: No Intellectual Property Protection for Functional
Attributes Absent Significant Technological Advance
B. Proportionality Principle: Overcoming Excess Inertia without Undue
Protection
C. Deterrence Principle: Discouraging Overreach with Balanced Remedies
V. Intellectual Property Protection for Network Features
A. Trade Secret Protection
B. Copyright Protection
1. Software copyright legislation: the Copyright Act of 1976, the
CONTU Final Report, and the 1980 amendments
2. Software copyright jurisprudence: the first wave
3. Software licensing
4. Interoperability exception to the DMCA’s anti-circumvention
prohibition
5. Software copyright jurisprudence: the Oracle v. Google litigation
6. Standards and codes
C. Trademark Protection, Unfair Competition Law, and False Advertising
Protection
D. Patent Protection
1. Patentability requirements
2. Scope
3. Licensing
4. Remedies
* Koret Professor of Law and Director, Berkeley Center for Law and Technology, Berkeley
School of Law, University of California. Thanks to participants at the 2015 Economics of
Intellectual Property Research Handbook Conference, the 2017 Intellectual Property Scholars
Conference, and especially Michael Carrier for their comments on this project. I am grateful to
Alex Barata, Louise Decoppet, Amit Elazari, Andrea Hall, Reid Whitaker, and Samantha Vega for
research assistance.
157

5. Design patents
VI. Interplay of Intellectual Property Protection and Competition Policy in Network
Industries
A. Private Enforcement
1. Misuse doctrines
2. The principle of exhaustion
3. Ambush of standard setting processes
4. Breach of contract for failure to license SEPs on FRAND terms
5. Private antitrust liability
B. Public Enforcement
1. Intellectual property licensing guidelines
2. Significant network market enforcement actions
VII. Assessment of Intellectual Property Protection and Competition Policy for
Network Technologies
A. Institutional Considerations
B. Measuring Progress Based on the Normative Principles
1. Parsimony principle
2. Proportionality principle
3. Deterrence principle
VIII. Future Research Directions
References
I. INTRODUCTION
The economics of intellectual property begins with the classic appropriability problem: in
a competitive economy, imitators can enter markets for information goods after inventors
and authors have incurred research and development (R&D) costs and sell the innovative
or creative product at the cost of reproduction. Without means for appropriating an
adequate return on investment in R&D, the market will under-produce technological
advances and creative expression (Menell and Scotchmer, 2007).
The provision of intellectual property protection for technological advances and
creative expression affords inventors and authors a mechanism to recoup their invest-
ments, although not without imposing the deadweight loss of monopoly exploitation
and potentially interfering with cumulative creativity (Menell et al., 2018). Conventional
analysis of intellectual property seeks to balance the duration and scope of intellectual
property rights in order to optimize these tradeoffs (Nordhaus, 1969; Gallini, 1992). The
conventional framework applies to goods and services for which consumer demand is
independent—that is, where one consumer’s utility from consuming a good or service
does not depend on choices of other consumers.
Yet consumer demand for information goods and services can be interdependent,
especially in the digital age. The value to consumers of systems technologies—such as
telecommunication networks (including telephone networks, cable systems, satellite
systems, and Internet protocols), interconnected devices (e.g., mobile phones, operating
systems and application programs, printers and replacement cartridges, audio-video
devices and media), databases (e.g., Internet searches), and electric charging stations

Economic analysis of network effects and IP 159
(e.g., Tesla superchargers)—often depends upon other consumers’ choices. For example,
a smartphone platform with many adopters will attract more app developers, thereby
increasing the functionality and value of that platform for consumers, developers of
complementary goods (e.g., apps), and the platform sponsor.
Platforms function like a common language. Devices that ‘speak’ a common language
(such as a programming language, application program interface (API), or set of graphical
user icons) can communicate with other devices and humans familiar with that language.
Innovators can more easily design peripheral equipment that expands the functionality
of existing devices. Over time, users internalize how a computer language or application
program represents functions, often memorizing the most commonly used series of
keystrokes or developing macros customized to perform their most common tasks. These
human capital investments commit users to particular languages and platforms and
encourage employers to adopt systems that are widely known by prospective employees
so as to recruit promising candidates and reduce training costs (Gandal, 1994). Thus, it
is common for people seeking jobs in programming, accounting, and design fields to list
those computer languages and application programs that they have mastered on their
resumes. Network externalities arise from the enhanced labor mobility and reduced train-
ing costs produced by shared, or at least compatible, computer systems across different
work environments. When people in different places can communicate more efficiently
through compatible file formats, network externalities result.
The value of networks grows disproportionately with their adoption bases. Such posi-
tive feedback dynamics drive a growing number of markets in the information economy
(Shapiro and Varian, 1999), from computer operating systems to mobile phones, printers
(and ink cartridges), video game consoles, Internet search engines (such as Google),
Internet commerce (such as eBay and Amazon), social networks (such as Facebook,
LinkedIn, and Tinder), cloud computing, the Internet of Things, and shared economy
platforms (such as Airbnb and Uber).
Advances in digital and network technologies have dramatically reshaped the com-
petitive and innovative landscape. As a consultant for the Internet dating industry has
remarked, ‘It’s never been cheaper to start a dating site and never been more expensive to
grow one’ (Tugend, 2016). Dating apps usually start by offering free services to new users,
seeking to build a viral bandwagon. If they gain traction through innovative features
or marketing, they then face the daunting task of monetizing the network, typically
through advertising or membership fees. Monetization, however, can reverse the positive
feedback effects, thereby reducing the network’s size, unraveling the network’s benefits,
and jeopardizing the platform’s sustainability. Finding the right balance between viral
growth and monetization is the principal challenge of a growing range of enterprises in
the Internet Age.
The interdependence of consumer demand has important ramifications for the design
of intellectual property and competition policy. In a static economic model (i.e., one
without innovation), consumers benefit from robust competition within product stand-
ards. Open access to product standards encourages realization of network externalities.
Although bandwagon effects can enhance consumer welfare in a static context, they can
also make it more difficult for developers of improved platforms to enter the market.
Consumers and suppliers of complementary products can face significant switching costs
in migrating from one platform to another. For example, once businesses have invested

heavily in developing programs to run on a software platform (e.g., macros for the Lotus
1-2-3 spreadsheet), it becomes much more difficult for a competitor offering an enhanced
spreadsheet (e.g., Borland Quattro Pro) to enter the market unless they can provide a
low-cost migration path. Facebook’s widespread success and user investment made it
difficult for even Google to build a sustainable competing social network. Orkut, Google
Buzz, Google Friend Connect, and Google+ have failed or languished.
The technical standards governing access to platforms, commonly referred to as applica-
tion program interfaces (APIs) in the software industry, play a critical role in consumer
and programmer adoption decisions, market entry, and competition. Those who control
a widely adopted platform can obstruct new innovative platforms and complementary
products and services (such as refilling and repair). Familiarity with the user interface and
features, connections to other network adopters (such as Facebook friends), and investments
in complementary assets (such as macros that run on the platform) can keep consumers on
an otherwise inferior platform. The human capital investment in learning an API can lock
programmers into a platform, and sunk costs in manufacturing facilities, fabrication designs,
and contracts with suppliers and customers can lock manufacturers into design choices.
At the same time, the ability to secure an innovative platform can be vital to investing
in the R&D needed to advance systems technologies. Without the prospect of earning a
significant return on research, development, and marketing of a new platform, investors
have little incentive to take on the risk of investing the substantial resources necessary to
challenge an entrenched platform. Therefore, the availability, scope, and remedies for intel-
lectual property protection for network features of systems technologies and platforms
(e.g., interface specifications) provide both a key strategic asset for controlling network
markets and a critical mechanism for promoting advances in network technologies.
Demand-side or network effects, therefore, complicate the design of an optimal intel-
lectual property regime. Control of interface specifications and other network features
of computer technologies through intellectual property protection has become the key
to market dominance in a growing number of important Information Age markets.
Nearly all of the major software copyright disputes, as well as a key exception to the
Digital Millennium Copyright Act of 1998’s (DMCA) anti-circumvention provisions,
have revolved around the protectability of interface specifications. Patent protection
for network technologies has also become a critical battleground with some disputes
centered on licensing network technologies through standard setting organizations. Trade
secrecy, trademark protection, and contract law are also important tools for regulating
competition in network markets.
This chapter explores the critical role of intellectual property in network markets as
well as the ramifications of network effects for the design of intellectual property regimes.
Section II describes the functioning of network markets. Section III examines the interplay
of business strategy, contract, standard setting organizations, intellectual property, and
competition policy. Section IV presents three principles for tailoring intellectual property
regimes and competition policy for network technologies. Section V traces the evolution
of intellectual property protection for network features of systems and platforms. Section
VI discusses the interplay of intellectual property protection and competition policy.
Section VII assesses the extent to which intellectual property protection and competition
policy align with normative design principles. Section VIII identifies promising areas for
future research.

II. FUNCTIONING OF NETWORK MARKETS
In many market settings, consumers’ utility functions are independent. Take, for example,
the market for ice cream. My enjoyment of a particular flavor (e.g., hazelnut chocolate
chip), style (e.g., gelato), or brand (e.g., Talenti) does not depend significantly on the
utility that other consumers derive from the purchase and consumption of ice cream. It is
possible that greater popularity of a flavor, style, or brand makes that combination more
widely available or lowers the price due to economies of scale on the production side,
but competition usually ensures efficient allocation of resources in these circumstances.
The effects are more likely pecuniary, which work through the market and only affect
the distribution of value, than technological, which affect the economic efficiency of the
economy (Liebowitz and Margolis, 1994).
By contrast, some market equilibria depend critically on the number of consumers
that have joined or are likely to join a particular platform. Take, for example, a social
network like Facebook. A new entrant to this market, say Google+, might offer enhanced
functionality. But if most of my social network is already on Facebook and I cannot easily
bridge the two networks, then I am far less likely to switch.
Network effects have long been central to human civilization and market economies.
Languages, measurement systems (metric versus imperial), electrical equipment standards
(alternating current versus direct current; computer networking protocols), driving con-
ventions (left side versus right side), and railroad gauges (the width between and across
rails) are notable examples where demand-side coordination greatly influences consumer
welfare, economic efficiency, and social discourse. Standardized railroad gauge, for
example, supported far-reaching railroad networks, promoted competition in locomotive
and railcar markets, and enabled interconnected rail services (Puffert, 2000, pp. 944–7).
Section III focuses on how such coordination or standardization occurs through business
strategy, technological innovation, intellectual property law, industry and consumer
coordination, and government policies (including antitrust law).
The economic and social value of network effects can be substantial. According to
Metcalfe’s Law—attributed to Robert Metcalfe, co-inventor of the Ethernet, a local
computer network platform that foreshadowed and ushered in the Internet—the value
of a telecommunications network is proportional to the square (n2) of the number of
devices (or nodes (n)) of the system. This economic ‘law’ reflects the potential number of
contacts within a network and assumes that they are each of equal value. Even though this
theoretical maximum is unlikely to be obtained in the real world (Briscoe et al., 2006), the
powerful growth potential of network systems drives much of the information economy.
The net value, of course, also depends on the cost per user. In many telecommunications
and computer applications, such costs are low and have declined over time because of
Moore’s Law—Intel co-founder Gordon Moore’s audacious, yet remarkably accurate,
prediction that the number of transistors on an integrated circuit would double every two
years (later reduced to 18 months) (Moore, 1965; Borwein and Bailey, 2015).
Both real and virtual networks can produce these effects (Shapiro and Varian, 1999,
p. 183; Katz and Shapiro, 1985, p. 424). Real networks entail physical connectivity ena-
bling a user to interact or communicate directly with others. They include transportation
systems (such as railroad gauges), telecommunication systems (such as a telephone or
broadcast network), and media systems (such as data storage devices). By contrast, virtual

networks operate through the evolution of markets for complementary products. The
supply of complementary goods typically drives these markets. For example, by enabling
programmers to develop apps for the iOS platform, Apple promotes a virtual network
surrounding its iPhone and other computer devices. The availability of apps on the iOS
drives demand for iOS devices, which in turn attracts app developers. More apps generate
a wide range of functionality, thereby spurring increased demand for iPhones. Other
examples of virtual networks include application programs that enable users to share data
files with other programs and users, ATM cards and automatic teller machines, credit
cards, and the merchants who accept them, and next generation payment systems such as
Apple Pay and Square. The defining feature of virtual networks is that the demand for the
product depends significantly on the availability of complementary goods and services.
The magnitude of network effects depends on several considerations: interdependencies
of consumer utility functions, range of complementary products or services, availability
of alternative platforms, switching costs, business strategies, and legal limits (such as
intellectual property protection and competition policy) on leveraging network markets.
In some cases, physical limitations govern network access—for example, where a device
must physically or digitally interoperate with other devices. In others, the network is not
physically constrained, but instead driven by consumer familiarity or ease of use.
The design determinants of a network market—interoperability or compatibility
standards—are shaped by the type and degree of ownership, sponsorship, and govern-
ance of network access. Some network standards are established or authorized by a
government or international organization or formal standard setting organization (SSO).
These are sometimes referred to as de jure standards as they have an official backing
and can be enforced by law. Such enforcement can limit or afford access to standards.
Individual companies or consortiums sponsor many important network standards. These
are sometimes referred to as de facto standards, although they might be backed by patent,
copyright, trademark, or false advertising law.
A third important distinction in network markets relates to whether a standard is ‘free,’
open, closed (i.e., proprietary), or somewhere in the middle (Meeker, 2015, pp. 31–47;
West, 2007). The free software movement allows other users to run, study, share, copy,
and modify the software so long as these users permit use of any derivative works on the
same terms. ‘Open source’ software typically connotes that the software or interface is
freely available to any market participant, but there might or might not be restrictions on
the availability of complementary goods embodying the standard. A closed or proprietary
standard is one in which a sponsoring enterprise or organization regulates access, typically
through licensing of intellectual property rights.
Thus, the distinction between open and closed standards can be ambiguous. For
example, many SSOs require that participating enterprises license standard-essential
patents on fair, reasonable, and non-discriminatory (FRAND) terms. Which patents
are ‘standard-essential’ and the license terms on which they are available are rarely fully
specified in advance, creating substantial uncertainty (Contreras and Gilbert, 2015). On
the other hand, ‘free’ software licensed pursuant to the General Public License (GPL)
requires users to make available any software incorporating the licensed code under the
same ‘share and share alike’ restriction (Carver, 2005).
The controversy over the Java API platform illustrates the complexity that can arise
surrounding intermediate—that is, partially open—platforms (Menell, 2018; Lemley and

McGowan, 1998). Sun Microsystems released the Java programming language without
restriction in part to prevent Microsoft from leveraging its Windows desktop computer
operating system monopoly into dominance of website functionality. Sun’s Java strategy
promoted the ‘Write Once, Run Anywhere’ (WORA) principle: the notion that any
browser can execute Java applets (small application programs, such as those used for
animated web pages)—including on Microsoft Windows, Unix, macOS, and Linux.
Over time, Sun developed pre-written API packages to facilitate Java programming.
Sun developed the Java Community Process (JCP), a quasi-public formalized administra-
tive process, for developing technical specifications for Java technology and extensions.
Sun used the JCP and licensing of the Java trademark to promote collaboration and
commitment to the WORA principle. When Google sought to use some, but not all, of the
Java APIs to develop the Android platform and licensed Android using a less restrictive
licensing regime (i.e., not requiring that derivative works be shared on a ‘free’ basis), Sun
and Oracle (which acquired Sun Microsystems in 2010) objected, resulting in one of the
costliest intellectual property battles (Menell, 2018).
Network effects arise whenever the value that consumers place on a product or service
depends upon the number of other consumers or programmers purchasing that product
or using that service. As the number of adopters (or the installed base) of a platform
grows, the benefits of being part of that platform increase. For example, consumers
generally prefer telephone networks or protocols offering the largest user bases.
Like economies of scale (declining unit costs with increased production) on the supply
side of a market, the value of a network generally increases with widespread adoption.
The availability of better application programs to run on an operating system platform
will lead more consumers to prefer that operating system, which in turn will spur a
greater quantity and quality of application programs for that operating system. Whereas
economies of scale typically fall off at some point due to technical or organizational limits,
positive feedback on the demand side generally continues to increase with the size of
the installed base. For this reason, one or a very small number of standards are likely to
predominate in markets with strong network effects, as reflected in Microsoft’s dominance
in microcomputer operating systems, Google’s dominance among Internet search engines,
and Facebook’s dominance as a social network.
The high value that consumers place upon standardization, however, can make it
particularly difficult for improved products to break into the market. Such bandwagon
effects can stifle development and diffusion of improved technology platforms (Farrell
and Saloner, 1985).
III. INTERPLAY OF BUSINESS STRATEGY, CONTRACT,

STANDARD SETTING, INTELLECTUAL PROPERTY, AND
COMPETITION POLICY
The dynamics of network technologies produce a particularly complex strategic playing

field. Firms typically choose among three strategies when competing in network markets:
(1) market dominance through establishing and controlling a new proprietary standard;
(2) adopting an existing standard either through imitation (where it is legally permis-
sible) or licensing; or (3) working with other firms in the industry—either informally,

c ontractually, through formal industry organizations, or through governmental stand-

ardization bodies—to develop an open or quasi-open standard (Farrell and Saloner, 1998;
Besen and Farrell, 1994; Shapiro and Varian, 1999; Farrell and Simcoe, 2012).
Among the strategies firms use to establish their product or service as the de facto
industry standard are: massive advertising campaigns; penetration pricing (pricing
products or services below cost or giving them away in order to speed adoptions by
consumers); issuing impressive product preannouncements to entice consumers and
discourage competitors; providing adopters with various forms of insurance (such as
short term leases or pricing arrangements that tie the price of the system to the number
of adopters); licensing of the product in order to grow the network more rapidly (and to
create competition in the expansion of the network); and vertical integration and strategic
investments into markets for complementary products to assure consumers that valuable
application programs will be available (Farrell and Simcoe, 2012; Farrell and Klemperer,
2007; Shapiro and Varian, 1999; Baseman et al., 1995; Farrell and Saloner, 1986; Dybvig
and Spatt, 1983).
Adopting an existing standard enlarges the size of a network comprising both the
entrant’s product and its rival’s—the existing platform—products. This increases the
desirability of the rival’s products to consumers, thereby reducing the adopter’s market
share (although of a larger market) relative to what it would have been had the firm
adopted an incompatible product standard. Thus, even though the net social welfare of
adopting a rival’s standard may exceed the net social welfare of introducing an incompat-
ible standard, the entrant may nonetheless prefer to adopt an incompatible standard
because the entrant cannot appropriate all of the benefits of compatibility, some of which
accrue to past and present purchasers of the rival’s products (Katz and Shapiro, 1985,
p. 435; Katz and Shapiro, 1986).
Firms often pursue Strategy 2 (adopting an existing standard) and Strategy 3 (collabo-
rating with other firms in establishing a standard) in tandem. Both strategies create a more
traditional market setting in which firms compete over price, quality, and services to win
market share on a common platform. This achieves greater competition on a particular
platform and fosters the realization of network externalities, but may impair competition
to innovate better platforms (Katz and Shapiro, 1994). The market dominance strategy is
often riskier but can produce the highest payoff for the winner.
A firm’s strategy will depend on a range of factors, including its reputation among con-
sumers for serving the type of network market that it has targeted, its available resources
(and access to capital markets) to make the investments in distribution and marketing
necessary to persuade consumers that the firm will prevail in the standard battle, the
strength of its technology for establishing a standard (although such technology need not
be superior to others on the market), and complementary assets within the firm or strong
strategic alliances in vertical markets. The firm’s strategy will also depend upon the avail-
ability of intellectual property protection, contractual means, and technological controls
(e.g., encryption technology) for precluding, limiting, or delaying access by competitors
to the firm’s standard.
IBM successfully pursued the market dominance strategy when it entered the
microcomputer market in the early 1980s. IBM combined its reputation for serving the
mainframe market and technological and marketing capabilities with copyright and
trade secrecy protection for its basic instruction operating system (BIOS) chip. IBM’s

strategic hold on the industry quickly unraveled, however, when competitors successfully
reverse engineered the BIOS chip (Moy, 2000, pp. 72–3; Menell, 1987), making much less
expensive, fully compatible IBM clones available on the market by the mid to late 1980s.
IBM exited the microcomputer hardware industry soon thereafter.
Microsoft emerged as the winner in the microcomputer industry during this upheaval.
Its DOS operating system, on which IBM had previously built its microcomputers,
emerged as the de facto standard. Robust competition in microcomputers using DOS and
a growing array of application programs (including several Microsoft flagship products
such as Word and later Excel) drove adoption of DOS-based computers and fueled
Microsoft’s dominance. Microsoft skillfully migrated users from DOS to Windows, with-
standing Apple’s assertion of intellectual property control of the Mac desktop graphical
user interface. By the mid-1990s, Microsoft dominated the microcomputer industry
through its control of the Windows platform. Apple was a distant second and fading.
The emergence of the Internet in the mid-1990s opened new modes of competition in
computer markets. Netscape’s Navigator Internet browser and Sun’s highly interoperable
Java platform threatened Microsoft’s dominance in the microcomputer and software
marketplace (Lemley and McGowan, 1998). Microsoft responded by integrating its
browser technology, Internet Explorer, into the Windows operating system and engaging
in restrictive licensing agreements with microcomputer manufacturers, thereby reducing
the effective price of its browser to zero. Consequently, the market for Netscape’s browser
evaporated. Microsoft also undermined Java’s efforts to establish a universal meta-
platform for software application programs by offering a proprietary, non-interoperable
version (Sun Microsystems, Inc. v. Microsoft Corp., 46 U.S.P.Q. 2d 1531 (N.D. Cal. March
24, 1998)).
Network effects have allowed particular firms to dominate many Internet markets,
including search (Google), social networks (Facebook), mobile (iOS, Android), and shar-
ing networks (Airbnb, Uber). Apple successfully regained prominence in critical digital
markets through its mobile and App Store network market strategies.
Formal standardization plays a tremendous role in many electronics and telecommu-
nications markets (Contreras, 2019; Lemley, 2002). Russell (2006) traces electrical stand-
ardization through formal standard-setting organizations over more than a century. These
processes have relied on engineers and scientists seeking to promote the best engineering
solutions to technical challenges. They form key infrastructure for the electronics and
telecommunications industries. Engineers from major technology companies participate
in dozens of standard-setting organizations, including many of the leading professional
engineering societies, such as the Institute of Electrical and Electronics Engineers (IEEE)
Standards Association and the Internet Engineering Task Force (IETF). These processes
have carried over to semiconductor designs, mobile phones, Internet protocols, and
computer devices. A typical laptop computer today embodies more than 250 technical
standards (Contreras, 2019).
Intellectual property protection for network technologies can significantly influence
the development of standards, follow-on innovation, and market competition. Patents in
the information and communication technology fields (semiconductors, computers, and
mobile phones) have presented the most salient concerns.
Building on Williamson’s (1985) classic treatment of economic holdup—whereby
asymmetric information, transaction costs, and incomplete contracts create the potential

for a contracting party to extract the value of sunk or locked in, relationship-specific
investments—Lemley and Shapiro (2007a) posit a patent bargaining model in the shadow
of strong potential remedies (automatic injunctive relief and large monetary awards) that
generates an analogous inefficient dynamic. Companies that unwittingly sink large invest-
ments into infringing products are subject to having such investments extracted through
patent infringement litigation. Such extraction can greatly exceed the contribution of
the patented technology relative to the best non-infringing alternative. The presence of
multiple patents covering a single product—what has been referred to as the patent thicket
problem (Shapiro, 2001)—exacerbates holdup effects, creating a royalty stacking problem:
total patent royalty demands may exceed the contribution of patented technologies to the
market demand for the product.
Various scholars have questioned Lemley and Shapiro’s assumptions and empirical
basis for royalty stacking (Elhauge, 2008; Sidak, 2008; Geradin et al., 2008; Geradin
and Rato, 2007; Geradin, 2007; Golden, 2007; but see Lemley and Shapiro, 2007b
(responding to Golden, 2007)). They note that royalty stacking is unlikely to occur with
full information and low transaction costs. There is good reason, however, to question
optimism about ex ante bargaining. Ziedonis (2004), for example, finds that firms acquire
patents more aggressively when the patents for numerous component technologies of an
industry—like the semiconductor industry—are widely distributed. The proliferation
of patent litigation over information and communication technology indicates that
intellectual property protection imposes at least some implicit tax on these network
industries. Nonetheless, more recent empirical research raises doubts about the severity
of royalty stacking. Galetovic and Gupta (2016), for example, find that mobile wireless
prices have fallen, quantities have grown, and the industry has become less concentrated
over time, indicating that royalty stacking may not be as serious as prior research had
claimed. Barnett (2017) surveys the growing literature and concludes that the evidence of
royalty stacking is weak. All would agree, however, that industry coordination through
patent pooling and SSOs can alleviate these problems (National Research Council of the
National Academies, 2013). Such pools, however, can facilitate collusion, raise barriers to
entry, and spark other public policy concerns.
Notwithstanding the widespread use of standard-setting processes and agreements on
technical standards, the rules governing access to standards and the licensing of patented
technologies are rarely specified in advance. SSOs exercise caution to avoid violating
antitrust laws barring price-fixing. In addition, many companies participating in standard
setting processes do not wish to reveal their patent prosecution strategies or pre-commit
to price terms. Thus, most technical standard setting organizations require only that
participants disclose their patented technologies and agree to license standard-essential
patents (SEPs) on FRAND terms. The potential for holdup and royalty stacking remains
(Contreras, 2019; Federal Trade Commission, 2011; Lemley and Shapiro, 2007a).
Some sectors of the software industry have alleviated or avoided these risks by com-
mitting to open source policies (Merges, 2004). Viral forms of open source licensing, such
as the GPL, however, can discourage investment in downstream innovation by limiting
direct appropriability for technological advances. For this reason, Google chose a more
permissive open source license for Android (Menell, 2018). This fostered collaboration
and rapidly expanded the Android network while encouraging innovation by handset
makers and telecommunications companies.

IV. RAMIFICATIONS FOR INTELLECTUAL PROPERTY AND

COMPETITION POLICY
As the preceding analysis suggests, intellectual property protection can play a critical role
in network markets. As one software entrepreneur metaphorically explained, creating an
API is analogous to building a city:
First you try to persuade applications programmers to come and build their businesses on [your
tract of land]. This attracts users, who want to live there because of all the wonderful services
and shops the programmers have built. This in turn causes more programmers to want to rent
space for their businesses, to be near the customers. When this process gathers momentum, it’s
impossible to stop.
Once your city is established, owning the API is like being the king of the city. The king gets
to make the rules: collecting tolls for entering the city, setting the taxes that the programmers
and users have to pay, and taking first dibs on any prime locations (by keeping some APIs
confidential for personal use). (Kaplan, 1996)
This section discusses the general economic considerations bearing on whether and
to what extent intellectual property ought to protect network features of systems
technologies—those features that affect access to or interoperability with a system.
There are two market failures in play in optimizing intellectual property protection.
First, network features of system technologies, like any other technology, are subject to
the classic appropriability problem. Without intellectual property protection, inventors of
more advanced platform technologies will be subject to being undercut by new entrants
who imitate the innovations without bearing R&D costs. First-mover advantages, effec-
tive marketing, trade secrecy, and other strategies might provide sufficient motivation for
some R&D, but there is reason to be concerned that the unregulated market will under-
produce potentially high value, but risky and costly, innovation in network technologies.
Demand-side effects in network markets, however, complicate the conventional analysis
of intellectual property protection. Because of the dynamics of network markets, some
firms might be motivated to limit access to their platforms to reap the outsize profits
from controlling a network market. This strategy, however, can hinder the realization
of network benefits by raising prices, limiting access by third parties, and discouraging
innovation because of the high barriers to entry. Consumers benefit when they and their
devices, systems, and programs ‘speak’ the most widely adopted platform—the lingua
franca—or can translate that code into language their devices understand. This often
provides for greater functionality, such as more software that will run on their platform
and larger communication networks.
At the same time, widely adopted product standards can strand the industry on an
obsolete platform (Farrell and Saloner, 1985). Consumers resist switching costs—from
learning new tools and languages to acquiring new devices. They demand substantial
improvements in efficiency or functionality to jettison comfortable, well-worn devices and
software tools for new tools and systems.
Thus, the installed base built upon the dominant platform—reflected in durable goods
and human capital (training) specific to the old standard—can create an inertia that
makes it much more difficult for any one producer to break away from the prevailing
standard by introducing a noncompatible product, even if the new standard offers a

significant technological improvement over the current standard (Farrell and Saloner,
1986). In this way, network externalities can retard innovation and slow or prevent adop-
tion of improved product standards.
Therefore, companies seeking to leapfrog a widely adopted standard face substantial
risk. They must not only invent an improved platform, but they must also devise and
execute a successful strategy to migrate consumers from the dominant platform. They also
face the challenge of encouraging other software and complementary product developers
to build for the new platform. One strategy is to steeply discount the costs of the new plat-
form or provide free access. This strategy is not sustainable unless the platform developer
has ancillary revenue streams—such as bundled advertising or ties to other products and
services—to cover their research, development, product, and support costs.
Intellectual property protection can contribute to and alleviate the network externality
dilemma. On the one hand, intellectual property protection for the network features of
computer technology can discourage realization of positive network externalities by
limiting access to network technologies. The sponsor of a particular network technology
can use intellectual property protection to exclude competitors or charge a high licensing
fee for access, thereby raising costs. The intellectual property owner can also limit innova-
tion by restricting how the network technology evolves. On the other hand, intellectual
property protection can provide valuable incentives for overcoming bandwagon effects
that entrench obsolete standards (Menell, 1987, p. 1343). Without the potential for a large
reward, inventors contemplating innovative new platforms might not be willing to make
the substantial, risky R&D and marketing investments needed to challenge, and hopefully
leapfrog, the incumbent platform.
These considerations suggest three principles for intellectual property protection of APIs
and other functional features of platform technologies: (A) a parsimony principle to prevent
firms from establishing protection for product standards without providing a significant
technological advance; (B) a proportionality principle to ensure that firms can appropriate
a fair return on technological advances in platform innovation sufficient to overcome the
excess inertia of network markets, but not so large as to stunt network externalities; and (C)
a deterrence principle to discourage deceptive practices and overreach in network markets.
A. Parsimony Principle: No Intellectual Property Protection for Functional Attributes

Absent Significant Technological Advance
Consumers benefit from access to platforms that produce network benefits. Those
benefits can increase over time through positive feedback effects and the development of
aftermarket enhancements and complementary products. The incentives for firms adopt-
ing product standards, however, are distorted. New entrants might choose an incompat-
ible standard to differentiate their products from established brands, even where growing
the established network would enhance consumer welfare (Katz and Shapiro, 1985, 1986).
Intellectual property protection affects such choices by setting the ground rules for
establishing proprietary platforms. Firms will be more inclined to build competing plat-
forms where the thresholds for acquisition of intellectual property protection—and hence
the power to exclude subsequent entrants and those seeking to bridge platforms—are low.
Thus, intellectual property regimes should discourage platform adoption choices that
undermine realization of network externalities unless there is a large countervailing ben-

efit, such as substantial technological advance. Affording meaningful intellectual property

protection for network technologies without requiring a significant technological advance
encourages wasteful differentiation and increases the risk of undeserved monopoly
power. With easy access to intellectual property protections—for example, by merely
using arbitrary lock-out codes—firms can fragment platforms that would otherwise foster
competition in the non-network product features and in downstream products competing
on the platform. Through serendipity, first-mover advantage, clever marketing, or simply
luck, market power can emerge through positive feedback effects without discernible
consumer benefits. Therefore, intellectual property law should not simply reward novel
(but obvious) or expressive functional features of network goods or services. Rather,
strong protection should be reserved for substantial advances.
B. Proportionality Principle: Overcoming Excess Inertia without Undue Protection
While low thresholds for intellectual property protection for network technologies under-
mine realization of network externalities, balanced protection for substantial technologi-
cal advances may be necessary for entrants to overcome the strong inertial forces driving
network markets. Switching costs discourage consumers from making the leap to a new
platform. For network products and services, those costs can be particularly high due to
network effects. The leap is likely not worth the cost for modest technological improve-
ments. At some point, however, overall consumer welfare will be enhanced by migration
to an alternative platform. The efficient tipping point depends on R&D and marketing
costs as well as the contours of consumer demand.
The excess inertia of network effects can hinder, delay, and possibly prevent the techno-
logical shift to a substantially more advanced technological platform. If all such advances
were freely available to entrants, the free-rider problem would discourage the needed
R&D and marketing investment needed to displace the obsolete platform. Yet providing
strong intellectual property protection for such advances can lead to robust returns as the
market tips to the new platform.
The shift from ‘feature phones’—mobile phones ‘featuring’ voice and text messaging
with rudimentary Internet access—to true ‘smartphones’ with email and robust web
functionality illustrates the challenges and opportunities surrounding network markets.
Through the 1990s, Motorola, Nokia, and a few other vendors established the first
generation of mobile devices. Sun, Microsoft, and Symbian vied to establish the platform
for mobile devices that integrated email and Internet capabilities. By 2005, Java’s Micro
Edition (ME) was faring well, with adoption by Palm and Blackberry.
As the first-generation smartphone battle was resolving, Apple was secretly investing
heavily in an ambitious new platform. Intellectual property played a significant role in
motivating Apple’s R&D. As Steve Jobs noted during the historic January 2007 iPhone
announcement, ‘boy have we patented it!’
Meanwhile, Google was at work on its own skunkworks1 smartphone play: the Android
1
Derived from Lockheed’s code-named secret World War II project to develop a new fighter
jet (‘Skunk Works’), which was taken from Al Capp’s Li’l Abner comic strip, a ‘skunkworks’ project
brings together a small group of highly skilled researchers to pursue radical innovations.

smartphone platform. Given Google’s concern that its success in search and online adver-
tising could be displaced if Microsoft or Apple gained dominance in the shift to mobile
devices, Google sought to develop an open platform that would perpetuate Google’s
search and other services on mobile devices (Menell, 2018).
This standards war illustrates the dynamism of network markets as well as the complex
role of intellectual property protection. In the space of just a few years, the market shifted
dramatically from feature phones to rudimentary smartphones and then to advanced
smartphones. By 2011, Apple and Google dominated the market. Intellectual property
protection played a central role in encouraging investment, but also resulted in massive
resources devoted to intellectual property acquisition, coalition building, standard setting
on upstream technologies, and litigation.
There is no simple answer to the question of how much protection is enough, especially
given the range of business strategies, institutions, and intellectual property regimes that
can deliver appropriate returns on investment, the dynamism of network markets, and
concerns about anti-competitive leveraging network technology dominance. Lichtman
(2000) emphasizes strong property rights to promote platform competition, but this analy-
sis assumes low transaction costs, overlooks consumers’ cognitive limitations stemming
from lock-in, and risks leveraging monopoly power and inhibiting cumulative innovation.
The optimal level of intellectual property protection has a dynamic quality, with the level
of protection dissipating as network technologies and platforms become dominant. Menell
(1987) recommends a limited patent-type regime to protect the functional features of com-
puter software, although with shorter duration and more flexibility to promote access to
platforms that become widely adopted. Menell (1989) advocates a genericide-type doctrine,2
which could protect emerging platforms but give way to broader access when a platform
becomes dominant and risks affording the proprietor the ability to leverage that control to
hinder cumulative innovators (Langlois, 2001). This analysis anticipated Microsoft’s rise
and its abusive market tactics in undermining Netscape and Sun. At the same time, scholars
have opposed copyright protection for the functional and interoperable aspects of computer
technology so as to avoid large returns to first movers that win a standards battle without
offering significant technological innovation (Menell, 1987; Karjala, 1987; Samuelson et
al., 1994). Such limitations on copyright protection afford competitors freedom to use and
build on unpatented methods of operation. In some circumstances, compulsory licensing
of patents might be desirable. This can be achieved through injunctive relief.
The proportionality principle ensures that platform innovators who choose proprietary
strategies (as opposed to more collaborative approaches) have the potential to reap sig-
nificant rewards if they prevail in a standards competition, but that their ability to control
the platform (and charge monopoly prices) declines as the network becomes entrenched.
Such a regime creates optimal conditions for overcoming excess inertia while promoting
the realization of network benefits. It also allows for competition to enhance and improve
established platforms.
2
A trademark can become generic and thereby lose protection if it becomes associated in
the public’s mind with a category of product rather than the source of a particular brand of the
product. See, e.g., The Murphy Door Bed Co., Inc. v. Interior Sleep Systems, Inc., 874 F.2d 95 (2d
Cir. 1989) (‘Murphy bed’ for a bed that folds up into a wall cabinet); King-Seely Thermos Co. v.
Aladdin Indus., Inc., 321 F.2d 577 (2d Cir. 1963) (‘Thermos’ vacuum insulated bottle).

C. Deterrence Principle: Discouraging Overreach with Balanced Remedies
Intellectual property law and competition policy should also protect against deceptive
practices and leveraging intellectual property rights to control network markets. The
integrity of standard setting processes is particularly critical to efficient collaboration
among enterprises and innovators working in network industries. The choice of standards
depends on a range of factors, including potential restrictions on practicing technological
standards. Hence, standard setting bodies should require disclosure of all potential intel-
lectual property encumbrances or, at a minimum, advance commitment by SSO members
to licensing such technologies on fair and reasonable terms. Courts should penalize efforts
to reduce transparency in standard setting processes and take failure to abide by such
commitments into consideration in enforcing patent rights.
Antitrust law and competition policy should also take network effects into account in
assessing monopoly power, scrutinizing collaborations and contractual agreements, and
fashioning remedies. The consumer, competitive, and innovation ramifications of network
markets are especially complex. What might appear to be benign and welfare-improving
behaviors—such as integrating a ‘free’ browser into an operating system product or
bundled aftermarket services—might ultimately lead to monopolization of important
emerging and downstream markets. Hence, antitrust law must be vigilant in assessing
the dynamism and path-dependence of network technologies. For example, advance
determination of licenses for SEPs can promote competition in downstream products
and services. In some circumstances, antitrust authorities should tolerate some collusive
behaviors—such as ex ante negotiation of FRAND license rates by SSOs—that resemble
forbidden price-setting. The Sherman Antitrust Act bars contracts and conspiracies that
unreasonably restrain competition. In network markets, some collaboration promotes
economic efficiency.
The crafting of remedies to combat abusive and anti-competitive behavior in network
markets requires careful consideration of effects on consumers and competitors. Once
a standard has taken root and is generating substantial network benefits, traditional
remedies—such as enjoining the offensive activities of breaking up dominant firms—can
cause adverse effects on the consumers who have adopted the standard as well as other
downstream users—such as programmers and competitors who have incurred sunk costs
in joining the platform. Leveraging intellectual property rights to control network markets
might also produce countervailing innovative efficiencies. Hence, antitrust authorities
and courts should consider remedies that promote the realization of network benefits
while promoting enhanced competition and innovation. In some circumstances, these
considerations favor compulsory licenses, which can be flexible and adaptable, over
injunctive remedies.
V. INTELLECTUAL PROPERTY PROTECTION FOR NETWORK

FEATURES
In view of the tremendous economic significance of controlling access to systems

technologies by exploiting demand-side effects and excluding competition in comple-
mentary goods and services, such as repair services, replacement parts, and ancillary

markets (e.g., advertising and consumer data), platform developers and entrepreneurs
have sought to use intellectual property to protect APIs and other means to exclude
competitors from their platforms and systems. As an alternative approach, computer
programmers and a growing number of commercial enterprises in the open source com-
munity have deployed intellectual property protection as a tool for sharing technology
and precluding proprietary control of core Internet and computer operating system
technologies.
Since the principal forms of intellectual property protections developed long before the
advent of digital technology, which made network effects so important, the intellectual
property statutes do not expressly reflect the aforementioned policy principles for APIs
and other functional features of platform technologies. Nonetheless, the mixed statutory/
common law heritage of intellectual property law (Menell, 2013) has afforded courts
discretion to interpret statutory provisions, adapt common law doctrines, and apply
equitable enforcement principles to address network effects. Moreover, more recent
legislation has integrated network economics into intellectual property law (see Section
V(B)(4) exploring the DMCA interoperability exemption).
Although patents have long protected platform technologies, such as electrical stand-
ards (e.g., AC/DC, phonogram, color television, and telecommunications) (Shapiro and
Varian, 1999, pp. 210–23), the contours of intellectual property protection for network
features of systems and platforms centers around software technology. Trade secrecy and
contract law provided relatively effective protection for much of the software developed
during the mainframe and minicomputer eras. And although advances in computer
hardware fell squarely within the patent domain, there were significant doubts about the
patentability of computer software into the 1990s. Hence, as microcomputers emerged,
which spurred retail distribution of computer software, copyright law emerged as the
primary battleground for computer software by the mid-1980s.
This section begins by discussing how trade secrecy can protect the network features of
systems technologies. It then traces the evolution of copyright protection for computer
software. Almost all of the major computer software battles have focused on the extent
to which copyright protection afforded protection to the network features of computer
software. Section C discusses the role of trademark and related protections for network
technologies. Section D examines the role of patent protection for network technologies,
which emerged as a more robust and controversial form of protection for computer
software in the 1990s.
A. Trade Secret Protection
Trade secret protection protects against the misappropriation of confidential informa-

tion that is subject to reasonable efforts to maintain secrecy, such as security and
non-disclosure agreements with employees and contractors (Menell et al., 2017, ch. II).
Trade secret protection can last indefinitely, but once trade secrets become public, they
lose protection.
Trade secret protection came into common usage in the software industry as a tool
for protecting algorithms, software design, and coding—including APIs. Trade secret
protection of passwords is also commonly used today to control access to websites and
cloud servers.

Trade secret protection does not provide absolute protection for information. It only
protects against misappropriation through improper means and unauthorized disclosure.
Therefore, competitors do not violate trade secrecy protection through reverse engineer-
ing of publicly available products and websites. The reverse engineering limitation on
trade secret protection thus exposes the trade secret owner to free riding by others. This
limitation, however, strikes a salutary balance between protection on the one hand and
competition and the dissemination of knowledge on the other (Landes and Posner,
2003; Samuelson and Scotchmer, 2002; Chisum et al., 1989). The trade secret owner can
‘purchase’ greater protection against this risk by investing in higher levels of security (e.g.,
more effective encryption for software-encoded technology). The inventor can also pursue
patent protection, which proscribes reverse engineering, although only for the limited
duration of the patent, and mandates disclosure of the invention to the public. By declin-
ing to pursue patent protection (or failing to satisfy the requirements thereof), however,
inventors should not be able to secure potentially perpetual rights in technologies merely
by encrypting them or otherwise obscuring how they function. To do so would undermine
the larger balance of the federal intellectual property system.
As the next section explains, courts have interpreted copyright law to permit multiple
reproductions of copyrighted software programs as a means for reverse engineering
unprotected (by copyright), but secret, elements of code necessary for interoperability.
B. Copyright Protection
As the proliferation of microcomputers seeded a market for computer programs,

software entrepreneurs saw copyright as an effective strategy to protect their programs
from unauthorized reproduction and distribution. Computer software, however, does
not fit easily within the copyright mold. Copyright law had long denied protection to
functional elements. Although written in text, computer software provides the gears and
levers for digital machines—which fits more naturally within the utility patent system
(cf. Baker v. Selden, 101 U.S. 99, 102 (1879) (‘The claim to an invention or discovery of
an art or manufacture must be subjected to the examination of the Patent Office before
an exclusive right therein can be obtained; and it can only be secured by a patent from
the government’).
The rapid emergence of the computer software marketplace in the early 1970s posed a
dilemma for intellectual property policymakers. Computer software could be expensive
to develop and was easily pirated, creating a severe appropriability problem for the
nascent, yet critical, software industry (Gates, 1976). Patent law, which had long served
as the primary form of protection for technological advances in machines and processes,
was thought to be too costly, time-consuming, stringent, and uncertain as a means for
protecting software products against piracy (Menell, 1987, pp. 1347–51). Copyright law
had long provided an effective means of protecting literary works from piracy, but its
doctrines excluding ideas and functional elements from protection raised serious ques-
tions about its appropriateness for protecting inherently utilitarian works. Copyright’s
low threshold for protection (mere originality), broad array of rights (including the right
to adapt), and long duration created a high risk of overbroad protection for computer
software products, in direct opposition to the parsimony principle. On the other hand,
copyright law’s limiting principles, such as the idea-expression dichotomy (denying

copyright protection to expression that encumbers the use of ideas) and the fair use
doctrine, provided tools for aligning copyright protection with the parsimony principle.
The interplay of copyright protection and network effects has played out on several
fronts during the past four decades. Section 1 explains the principal legislation undergird-
ing copyright protection for computer software. Section 2 traces the development of soft-
ware copyright jurisprudence relating to APIs through 2010. Section 3 explores software
licensing and the emergence and growth of the free and open source movements—key
drivers of network technology markets. Section 4 explores the interoperability excep-
tion to the anti-circumvention provisions added to the copyright law in 1998. Section 5
picks up where Section 2 left off by examining the Oracle v. Google litigation. Section 6
examines copyright protection for standards and codes.
1. Software copyright legislation: the Copyright Act of 1976, the CONTU Final Report,
and the 1980 amendments
The software protection controversy of the early 1970s emerged at an inopportune time.
Congress had been working for nearly two decades to overhaul the Copyright Act of
1909 and was nearing closure in the early to mid-1970s. Faced with the challenge of
fitting computer and other new information technologies under the existing umbrella
of intellectual property protection, Congress established the National Commission on
New Technological Uses of Copyrighted Works (CONTU) to study the implications of
the new technologies and recommend revisions to federal intellectual property law. As a
stopgap, Congress included computer software within the scope of ‘literary works’ in the
Copyright Act of 1976 (1976 Act). The House Report explains that
[t]he term ‘literary works’ does not connote any criterion of literary merit or qualitative value: it
includes catalogs, directories, and similar factual, reference, or instructional works and compila-
tions of data. It also includes computer data bases, and computer programs to the extent that they
incorporate authorship in the programmer’s expression of original ideas, as distinguished from
the ideas themselves. (H.R. Rep. No. 94-1476, 94th Cong., 2d Sess. 54 (1976) (emphasis added))
Other provisions of the 1976 Act, however, maintained traditional exclusions for ideas
and functional features. (17 U.S.C. § 102(b) (‘In no case does copyright protection for an
original work of authorship extend to any idea, procedure, process, system, method of
operation, concept, principle, or discovery’)).
The CONTU Final Report concluded that copyright law should protect the intellectual
work embodied in computer software, notwithstanding the fundamental principle that
copyright cannot protect ‘any idea, procedure, process, system, method of opera-
tion, concept, principle, or discovery’ and the Supreme Court’s foundational decision
in Baker v. Selden (CONTU, 1979, p. 1). Nonetheless, CONTU recommended that
Congress immunize rightful possessors of a computer program from liability for using
the program (which typically results in reproduction of computer code) and making a
backup copy of computer programs, which Congress largely adopted in 1980 legislation
(Act to amend the patent and trademark laws, Pub. L. No. 96-517, 94 Stat. 3015, 3028
(1980) (codified at 17 U.S.C. §§ 101, 117)).
In keeping with copyright law’s fundamental limiting principles, the CONTU Final
Report explained that while ‘one is always free to make a machine perform any conceiv-
able process (in the absence of a patent), [] one is not free to take another’s program,’

subject to copyright’s limiting doctrines—originality and the idea/expression dichotomy

(CONTU, 1979, p. 20). The Report further explained that
The ‘idea-expression identity’ exception provides that copyrighted language may be copied
without infringing when there is but a limited number of ways to express a given idea. This
rule is the logical extension of the fundamental principle that copyright cannot protect ideas.
In the computer context this means that when specific instructions, even though previously
copyrighted, are the only and essential means of accomplishing a given task, their later use by
another will not amount to an infringement. (CONTU, 1979, p. 20 (footnote omitted))
Thus, while recognizing important limitations on copyright protection for computer soft-
ware, including the § 102(b) limitations, Congress intended that software programmers
would garner protection for their programming design and coding choices to the extent
that the expression was separable from the underlying ideas. In this way, the general pro-
gramming ideas and unoriginal programming choices remain free for others to use while
the creative effort in particularized programming choices and compilations, especially in
complex programs, gains protection from copyists.
2. Software copyright jurisprudence: the first wave

The 1976 Copyright Act, as well as the CONTU Final Report, pushed the availability and
scope of copyright protection for computer software to the courts. The treatment of APIs
under copyright law emerged over the next two decades as courts interpreted and applied
the § 102(b) limitations (including the idea-expression dichotomy), infringement standards,
the fair use defense, and other legal doctrines standards. Courts confronted battles across
various software markets—from microcomputer operating systems to job scheduling
software for mainframe computers, mobile phone networks, computer-user interfaces,
video game devices, printer cartridges, garage door openers, and all manner of application
programs (such as business systems, design programs, video games, and spreadsheets).
Nearly every major software copyright litigation involved interoperability elements.
After an inauspicious start, the federal courts implemented a balanced framework for
both protecting computer software against piracy and interpreting the idea/expression
doctrine to ensure that copyright law excludes functional features of computer technol-
ogy (Menell, 1998). These decisions effectuated the subtle balance to which the CONTU
Final Report referred. The courts came to appreciate that ‘creativity’ must be understood
contextually. While programming a computer can unquestionably be termed ‘creative’ in
a general sense, it is not necessarily ‘creative’ in a copyright sense. Just as the design of an
efficient mechanical machine can be creative, such devices are not eligible for copyright
protection unless the aesthetic features can be separated from the functional attributes
(17 U.S.C. § 101 (‘“Pictorial, graphic, and sculptural works” include two-dimensional
and three-dimensional works . . .; the design of a useful article . . . shall be considered
a pictorial, graphic, or sculptural work only if, and only to the extent that, such design
incorporates pictorial, graphic, or sculptural features that can be identified separately
from, and are capable of existing independently of, the utilitarian aspects of the article’).
Lines of code are the gears and levers of digital machines. The fact that computer
software, like a sculptural work, is eligible for copyright protection does not authorize
protection for functional features.
The courts came to recognize that APIs have significant functional dimensions.
They serve in many contexts as the basis for interoperability of computer technologies.

The First Circuit held that the particular functional specifications, as opposed to the
implementing code, can be fairly characterized as ‘methods of operation.’ Although
the Supreme Court’s split decision in Lotus v. Borland (Lotus Dev. Corp. v. Borland
Int’l, Inc., 516 U.S. 233 (1996) (affirming, without opinion by an equally divided vote,
the First Circuit’s decision holding that the menu command structure for a spreadsheet
is an uncopyrightable method of operation under § 102(b))) left some uncertainty, the
resolution of that litigation marked the end of the major API copyright litigations that
had raged since the early 1980s.
This section traces that evolution. Section (i) examines the emergence of jurisprudence
excluding functional and network features of computer software. Section (ii) explores
the related issue of whether competitors can reproduce computer software as a means of
learning unprotectable code elements.
i. Unprotectability of functional and network features The first major cases to address
copyright protection for interoperable features of computer software pitted Apple
Computer Corporation, then a young, break-out microcomputer company, against
cavalier, unscrupulous competitors offering discount ‘interoperable’ Apple II clones
(Apple Computer, Inc. v. Franklin Computer Corp., 545 F. Supp. 812 (E.D. Pa. 1982),
rev’d, 714 F.2d 1240 (3d Cir. 1983); Apple Computer, Inc. v. Formula Int’l, Inc., 562
F. Supp. 775 (C.D. Cal. 1983), aff’d, 725 F.2d 521 (9th Cir. 1984)). The clone makers
quickly entered the market by simply copying, bit by bit, Apple’s operating system and
application programs.
The defendants in these cases argued that copyright protection did not extend to non-
human readable (object code) formats of computer software and that the idea-expression
doctrine barred copyright protection for operating system programs. They further
argued that copyright protection should not stand in the way of their selling computers
that can run programs written for the Apple II. The courts had little trouble validating
Apple’s complaint that verbatim copying of millions of bits of code constituted copyright
infringement. The 1976 Act, in conjunction with the CONTU Final Report, clearly
extended copyright protection in this circumstance.
Unfortunately, the Third Circuit’s decision included language suggesting that copyright
protection could encompass the functional requirements for interoperability: ‘total
compatibility with independently developed application programs . . . is a commercial
and competitive objective which does not enter into the somewhat metaphysical issue of
whether particular ideas and expressions have merged’ (Apple Computer, Inc. v. Franklin
Computer Corp., 714 F.2d at 1253 (3d Cir. 1983)). Since two entirely different programs
can achieve the same ‘certain result[s]’—for example, generate the same set of protocols
needed for interoperability—the court was not justified in making such an expansive state-
ment about the scope of copyright protection for computer program elements. CONTU
was clear that ‘[o]ne is always free to make the machine do the same thing as it would
if it had the copyrighted work placed in it, but only by one’s own creative effort rather
than by piracy’ (CONTU, 1979, p. 21). Given the verbatim copying of millions of bits of
object code, there was no need to address the interoperability issue. The defendant offered
no explanation for which elements of the program were protectable and which were not.
The Third Circuit’s decision in Whelan Associates, Inc. v. Jaslow Dental Laboratory,
Inc., 797 F.2d 1222 (3d Cir. 1986), further expanded copyright protection for computer

software. In that case, Jaslow Dental Laboratory had hired Whelan Associates, a custom
software company, to develop a computer program to organize its bookkeeping and
administrative tasks. When Jaslow developed and marketed its own program for manag-
ing a dental laboratory, Whelan sued Jaslow for copyright infringement. The evidence
at trial showed that, although Jaslow had not literally copied Whelan’s code, there were
overall structural similarities between the two programs. As a means of distinguishing
protectable expression from unprotectable idea, the court reasoned:
[T]he purpose or function of a utilitarian work would be the work’s idea, and everything that is not
necessary to that purpose or function would be part of the expression of the idea. Where there are
many means of achieving the desired purpose, then the particular means chosen is not necessary
to the purpose; hence, there is expression, not idea (797 F.2d at 1236 (emphasis in original;
citations omitted)).
In applying this rule, the court defined the idea as ‘the efficient management of a dental
laboratory,’ which countless programs could express (797 F.2d at 1236, n. 28). Drawing the
idea/expression dichotomy at such a high level of abstraction implied an expansive scope
of copyright protection. Although the case did not directly address copyright protection
for interoperable features of computer code, the court’s mode of analysis expanded the
scope of copyright protection to all aspects of computer programs. If everything below
the general purpose of the program were protectable under copyright law, then it would
follow that particular protocols were protectable because there would be other ways
to accomplish the program’s same general purpose. Such a result would effectively bar
competitors from developing interoperable programs and computer systems.
Commentators roundly criticized the Whelan test (Chisum et al., 1989, pp. 20–21;
Menell, 1989, p. 1074; Englund, 1990, p. 881), and other courts developed alternative
approaches. A few months after Whelan, the Fifth Circuit confronted a similar claim of
copyright infringement based upon structural similarities between two programs designed
to provide cotton growers with information regarding cotton prices and availability,
accounting services, and a means for conducting cotton transactions electronically (Plains
Cotton Coop. Ass’n v. Goodpasture Computer Serv., Inc., 807 F.2d 1256 (5th Cir. 1987)).
In declining to follow the Whelan approach, the court found that the similarities in the
programs were dictated largely by standard practices in the cotton market—what the
court called ‘externalities’—such as the ‘cotton recap sheet’ for summarizing basic trans-
action information, which constitute unprotectable ideas. The court found persuasive the
decision in Synercom Technology, Inc. v. University Computing Co., 462 F. Supp. 1003,
1013 (N.D. Tex. 1978), which analogized the ‘input formats’ of a computer program (the
organization and configuration of information to be inputted into a computer) to the
‘figure-H’ pattern of an automobile stick shift.
Drawing on the Fifth Circuit’s approach and Judge Learned Hand’s foundational test
for analyzing copyright infringement (Nichols v. Universal Pictures Corp., 45 F.2d 119
(2d Cir. 1930)), the Second Circuit crafted what has become the leading framework for
analysing infringement of computer software code (Computer Assocs. Int’l v. Altai, Inc.,
982 F.2d 693 (2d Cir. 1992)). Computer Associates (CA), a leading mainframe software
provider, had developed a job scheduling program (SCHEDULER) for IBM mainframe
computers. Part of the success of this program was that it had a sub-component
(ADAPTER) which interoperated with any of the three IBM mainframes. Thus, the user

did not need to customize its programs for each of the IBM mainframes. CA’s ADAPTER
program ensured that programs written for SCHEDULER would run on any of the three
IBM mainframes.
CA sued Altai, a competitor that pursued a similar strategy for designing its job
scheduling software for the IBM mainframes. Unbeknownst to Altai’s management, one
of its key programmers copied 30 percent of ADAPTER code into Altai’s job scheduling
software product. When Altai management learned of the copying, the company initiated
a ‘clean room’ process to insulate its programmers from copyright-protected code so as
to ensure that the resulting program interoperated with the IBM mainframes without
copying any ADAPTER code (Sammi et al., 2013).
Altai accepted responsibility for copyright infringement based on the early version.
Nonetheless, drawing on the Third Circuit’s Whelan decision, CA claimed that the clean
room version was also infringing due to structural similarities at various levels, such
as flow charts, inter-modular relationships, parameter lists, and macros. The Second
Circuit rejected Whelan’s approach. The Second Circuit fleshed out a detailed analytical
framework for determining copyright infringement of computer code:
In ascertaining substantial similarity . . . a court would first break down the allegedly infringed
program into its constituent structural parts. Then, by examining each of these parts for such
things as incorporated ideas, expression that is necessarily incidental to those ideas, and elements
that are taken from the public domain, a court would then be able to sift out all non-protectable
material. Left with a kernel, or perhaps kernels, of creative expression after following this process
of elimination, the court’s last step would be to compare this material with the structure of an
allegedly infringing program (982 F.2d at 706).
The court’s ‘abstraction-filtration-comparison’ test recognized that an idea could exist

at multiple levels of a computer program and not solely at the most abstract level.
Furthermore, the ultimate comparison is not between the programs in their entirety. Rather,
courts must focus solely on whether protectable elements of the program were copied. Of
most importance for fostering interoperability, the court held that copyright protection did
not extend to those program elements where the programmer’s ‘freedom to choose’ is
circumscribed by extrinsic considerations such as (1) the mechanical specifications of the
computer on which a particular program is intended to run; (2) compatibility requirements
of other programs with which a program is designed to operate in conjunction; (3) computer
manufacturers’ design standards; (4) demands of the industry being serviced; and (5) widely
accepted programming practices within the computer industry (982 F.2d at 709–10).
Directly rejecting the Third Circuit’s dictum in Apple v. Franklin that achieving ‘total
compatibility with independently developed application programs . . . is a commercial
and competitive objective which does not enter into the somewhat metaphysical issue of
whether particular ideas and expressions have merged,’ the Second Circuit recognized that
external factors such as interface specifications, de facto industry standards, and accepted
programming practices are not protectable under copyright law. The formulation of the
Second Circuit test judges these external factors when the allegedly infringing activities
(i.e., ex post) occur, not when the first program is written. The court emphasized that the
first company to write a program for a particular application should not be able to ‘“lock
up” basic programming techniques as implemented in programs to perform particular
tasks’ (982 F.2d at 712 (quoting Menell, 1989, p. 1087)).

Other circuits embraced the Second Circuit’s Altai framework (Gates Rubber Co. v.
Bando Chem. Indus., Ltd., 9 F.3d 823, 836–43 (10th Cir. 1993); Eng’g Dynamics, Inc. v.
Structural Software, Inc., 26 F.3d 1335 (5th Cir. 1994); Apple Computer, Inc. v. Microsoft
Corp., 35 F.3d 1435 (9th Cir. 1994); Bateman v. Mnemonics, Inc., 79 F.3d 1532, 1547
(11th Cir. 1996); Mitel, Inc. v. Iqtel, Inc., 124 F.3d 1366 (10th Cir. 1997)). The Altai case
addressed programmers’ freedom to write code to interoperate with externally established
APIs—in that case by IBM. IBM had not challenged CA’s or Altai’s use of its interface
specifications. It welcomed other companies to develop software for its mainframes. Thus,
the case did not specifically address whether the API developer could assert a copyright
infringement claim based on unauthorized use of their interface specifications. That issue
would emerge in a series of cases involving video games and spreadsheets.
The Ninth Circuit’s decision in Sega Enterprises Ltd. v. Accolade, Inc., 977 F.2d 1510
(9th Cir. 1992), expressly recognized the legitimacy of deciphering and copying particular
lock-out codes for purposes of developing interoperable products. Sega developed a
successful video game platform (Genesis) for which it licensed access to video game
developers. Accolade, a video game manufacturer, wanted to distribute versions of its
game on the Genesis platform. It did not, however, want to limit distribution exclusively
to Genesis, as Sega required. Rather than license access to Sega’s code, Accolade reverse
engineered the access code through a painstaking effort that entailed making hundreds
of intermediate copies of Sega’s computer code. Accolade then incorporated only those
code elements (approximately 25 bytes in games containing between 500,000 and 1.5
million bytes) that were necessary to achieve interoperability with the Genesis platform
into Accolade game cartridges.
Sega sued Accolade for copyright infringement. Given the relatively small amount
of Sega code in the Accolade game cartridges, Sega focused its copyright claim on
Accolade’s reproduction of the entirety of Sega’s program code for purposes of isolating
those code elements needed to interoperate with the Genesis console. The district court
rejected Accolade’s argument that such intermediate copies—made solely for the purpose
of reverse engineering the platform—constituted fair use and granted a preliminary
injunction. The Ninth Circuit held that ‘disassembly of object code in order to gain an
understanding of the ideas and functional concepts embodied in the code is a fair use that
is privileged by section 107 of the Act’ (977 F.2d at 1518). Balancing these factors, the
Ninth Circuit ruled that ‘the functional requirements for compatibility with the Genesis
[video game console are] aspects of Sega’s programs that are not protected by copyright’
(977 F.2d at 1522 (citing 17 U.S.C. § 102(b))). In effect, the court held that copyright law
does not protect the particular code or process needed for interoperating with a copy-
righted computer program (such as lock-out code). The Ninth Circuit reaffirmed and
expanded the Sega decision in Sony Computer Entertainment, Inc. v. Connectix Corp.,
203 F.3d 596 (9th Cir. 2000).
The Northern District of California and the Ninth Circuit applied the Altai framework
to the graphical user interface features of a computer program in Apple Computer, Inc. v.
Microsoft Corp., 799 F. Supp. 1006 (N.D. Cal. 1992), aff’d in part, rev’d in part, 35 F.2d
1435 (9th Cir. 1994). Apple alleged that Microsoft’s Windows operating system infringed
copyrights in the desktop graphical user interface of its Macintosh computer system.
A licensing agreement authorizing Microsoft to use aspects of Apple’s graphical user
interface muddied the copyright issue. The court determined, however, that the licensing

agreement was not a complete defense to the copyright claims and therefore undertook
an analysis of the scope of copyright protection for a large range of audiovisual elements
of computer screen displays.
In framing the analysis, the district court expressly recognized the relevance of network
externalities and the cumulative nature of innovation to the scope of copyright protection:
Copyright’s purpose is to overcome the public goods externality resulting from the non-
excludability of copier/free riders who do not pay the costs of creation. Peter S. Menell, An
Analysis of the Scope of Copyright Protection for Application Programs, 41 Stan. L. Rev. 1045,
1059 (1989). But overly inclusive copyright protection can produce its own negative effects by
inhibiting the adoption of compatible standards (and reducing so-called ‘network externalities’).
Such standards in a graphical user interface would enlarge the market for computers by making
it easier to learn how to use them. Id. at 1067–70. Striking the balance between these consid-
erations, especially in a new and rapidly changing medium such as computer screen displays,
represents a most ambitious enterprise. Cf. Lotus Dev. Corp. v. Paperback Software Int’l, 740 F.
Supp. 37 (D. Mass. 1990).
While the Macintosh interface may be the fruit of considerable effort by its designers, its
success is the result of a host of factors, including the decision to use the Motorola 68000 micro-
processor, the tactical decision to require uniform application interfaces, and the Macintosh’s
notable advertising. And even were Apple to isolate that part of its interface’s success owing to its
design efforts, lengthy and concerted effort alone ‘does not always result in inherently protectible
expression.’ [quoting Computer Assocs. Int’l v. Altai, Inc. 982 F.2d at 711]
By virtue of having been the first commercially successful programmer to put these general-
ized features together, Apple had several years of market dominance in graphical user interfaces
until Microsoft introduced Windows 3.0, the first DOS-based windowing program to begin to
rival the graphical capability of the Macintosh. . ..To accept Apple’s ‘desktop metaphor’/‘look
and feel’ arguments would allow it to sweep within its proprietary embrace not only Windows
and NewWave but, at its option, also other desktop graphical user interfaces which employ the
standardized features of such interfaces, and to do this without subjecting Apple’s claims of
copyright to the scrutiny which courts have historically employed. Apple’s copyrights would
hold for programs in existence now or in the future—for decades. One need not profess to
know for sure where should lie the line between expression and idea, between protection and
competition to sense with confidence that this would afford too much protection and yield too
little competition.
The importance of such competition, and thus improvements or extensions of past expres-
sions, should not be minimized. The Ninth Circuit has long shown concern about the uneasy
balance which copyright seeks to strike: ‘What is basically at stake is the extent of the copyright
owner’s monopoly—from how large an area of activity did Congress intend to allow the
copyright owner to exclude others?’ 799 F. Supp. at 1025–26 (quoting Herbert Rosenthal Jewelry
Corp. v. Kalpakian, 446 F.2d 738, 742 (9th Cir.1971)).
The court found that most of the similar icons between Apple’s graphical user interface
and Microsoft’s Windows that were not authorized by the licensing agreements were
either not lacking originality or subject to one or more of copyright’s limiting doctrines.
Drawing on the principle that compilations of largely uncopyrightable elements are
only protected against ‘bodily appropriation of expression’ (see Harper House, Inc.
v. Thomas Nelson, Inc., 889 F.2d 197, 205 (9th Cir. 1989)) the court applied a ‘virtual
identity’ standard to compare the works as a whole and determined that no infringement
had occurred. On appeal, the Ninth Circuit affirmed the district court’s dissection of the
works to determine which elements are protectable, its filtering of unprotectable elements,
and its application of the ‘virtual identity’ standard.
The copyrightability of command systems for computer software arose most directly in

Worksheet Range Copy Move File Print Graph Data Quit
Global Insert Delete Column-Width Erase Titles Window Status
Format Label-prefix Column-Width Recalculation Protection Default Zero
Natural Columnwise Rowwise Automatic Annual Iteration
Figure 7.1 Lotus 1-2-3 menu command hierarchy
litigation surrounding spreadsheet technology. Building upon the success of the VisiCalc
program developed for the Apple II computer, Lotus Corporation marketed an enhanced
operating spreadsheet program incorporating many of VisiCalc’s features and commands
into its 1-2-3 program for the IBM PC platform. Lotus 1-2-3 quickly became the market
leader for spreadsheets running on IBM and IBM-compatible machines, and knowledge
of the program became a valuable skill in the accounting and management fields. The
1-2-3 command hierarchy was particularly attractive because it logically structured more
than 200 commands (see Figure 7.1). Users could create custom programs (called macros)
to automate particular accounting and business planning tasks. Businesses and users
increasingly became ‘locked-in’ to the 1-2-3 command structure as they invested time to
learn the system and their libraries of macros grew (Gandal, 1994). By the late 1980s,
software developers seeking to enter the spreadsheet market could not ignore the large
premiums that consumers placed on their investments in the 1-2-3 system (Menell, 1998).
After three years of intensive development efforts, Borland International, developer
of several successful software products including Turbo Pascal and Sidekick, introduced
Quattro Pro, its entry into the spreadsheet market. Quattro Pro offered improved design
and graphics over Lotus 1-2-3. Computer magazines praised its innovation. Quattro
Pro offered a new interface for its users, which many preferred over the 1-2-3 interface.
Nonetheless, because of the large number of users already familiar with the 1-2-3
command structure and those who had made substantial investments in developing
1-2-3 macros, Borland considered it essential to offer an operational mode based on the
1-2-3 command structure as well as macro compatibility. Nonetheless, Borland’s visual
representation of the 1-2-3 command mode substantially differed from the 1-2-3 screen
displays.
Lotus sued Borland for copyright infringement based on Quattro Pro’s emulation of
the 1-2-3 menu command hierarchy. The First Circuit viewed the case as one of first
impression: ‘[w]hether a computer menu command hierarchy constitutes copyrightable
subject matter.’ (Lotus Dev. Corp. v. Borland Int’l, Inc., 49 F.3d 807, 813 (1st Cir. 1995)).
The court distinguished Altai as dealing with protection of computer code as opposed to
the results of such code. Instead, the First Circuit saw the subject matter of the Lotus case
as a ‘method of operation’ falling directly within the exclusions from copyright protection
set forth in 17 U.S.C. § 102(b). The court held the Lotus menu command hierarchy is an
uncopyrightable ‘method of operation.’

The Lotus menu command hierarchy provides the means by which users control and operate
Lotus 1-2-3. If users wish to copy material, for example, they use the ‘Copy’ command. If users
wish to print material, they use the ‘Print’ command. Users must use the command terms to tell
the computer what to do. Without the menu command hierarchy, users would not be able to
access and control, or indeed make use of, Lotus 1-2-3’s functional capabilities.
The Lotus menu command hierarchy does not merely explain and present Lotus 1-2-3’s
functional capabilities to the user; it also serves as the method by which the program is operated
and controlled. . ..(49 F.3d. at 815)
The US Supreme Court affirmed without opinion by an equally divided vote (Lotus Dev.
Corp. v. Borland Int’l, Inc., 516 U.S. 233 (1996)).
Subsequent appellate decisions reached similar outcomes, although they did not
fully adopt the First Circuit’s categorical exclusion of menu command hierarchies from
copyright protection. In MiTek Holdings, Inc. v. ARCE Engineering Co., 89 F.3d 1548
(11th Cir. 1996), the holder of a copyright in an application program that designed and
arranged wood trusses for the framing of building roofs brought an infringement action
against the maker of a competing program that featured a similar menu command tree
and user interface. Affirming the lower court’s decision, the Eleventh Circuit held that the
menu and submenu command structure of the truss design program was uncopyrightable
under § 102(b) of the 1976 Act because it represents a process. The court did not need to
reach the broader question, addressed in Lotus, of whether all menu command structures
are uncopyrightable as a matter of law.
In Mitel, Inc. v. Iqtel, Inc., 124 F.3d 1366 (10th Cir. 1997), Mitel, the maker of a widely
adopted computer system for automating the selection of a particular telephone long
distance carrier and remotely activating optional telecommunications features such as
speed dialing, sued Iqtel, a competing firm that used the identical command codes, for
copyright infringement. Because Mitel’s system had become a de facto standard, Iqtel
defended its use of compatible controller codes on the ground that ‘technicians who install
call controllers would be unwilling to learn Iqtel’s new set of instructions in addition to
the Mitel command code set, and the technician’s employers would be unwilling to bear
the cost of additional training.’ As Borland had done, Iqtel’s product included both its
own set of command codes as well as a ‘Mitel Translation Mode.’ While commenting
that a method of operation may in some circumstances contain copyrightable expres-
sion, the Tenth Circuit nonetheless concluded that the Mitel command codes, which
were arbitrarily assigned, lacked the minimal degree of creativity necessary to qualify
for copyright protection. The court further held that Mitel’s command codes should be
denied copyright protection under the scènes à faire doctrine because external factors such
as compatibility requirements and industry practices largely dictated the codes.
There were no further cases reported addressing copyright protection for APIs over the
next 15 years. We address the Federal Circuit’s decision upholding copyright protection
for APIs in the Oracle v. Google case in subsection 5.
ii. Permissibility of reverse engineering As discussed in section V(A), network system

developers can use encryption and trade secret law to protect computer code. Distributing
computer programs in object code (binary) format typically constitutes a reasonable effort
to maintain secrecy. As noted, however, competitors can lawfully gain access to such infor-
mation through reverse engineering. One such method is to experiment with object code

to determine which bits are necessary for interoperability. Such forensic work typically
requires the investigator to make many copies, raising the risk of copyright infringement.
The LaST Frontier Final Report (Chisum et al., 1989), a consensus statement of
leading intellectual property scholars, opined that ‘limited copying of programs for the
purpose of examination and study . . . falls within the rigorous terms of the fair use
provisions in section 107 of the Copyright Act’ (Chisum et al., 1989, p. 25; Samuelson
and Scotchmer, 2002). In addition to holding that computer code necessary for interoper-
ability is unprotectable under § 102(b), the Ninth Circuit’s Sega decision authorized the
copying of entire computer programs for purposes of deciphering unprotectible code
elements. In explaining why disassembly and reproduction of object code constitute fair
use, the court reasoned that the ‘functional specifications’ of a computer program are
unprotectable. The Ninth Circuit based its analysis on the architecture of the intellectual
property system:
[D]isassembly of the object code in Sega’s video game cartridges was necessary in order to
understand the functional requirements for Genesis compatibility. The interface procedures for
the Genesis console are distributed for public use only in object code form, and are not visible
to the user during operation of the video game program. Because object code cannot be read by
humans, it must be disassembled, either by hand or by machine. If disassembly of copyrighted
object code is per se an unfair use, the owner of the copyright gains a de facto monopoly over
the functional aspects of his work—aspects that were expressly denied copyright protection by
Congress. 17 U.S.C. § 102(b). In order to enjoy a lawful monopoly over the idea or functional
principle underlying a work, the creator of the work must satisfy the more stringent standards
imposed by the patent laws. Bonito Boats, Inc. v. Thunder Craft Boats, Inc., 489 U.S. 141, 159–64
(1989). Sega does not hold a patent on the Genesis console. (Sega Enterprises Ltd. v. Accolade,
Inc., 977 F.2d 1510, 1526 (9th Cir. 1993))
The Ninth Circuit reaffirmed and expanded the Sega analysis in Sony Computer
Entertainment, Inc. v. Connectix Corp., 203 F.3d 596 (9th Cir. 2000).
3. Software licensing

Copyright law grants authors exclusive rights to copy, adapt, distribute, publicly perform,
and publicly display protected works, subject to various limitations. The early computer
industry, however, did not rely on proprietary control over their customers’ use or adapta-
tion of their software programs. Nor did companies restrict customers’ access to source
code. Rather, the industry—led by IBM and followed by Burroughs, UNIVAC, NCR,
Control Data, General Electric, and RCA (often referred to as the ‘Seven Dwarfs’ due to
IBM’s dominance in the computer industry)—bundled software with their mainframes
and derived revenues from leasing computer usage and sales of complementary products
and services (Ceruzzi, 2003, ch. 5). In this era, IBM actively facilitated sharing of software
among its users as a way of increasing usage of its computers.
The structure of the computer industry and copyright’s role dramatically changed
during the 1970s. With technological advances creating a minicomputer market and
IBM’s 1969 decision to unbundle software from mainframe leasing in the face of antitrust
charges, computer hardware vendors and independent software developers came to use
copyright licenses to protect computer programs. The opening of a competitive propri-
etary software marketplace ended an era in which software was freely shared (Phillips,
2009, pp. 113–15).

This shift produced a backlash within the programmer community that continues to
reverberate throughout the computer and software industries. The rapid rise of a robust
microcomputer industry followed by the creation of the Internet generated a robust,
independent software marketplace. These technologies had strong and complex network
effects, which have been substantially affected by software licensing practices. While
many hardware and software enterprises continue to rely heavily on proprietary software
licensing agreements, the programmers’ backlash against restrictive software licensing as
well as business strategies aimed at disrupting proprietary standards have dramatically
reshaped software licensing institutions, practices, and patterns.
This section explores this evolving landscape. Section (i) traces the emergence of the
free software movement, which resourcefully uses copyright licensing to promote open
platforms. The movement’s innovative licensing framework produced a form of network
effects. Section (ii) examines the open source movement, based on a more permissive
licensing model, which broadened the shift away from proprietary software licensing.
Section (iii) discusses the use of dedication of software copyrights to the public domain
as a third alternative for promoting network effects. Section (iv) surveys federal copyright
preemption of licensing restrictions.
i. The free software movement (General Public License) Many independent and
academic programmers, who had long enjoyed free access to source code, viewed the
shift to proprietary software licensing as a debilitating restriction on collaborative
research, programming freedom, and software innovation. Beginning in the early
1980s, Richard Stallman, then a researcher in MIT’s Artificial Intelligence Laboratory,
began a grass-roots ‘free software’ movement. Although Stallman was vehemently
opposed to intellectual property protection for computer software, he came to see that
the same copyright protections that exclude competitors could be deployed to prohibit
restrictions on adaptation and reuse of code and to foster open platforms (Weber, 2004,
pp. 47–9).
Stallman established the Free Software Foundation (FSF) in 1985 to promote users’
rights to use, study, copy, modify, and redistribute computer programs. The FSF devised
the GPL to prevent programmers from building proprietary limitations into software.
The GPL guarantees end users the freedoms to run, study, share (copy), and modify
the software so long as the users permit use of any derivative works on the same terms
(Carver, 2005). In this way, GPL software ‘infects’ derivative works with user rights and
virally spreads these rights through the collaborative software ecosystem.
Stallman targeted the development of a viable UNIX-compatible open source operat-
ing as FSF’s initial goal. The UNIX operating system, developed by researchers at MIT,
AT&T’s Bell Labs, and General Electric in the late 1960s and early 1970s, offered innova-
tive time-sharing capability. It became a foundation for modern computer operating
system design (McKusick, 1999). In 1972, two Bell Labs researchers—Dennis Ritchie,
inventor of the C programming language, and Ken Thompson—rewrote UNIX in C,
enabling UNIX to be installed on any advanced computer system. AT&T held the copy-
right to UNIX, which restricted its use and adaptation. Stallman sought to liberate UNIX
through the GNU (‘GNU’s Not Unix’) GPL independent re-implementation project.
Many programmers throughout the world contributed to this effort on a voluntary
basis, and by the late 1980s most of the components had been assembled. The project

reached fruition in 1991 when Linus Torvalds developed a UNIX-compatible kernel—the

central core of the operating system. Torvalds structured the evolution of his component
on the GPL model. The resulting UNIX-compatible free software program, dubbed
‘Linux,’ has become widely used throughout the computing world.
While attractive to many independent, non-commercial programmers, the so-called
‘copyleft’ GPL licensing model posed a serious problem for many commercial software
vendors. Although it afforded free access to GPL software, it prevented these cumulative
developers from charging a royalty for their modifications and subjected further modifica-
tions by licensees to GPL restrictions (Determann, 2006, p. 1484).
ii. The open software movement (permissive licenses) The ‘open source’ movement
emerged as a middle ground between proprietary software distribution and the ‘free’
software movement. Like Linux, the open source movement traces its roots to efforts
to liberate UNIX. In the mid-1970s, Ken Thompson at the University of California,
Berkeley, spearheaded an effort by Berkeley faculty and students to enhance UNIX
capabilities. In contrast to the GPL, the Berkeley Software Development (BSD) project
offered its software on a ‘permissive’ basis: licensees could distribute modifications of
the BSD software whether or not the modifications were freely licensed. Nonetheless,
the licensee was still obliged to obtain a license from AT&T for the underlying UNIX
code.
As the Internet took off in the late 1990s, a growing number of hardware and software
vendors embraced ‘free’ and ‘open source’ development and distribution strategies.
They saw these non- or less-proprietary licensing models as means to prevent Microsoft
from expanding its influence into the Internet and other platform technologies while
simultaneously promoting competition and innovation (Lerner and Tirole, 2004, 2005;
Merges, 2004; Benkler, 2002, 2004; McGowan, 2001). There is now a wide variety of
permissive open source licensing models (Meeker, 2015, Appendix B). Free (GPL) and
open source software play strong and increasing roles in network technologies, such as
operating systems (e.g., Linux), Internet infrastructure (e.g., Apache Web Server) and
mobile devices (e.g., Android), but have been less successful in penetrating consumer as
opposed to programmer-centric product areas (Phillips, 2009, pp. 156, 158–68; Lerner and
Tirole, 2005). Notwithstanding the proliferation of free and open source licenses, there
have been relatively few litigated disputes (Jacobsen v. Katzer, 535 F.3d 1373 (Fed. Cir.
2008); Phillips, 2009, pp. 120–21).
iii. Dedication to the public domain A further distribution alternative that has been
especially important in the proliferation of network benefits is outright dedication of
computer software copyrights (and other forms of intellectual property) to the public
domain. Tim Berners-Lee, the developer of the World Wide Web (WWW), was initially
attracted to releasing his hypertext software platform under the GPL (Berners-Lee, 2000,
pp. 72–3). Internet engineers, however, raised the concern that any restrictions attached
to its usage could limit its adoption and use. Some large companies were rumored to be
opposed to allowing usage of any software that could trigger license restrictions, includ-
ing GPL copyleft requirements. Berners-Lee ultimately chose to dedicate the WWW
to the public domain. Notwithstanding concerns that unprotected software could be
fragmented and captured through proprietary extensions, the WWW has thrived and

remained remarkably stable (Phillips, 2009, p. 174). This is attributable to the very strong
network effects of Internet protocols and the community and technically driven, open,
standard setting processes administered by the WWW Consortium (W3C) headed by
Berners-Lee and the IETF.
iv. Federal preemption of contractual restrictions In contrast to the free and open
software movements, some software developers use licensing provisions to restrict the
use of their copyrighted software. Some licenses, for example, bar reverse engineering of
software programs. Such a restriction affords the copyright owner greater control over
the development of interoperable products. The courts are divided, however, on whether
federal copyright law and intellectual property policies preempt such state law contractual
provisions.
In Vault Corp. v. Quaid Software, Ltd., 847 F.2d 255 (5th Cir. 1988), the Fifth Circuit
Court of Appeals held that the Louisiana Software License Enforcement Act clause
permitting a copyright owner to prohibit software decompilation or disassembly was
preempted by the Copyright Act, and was therefore unenforceable. A more recent case
interpreted the scope of federal copyright protection more narrowly, enforcing licensing
restrictions that bar activities that would otherwise fall within copyright’s fair use privilege
(Bowers v. Baystate Techs., 320 F.3d 1317 (Fed. Cir. 2003)). The dissenting opinion in
that case, however, indicates that the scope of federal preemption of licensing restrictions
that contract around the fair use privilege remains unsettled (Band and Katoh, 2011,
pp. 121–33). Section VI examines the related questions of whether antitrust law or misuse
doctrines further restrict licensing provisions that leverage intellectual property rights to
hinder downstream innovation or competition.
4. Interoperability exception to the DMCA’s anti-circumvention prohibition

The permissibility of reverse engineering software to achieve interoperability arose
during the legislative deliberations over the enactment of anti-circumvention prohibi-
tions. With the emergence of the Internet in the mid-1990s, motion picture studios,
record labels, publishers, and other content owners came to see encryption and other
digital rights management technologies as a promising self-help means to discourage
unauthorized distribution of their works. They recognized, however, that such technolo-
gies would be vulnerable to unauthorized circumvention of technological protection
measures. Thus, they sought to expand copyright protection beyond its traditional
prohibitions against infringement to include limits on the decrypting or circumventing
of technological protection systems and the trafficking in such decryption tools. They
contended that without such protection, they would be unwilling to release content onto
the Internet, which in turn would hamper the adoption of broadband services. Various
other interests—ranging from consumer electronics manufacturers, library associations,
computer scientists, and law professors—expressed concern about potential chilling
effects of such an expansion of copyright law upon those who wish to make fair use of
copyrighted works.
Congress crafted a compromise in the DMCA (17 U.S.C. § 1201). Section 1201(a) bans
circumvention of technological protection measures put in place by copyright owners to
protect copyrighted works. Section (b) prohibits trafficking in anti-circumvention tools.
Section 1201(f)(1) provides that:

a person who has lawfully obtained the right to use a copy of a computer program may cir-
cumvent a technological measure that effectively controls access to a particular portion of that
program for the sole purpose of identifying and analyzing those elements of the program that
are necessary to achieve interoperability of an independently created computer program with
other programs, and that have not previously been readily available to the person engaging in
the circumvention, to the extent any such acts of identification and analysis do not constitute
infringement under this title.
The legislative history notes that this provision is
intended to allow legitimate software developers to continue engaging in certain activities for
the purpose of achieving interoperability to the extent permitted by law prior to the enactment
of this chapter. The objective is to ensure that the effect of current case law interpreting the
Copyright Act is not changed by enactment of this legislation for certain acts of identification
and analysis done in respect of computer programs. See, Sega Enterprises Ltd. v Accolade, Inc.,
977 F.2d 1510, 24 U.S.P.Q. 2d 1561 (9th Cir. 1992.). The purpose of this section is to foster
competition and innovation in the computer and software industry. (S. Rep. No. 105-190, p. 13
(1998))
Because violations of the DMCA are not acts of copyright infringement, but rather
separate offenses, courts have held that the defenses available under the Copyright Act,
including fair use, do not apply to anti-circumvention violations (Universal City Studios,
Inc. v. Corley, 273 F.3d 429 (2d Cir. 2001); 321 Studios v. Metro-Goldwyn-Mayer Studios,
Inc., 307 F. Supp. 2d 1085 (N.D. Cal. 2004)). While § 1201(c)(1) provides that ‘nothing in
this law’ shall interfere with ‘fair use’ among other defenses, the courts have reasoned that
the DMCA does not interfere with fair use but merely renders it irrelevant by allowing
copyright owners to bring a non-copyright claim. Furthermore, the larger structure of
the DMCA provides additional safeguards to address free expression and other concerns.
Beyond the statutory exemptions to the anti-circumvention ban, the DMCA estab-
lished a triennial rulemaking process for exempting particular categories of works from
the anti-circumvention ban for which ‘noninfringing uses by persons who are users of a
copyrighted work are, or are likely to be, adversely affected’ (17 U.S.C. § 1201(a)(1)(B)–
(D)). Several of the granted exemptions authorize decryption for purposes of developing
interoperable products.
Smartphones, tablets, other mobile computing devices, and smart TVs, all of which
have networking aspects, have attracted particular attention. Several major manufacturers
of these products have sought to use encryption technologies to bundle the devices in
telecommunications service plans. In a series of rulemaking proceedings, the Copyright
Office has exempted unlocking or ‘jail-breaking’ of these products from the anti-
circumvention ban. (Library of Congress, 2018). Congress and the FCC have reinforced,
extended, and expanded these exemptions (Band, 2017, pp. 79–109; Unlocking Consumer
Choice and Wireless Competition Act, Pub. L. 113-144, 128 Stat. 1751 (2014)).
The DMCA’s anti-circumvention provisions have generated several cases involving the
use of technological protection measures to exclude competitors from aftermarkets—
goods or services supplied for a durable product after its initial sale (e.g., replacement
ink for printers). Several companies embedded digital code into their products and
aftermarket components that must interoperate to function as a means of exerting control
over such aftermarkets. When competitors in these aftermarkets decrypted such digital

codes to manufacture their own components, these durable product manufacturers sued,
alleging violation of the anti-circumvention provisions of the DMCA. Some courts have
declined to find liability, emphasizing that the careful balance that Congress sought to
achieve between the interests of content creators and information users would be upset
if the anti-circumvention prohibitions could be applied to activities that did not facilitate
copyright infringement (Chamberlain Group, Inc. v. Skylink Techs., Inc., 381 F.3d 1178,
1202 (Fed. Cir. 2004) (holding that ‘section 1201 prohibits only forms of access that bear
a reasonable relationship to the protections that the Copyright Act otherwise affords
copyright owners’); Lexmark Int’l, Inc. v. Static Control Components, Inc., 387 F.3d 522
(6th Cir. 2005) (holding that the lock-out technology at issue did not effectively control
access to a copyrighted work); Storage Technology Corp. v. Custom Hardware Eng’g &
Consulting, Inc., 421 F.3d 1307 (Fed. Cir. 2005) (holding that decryption (by third-party
software repair entity) to perform software maintenance activities is not actionable);
but see MDY Indus., LLC v. Blizzard Entm’t, Inc., 629 F.3d 928, 948‒52 (9th Cir. 2010)
(finding that only an incidental relation must exist between circumvention and copyright
infringement)).
i. GPL 3.0—DRM provision To bar intellectual property restrictions on software use
and promote sharing of code, the FSF added a provision to the GPL 3.0 (released in 2007)
barring licensors and those who use the licensed code from enforcing anti-circumvention
prohibitions (GNU General Public License 3.0, 2007, § 3). GPL 3.0 has not been as widely
adopted as prior GPL versions, particularly among commercial enterprises (Section D(3)
(iii); Meeker, 2015, ch. 10).
5. Software copyright jurisprudence: the Oracle v. Google litigation

After the Lotus v. Borland case resolved, litigation subsided over copyright protection for
the functional specifications of APIs and other network features of computer software
(Menell, 1998). The Sega, Altai, and Borland decisions and software industry norms
accorded competitors the ability to develop interoperable code and devices so long as
they independently implemented the functional specifications of the target platform
(Menell, 2018). If the programs were encrypted or only released in object code form,
the competitor would need to reverse engineer the code, which could be costly and time-
consuming. Beyond the drudgery of reverse engineering, copyright did not stand in the
way of developing and distributing interoperable code and devices.
A shift in business strategy in the Internet Age reinforced these legal principles and
industry norms. Whereas most software vendors in the pre-Internet era sought to
appropriate a return on their investments directly through software and device sales
and licenses, the Internet expanded the potential for multi-sided markets and indirect
appropriability—principally through advertising, service plans, and use of customer data
(Campbell-Kelly et al., 2015; Evans, 2003; Shapiro and Varian, 1999). These strategies
harnessed the positive feedback effects of network technologies.
Beginning with Netscape, a growing number of Internet Age entrepreneurs valued
adoptions over revenues in the start-up phase of their enterprises. The Internet provided
a low-cost means of distributing information and software, goods that had zero marginal
reproduction cost. For example, Sun Microsystems released the Java programming
language to the public as a means of promoting its hardware sales and forestalling

Microsoft’s dominance of website development tools (Menell, 2018). Google developed a

robust revenue stream for its search technologies without ever charging users. It profited
handsomely from bundling search results with keyword-generated advertisements.
Thus, many software and Internet companies welcomed adoption of their platforms,
including interoperability with their APIs. Sun Microsystems dedicated the Java pro-
graming language to the public domain early on, and in 2006 licensed the Java Standard
Edition, Enterprise Edition, and ME platforms—comprising packages of pre-written
APIs— under the GPL. Unlike Sega, it published its API specifications for the world to
see, adopt, and emulate. Its primary concern was maintaining the WORA interoperability
of these platforms. Hence, it required licensees to verify that implementations satisfied the
particular Java Technology Compatibility Kit (TCK) test.
When Google ventured into mobile platform development, it sought to take advantage
of the millions of programmers intimately familiar with Java, the most widely used
programming language and platform for web development. But unlike Borland, which
sought to achieve perfect interoperability with the Lotus 1-2-3 menu command hierarchy
so that Lotus macros could run on Borland’s Quattro system, Google sought to customize
Java for the smaller chip size of mobile handsets and add additional features, such as
location tools and a camera. Consequently, Google did not plan to include all the Java
APIs, which meant that the resulting system would not pass the Java TCK test. Moreover,
Google and its open handset alliance partners did not believe that the GPL would provide
sufficient flexibility for the range of players it believed would be needed to establish a
robust new mobile platform. They worried that the viral share and share alike provision
would discourage Google’s handset manufacturer and telecommunications partners from
investing in innovative features. The members of the Android Open Handset Alliance
believed that a more permissive licensing model, in which downstream suppliers could
make proprietary extensions on top of the base platform, would better promote robust
competition and innovation.
When licensing negotiations between Google and Sun reached an impasse, Google
chose to re-implement a subset of Java API packages independently to take advantage of
the vast Java programming community and the decade of testing that the Java APIs had
undergone. Google did not need to reverse engineer the Java API functional specifications
because Sun disclosed them. Nonetheless, Google had to devote substantial resources to
re-implementing the code using a clean room process.
When Google introduced Android in late 2007, Sun’s CEO publicly praised the adop-
tion of Java. Privately, however, he and other Sun leaders seethed at Google’s cavalier
approach and forking of the Java platform. Nonetheless, Sun refrained from blocking
Android through legal action (Menell, 2018).
With its hardware business in decline and unable to monetize Java, Sun’s viability as
an independent company came into question. Oracle Corporation, which had built many
of its software products on the Java platform, acquired Sun in 2010. Oracle immediately
pressured Google to license Java and when Google declined, sued alleging that Android
infringed Java-related patents and copyrights. Oracle focused its copyright claim on
Google’s copying of function labels, functional specifications (declarations), and the
structure, sequence, and organization of 37 Java API packages.
After the jury rejected Oracle’s patent causes of action, the district court ruled that the
Java APIs were not copyrightable (Oracle America, Inc. v. Google, Inc., 872 F. Supp. 2d

974 (N.D. Cal. 2012)). Judge Alsup cautioned that the ruling did not hold ‘Java API pack-
ages are free for all to use without license’ or that ‘the structure, sequence and organization
of all computer programs may be stolen.’ He grounded his decision in the particular and
distinctive functional attributes of the 37 Java APIs and that Google independently wrote
its own implementing code using a clean room process. The principal copying concerned
the lines of declarations, which are necessary to operate the particular methods of the
APIs. As Judge Alsup explained,
Significantly, the rules of Java dictate the precise form of certain necessary lines of code called
declarations, whose precise and necessary form explains why Android and Java must be identical
when it comes to those particular lines of code. That is, since there is only one way to declare a
given method functionality, everyone using that function must write that specific line of code in
the same way. (872 F. Supp. at 979 (emphasis in original))
While acknowledging that the overall structure of the Java API packages is creative,
original, and ‘resembles a taxonomy,’ Judge Alsup nonetheless concluded that it functions
as ‘a command structure, a system or method of operation—a long hierarchy of over six
thousand commands to carry out pre-assigned functions.’ Applying copyright’s limiting
doctrines as the Ninth Circuit has interpreted them, emphasizing the Sega decision, and
following CONTU’s guidance that when specific computer instructions, ‘even though
previously copyrighted, are the only and essential means of accomplishing a given task, their
later use by another will not amount to an infringement,’ (872 F. Supp. 2d at 986 (quoting
CONTU, 1979, p. 20) (emphasis added by Judge Alsup)), Judge Alsup determined that
Google was free to write code that accomplished the same functionality as the Java APIs
at issue even if it did not achieve complete compatibility with the full Java platform. Later
developers can achieve the particular functionality or method of operation of an API
subsystem (and even groups of subsystems) so long as they write their own code and no
patent protects that method.
Oracle appealed the copyright issues to the US Court of Appeals for the Federal
Circuit. (Menell, 2016 (explaining and questioning the Federal Circuit’s jurisdiction
over appeals from district court cases involving patent infringement allegations even if
neither party challenges the district court’s patent rulings)). The Federal Circuit is bound
by regional circuit law when reviewing questions that involve law and precedent not
exclusively assigned to the Federal Circuit.
Notwithstanding the Ninth Circuit’s holding in Sega and Sony v. Connectix that copy-
right law does not prohibit the precise coding necessary to achieve interoperability (Sega
Enterprises Ltd. v. Accolade, Inc., 977 F.2d 1510, 1525 (9th Cir. 1993); Sony Computer
Entertainment, Inc. v. Connectix Corp., 203 F.3d at 603 (‘There is no question that the
Sony BIOS contains unprotected functional elements’), the Federal Circuit reversed the
district court’s determination that the structure, sequence, and organization of the 37 Java
APIs were not copyrightable (Oracle America, Inc. v. Google, Inc., 750 F.3d 1339 (Fed.
Cir. 2014); Menell, 2018). The appellate court determined that even high-level API design
choices—including function labeling choices and compilation of functions—satisfy
copyright law’s low originality threshold. The court side-stepped the Sega and Sony cases
by construing Ninth Circuit law to hold that ‘copyrightability is focused on the choices
available to the plaintiff at the time the computer program was created,’ not the defend-
ant’s desire to achieve interoperability. The court concluded that Google’s interoperability

argument comes into play only as part of a fair use defense, an issue on which the jury had
hung. Consequently, the court remanded the case for a fair use trial.
On remand, the jury concluded that Android’s use of Java API declarations and
structure, sequence, and organization constituted fair use. The Federal Circuit once again
reversed, holding that the fair use balance tilted in Oracle’s favor (Oracle America, Inc.
v. Google LLC, 886 F.3d 1179 (Fed. Cir. 2018)). The Federal Circuit’s decision gives no
weight to the second fair use factor based on a questionable reading of Ninth Circuit
jurisprudence.3
The Federal Circuit’s decision rejecting Judge Alsup’s API copyrightability ruling is the
most significant recent federal appellate decision to confront the copyrightability of APIs.
Given the proliferation of software patents, there is a high likelihood that a company with
a widely-used set of APIs would be able to pursue both patent and copyright causes of
action in the same litigation, thereby bringing the Federal Circuit’s exclusive jurisdiction
over patent cases into play. Google is seeking Supreme Court review of both the Federal
Circuit’s 2014 API copyrightability decision and its 2018 fair use decision.
6. Standards and codes

Copyright protection extends to any work of authorship fixed in a tangible medium of
expression, subject to various limiting doctrines, such as the idea-expression dichotomy
and fair use. Standard setting bodies generally promote access to their standards and
codes. Sun (and later Oracle) published the Java API declarations. Their members typi-
cally wish to encourage widespread adoption of sponsored standards.
Some developers of standards seek to control access to their specifications. As reflected
in the Sega case, Sega controlled the access codes for the Genesis game platform through
trade secret law. After Accolade successfully reverse engineered the interoperability code,
Sega sought to bar its use by Accolade (and recover for copyright infringement). The
Ninth Circuit held, however, that software code elements necessary for interoperability
are unprotectable by copyright law.
Various technical, building, and other standards development seek to control access to
their work product principally to earn publication royalties. They contend that the royalty
income provides vital funding for coordinating standard development, resulting in better
formulated and maintained codes (Shannon, 2012).
Scholars have questioned the need for copyright protection to promote standards
developments. Professor Paul Goldstein contends that ‘[i]t is difficult to imagine an area
3
Oracle America, Inc. v. Google LLC, 886 F.3d 1179, 1205 (Fed. Cir. 2018) (explaining that
“[t]he Ninth Circuit has recognized . . . that th[e] second factor ‘typically has not been terribly sig-
nificant in the overall fair use balancing.’ Dr. Seuss Enters., L.P. v. Penguin Books USA, Inc., 109
F.3d 1394, 1402 (9th Cir. 1997) (finding that the ‘creativity, imagination and originality embodied
in The Cat in the Hat and its central character tilts the scale against fair use’); Mattel[, Inc. v.
Walking Mountain Prods., 353 F.3d 792, 803 (9th Cir. 2003)] (similar)”). The Federal Circuit’s
reliance on Dr. Seuss Enters. and Mattel is misplaced. Those cases addressed familiar children’s
stories and dolls; neither involved functional works, let alone computer software. By contrast, the
Ninth Circuit’s decisions in Sega (977 F.2d at 1524-27 (9th (extensive discussion of the second
factor connecting fair use to Baker v. Selden and § 102(b)) and Sony Comput. Entm’t (203 F.3d
at 602-05 (leading its discussion of fair use with the second fair use factor and affording it great
significance)), provide a far sounder footing for analyzing fair use in Oracle v. Google.

of creative endeavor in which the copyright incentive is needed less. Trade organizations
have powerful reasons stemming from industry standardization, quality control, and
self-regulation to produce these model codes; it is unlikely that, without copyright, they
will cease producing them’ (Goldstein, 2000, § 2.5.2).
The accessibility of edicts of law raises fundamental constitutional and policy ques-
tions (Malamud, 2014). Federal, state, and local laws, judicial opinions, and regulations
incorporate these codes. The Copyright Act expressly exempts works of the federal gov-
ernment from copyright protection (17 U.S.C. § 105). Court decisions on copyrightability
of non-federal edicts of law have been mixed.
The Fifth Circuit held that model codes enter the public domain when they enter into
law (Veeck v. S. Bldg. Code Congress Int’l, Inc., 293 F.3d 791 (5th Cir. 2002) (en banc)).
Building on that precedent, the Eleventh Circuit held that state law and the annotated
compilation of such law are sufficiently law-like to be regarded as sovereign work con-
structively authored by the citizens and thus not copyrightable. (Code Revision Comm’n
for General Assembly of Georgia v. Public.Resource.Org, Inc., 906 F.3d 1229, 1233,
1243-54 (11th Cir. 2018).
By contrast, the First Circuit recognized that copyright law could potentially protect
building codes (Building Officials & Code Admin. v. Code Technology, Inc., 628 F.2d
730, 736 (1st Cir. 1980)). The Ninth Circuit held that incorporation of a classification
system (taxonomy) for medical procedures in Medicare and Medicaid regulations does
not make them uncopyrightable (Practice Mgmt. Info. Corp. v. American Med. Assoc.,
121 F.3d 516, 518–20 (9th Cir. 1997)). Nonetheless, the court held that the copyright
misuse doctrine limited the ability of the AMA to enforce its copyright against a health
maintenance organization that used the taxonomy to comply with federal law (see
Section VI(A)(2)). Most recently, the D.C. Circuit overturned and remanded issuance
of a permanent injunction barring a non-profit organization from distributing copies of
technical standards produced by a private organization based on copyright and trade-
mark grounds. (American Society for Testing and Materials, et al. v. Public.Resource.
Org, Inc., 896 F.3d 437 (D.C. Cir. 2018)). As the court noted, ‘[f]ederal, state, and local
governments . . . have incorporated by reference thousands of these standards into law.’
(Id. at 440.). The court avoided a constitutional ruling by finding that the district court
‘failed to adequately consider whether, in certain circumstances, distributing copies of
the law for purposes of facilitating public access could constitute transformative use.’
(Id. at 450.).
C. Trademark Protection, Unfair Competition Law, and False Advertising Protection
In contrast to patent, copyright, and trade secret protection—which seek to promote

innovation—trademark, unfair competition law, and false advertising protection focus
primarily on ensuring the integrity of the commercial marketplace (Menell et al., 2017,
ch. V; Menell and Scotchmer, 2007).
The federal Lanham (Trademark) Act as well as analogous state statutes and common
law protects words, symbols, and other attributes, such as designs, slogans, and colors,
that serve to identify the source of goods or services. Certification marks certify con-
formity with centralized standards. Collective marks connote that a product or service
is manufactured or distributed by a member of a collective organization (e.g., Florists’

Transworld Delivery Association (FTD)) or that a product or service provider is a

member of a collective organization (e.g., American Automobile Association (AAA)).
To receive trademark protection, a mark need not be new or previously unused, but it
must represent a particular source of the good or service to consumers. It cannot merely
describe the good (e.g., hotel) or represent a generic term (e.g., thermos) for the class of
goods or services offered. Further, the identifying mark may not be a functional element
of the product itself but must serve a purely identifying purpose. Trademarks do not
expire, but continue in force unless their owner abandons them or they become generic.
Unlike patents or copyrights, trademarks do not directly protect the technology, good,
or work, but rather prevent others from creating a likelihood of consumer confusion as to
the source of goods. Thus, competitors may use the trademark of other companies in non-
confusing ways, such as comparative advertising and descriptive usages. Furthermore,
like copyright law, trademark law does not protect functional features of products.
Thus, trademark law does not protect the shape of shredded wheat biscuits (Kellogg Co.
v. National Biscuit Co., 305 U.S. 111, 122 (1938) (noting that the pillow-shaped form
reduces the cost of manufacturing the biscuit and affects its quality)). Patent law provides
the sole means of excluding competitors from utilitarian features of products. Similarly,
trademark law cannot protect aesthetically functional features of goods or packaging.
Thus, trademark law does not protect a red, heart-shaped box for packaging chocolates
(Restatement of Torts, § 742, comment a). The Lanham Act and state laws prohibit false
or misleading advertising.
The Supreme Court applied the functionality doctrine in a case involving network
effects (Inwood Labs., Inc. v. Ives Labs., Inc., 456 U.S. 844 (1982)). Ives Laboratories
manufactured and marketed a patented prescription drug using distinctively colored
capsules: a blue capsule for its 200-mg dosage and a combination blue-red capsule for its
400-mg dosage. Consumers and pharmacists came to associate the distinctive appearance
of the capsules with the particular patented compound and dosages. Thus, a consumer
could identify whether they were taking the proper drug and dosage from its appearance.
In that way, the packaging served as a very simple language.
Following expiration of the utility patent, generic drug manufacturers marketed the
chemical compound using the same color capsules. Ives sued generic drug makers for
indirect trademark infringement, alleging that they bore responsibility for pharmacists
that mislabeled the source of the drugs. Many pharmacies distribute capsules in
pharmacist-branded bottles. The pharmacists violated trademark law by filling requests
for Ives capsules with generic versions. The generic companies, however, only bore liability
if they intentionally induced pharmacists to infringe the Ives trademark or if it continued
to supply its product to pharmacists that it knew were engaging in infringement.
In finding that Ives had not proven that the generic manufacturers were indirectly liable
for trademark infringement, the Supreme Court observed that ‘a product feature is func-
tional if it is essential to the use or purpose of the article or if it affects the cost or quality
of the article’ (456 U.S. at 850, n. 10). A concurring opinion goes further, noting that
a finding of functionality offers a complete affirmative defense to a contributory infringe-
ment claim predicated solely on the reproduction of a functional attribute of the product.
A functional characteristic is ‘an important ingredient in the commercial success of the
product,’ and, after expiration of a patent, it is no more the property of the originator than the
product itself. It makes no more sense to base contributory infringement upon the copying of

functional colors than on the petitioners’ decision to use the same formulation of the drug, or
even to market the generic substitute in the first place. To be sure, the very existence of generic
drugs ‘facilitates’ illegal substitution. But Ives no longer has a patent for cyclandelate, ‘and the
defendants have a right to reproduce it as nearly as they can.’ Saxlehner v. Wagner, 216 U.S.
375, 380 (1910) (Holmes, J.). Reproduction of a functional attribute is legitimate competitive
activity. (Inwood Labs., Inc. v. Ives Labs., Inc., 456 U.S. at 862–3 (White, J., concurring) (some
citations omitted))
Trademark and unfair competition regimes play a variety of roles in controlling and
regulating information technology network markets by enabling platform sponsors to
regulate the usage of terms and symbols that signal interoperability and compatibility
with particular standards and interfaces (Lemley et al., 2011, ch. 4).
Platform sponsors and standard setting organizations routinely establish certification
and collective markets and use trademark law to police use of these designations. As
noted above, Sun Microsystems (and now Oracle Corporation) uses the Java TCK test as
well as certification marks to ensure that products using the Java trademark meet WORA
interoperability standards. In the mid to late 1990s, Sun used the ‘100% Pure Java’ initia-
tive to establish Java as a de facto industry standard (Floren, 1997). Sun successfully sued
Microsoft for violating its agreement not to adhere to Java’s standardized application
environment and compliance tests so as to ensure interoperability (Sun Microsystems,
Inc. v. Microsoft Corp., 87 F. Supp. 2d 992 (N.D. Cal. 2000)).
Platform sponsors have used trademark and false advertising law to combat confusing
product names or packaging and police compatibility and interoperability claims. Apple
Computer, for example, successfully prevented a competitor for using the term ‘Pineapple’
for its clone device (Apple Computer v. Formula Int’l, 562 F. Supp. 775 (C.D. Cal. 1983)).
As another example, Hewlett-Packard blocked an ink refiller from using confusingly
similar packaging for replacement cartridges (Hewlett-Packard Co. v. Nu-Kote Intern.,
Inc., 2000 WL 33992123 (N.D. Cal. 2000)).
In an interesting application of trademark’s genericide doctrine, Intel Corporation
sought to protect the ‘x86’ suffix from confusing use by a competitor. The court deter-
mined, however, that the ‘x86’ designation had become generic among buyers and sellers
of microprocessor chips (Intel v. Advanced Micro Devices, 765 F. Supp. 1292 (N.D. Cal.
1991)). Consequently, Intel designated its fifth generation design the Pentium. By con-
trast, notwithstanding the serious questions a court raised about whether ‘Windows’ was
generic for a graphical user interface, (Microsoft Corp. v. Lindows.com, Inc., 2002 WL
32153471 (W.D. Wash. 2002)), Microsoft obtained federal registration for the Windows
term. Google has successfully fended off claims that ‘google’ has become a generic term
for Internet search (Elliott v. Google, Inc., 860 F.3d 1151 (9th Cir. 2017)).
Platform sponsors and complementary product manufacturers have used trademark
and false advertising law to police use of compatibility and interoperability claims. In
Princeton Graphics Operating, L.P. v. NEC Home Electronics (U.S.A.), Inc., 732 F. Supp.
1258 (S.D.N.Y. 1990), the court applied a restrictive definition of compatibility because
of the importance of precise definitions in the computer industry (see also Creative Labs,
Inc. v. Cyrix Corp., 42 U.S.P.Q. 2d 1872 (N.D. Cal. 1997)).
In another interesting application of trademark law’s functionality doctrine, the Ninth
Circuit declined to allow Sega to use trademark law to prevent Accolade from selling
interoperable products that displayed Sega’s trademark as part of its lock-out code (Sega

Enterprises, Ltd. v. Accolade, Inc., 977 F.2d 1510 (9th Cir. 1992)). The basis for the
trademark claim was that the initialization code prompted a visual display for approxi-
mately three seconds that read ‘PRODUCED BY OR UNDER LICENSE FROM SEGA
ENTERPRISES LTD.’ The court rejected the false labeling claim as inconsistent with
the purposes of the Lanham Act. It also held that Sega could not use trademark law to
prevent competitors from marketing interoperable devices if the software design required
display of what might otherwise be confusing trademark information. The court ruled
that Sega failed to prove the existence of a feasible alternative to using the particular
lock-out code that produced the misleading label. Furthermore, Accolade had placed text
on its packaging materials disclaiming any association with Sega.
D. Patent Protection
Patents have long provided the potential for exclusive rights for network technologies. For
example, Alexander Graham Bell, who edged out Elisha Gray in a patent race over the
telephone, gained monopoly control over the quintessential network technology (Bruce,
1990). As the Supreme Court noted in Dolbear v. Am. Bell Tel. Co., 126 U.S. 1, 535
(1888), although an inventor’s claim might practically preempt all use of a discovery for
the duration of the patent, this fact will ‘show more clearly the great importance of his dis-
covery, [] it will not invalidate [the preempting] patent.’ Patents tracing back to Guglielmo
Marconi wireless communications technology played a central role in the developments
of the radio and television industries (Aitken, 1985). Xerox controlled the photocopy-
ing industry for several decades in the mid-20th century. Intel built its microprocessor
juggernaut on patents. Other network technology industries—from modems (Gandal et
al., 2007) to cell phones (Code Division Multiple Access (CDMA)) (Mock, 2005)—were
built on patent portfolios. Concern over patents affects many standard setting processes
(Contreras, 2019).
The extent to which patents enable control of network technologies depends on a range
of factors, including the extent to which the patent controls network features (patent
scope), the effective duration of patent protection, licensing structures (including patent
pools) (Mattioli, 2019), and antitrust constraints.
The advent of computer software introduced several additional complicating factors.
As courts limited copyright protection for network features of computer software
and the Federal Circuit expanded patent eligibility for software-related inventions in
the 1990s, the patent system emerged as a battleground for software-related network
technologies. Patent law’s higher protection threshold compared to other intellectual
property modes seeks to ensure that trivial advances remain available to the public while
potentially providing substantial advances robust protection, thereby motivating platform
developers to take on the challenge of overcoming the excess inertia of entrenched, but
obsolete, platforms. Patent law’s disclosure requirements enable the public to learn from
technological advances. Nonetheless, patent protection’s 20-year duration, although far
less than copyright protection, might still be excessive for software technologies (Menell,
1987). The uncertain scope of patent protection also poses some concern. Patent remedies
can be especially strong, although standard setting processes have tempered their effects
and promoted collaboration. Finally, design patent protection has recently added a new
weapon to the network technology arsenal.

This section examines patent protection for network technologies. It emphasizes the
most salient and contested area: computer software. Section 1 traces the evolution of
patent protection for software-related inventions. Section 2 examines the complicated
scope of patent protection. Section 3 discusses patent licensing. Section 4 explores patent
remedies. Section 5 examines design patents and their emergence in network markets.
1. Patentability requirements

The Patent Act sets forth five patentability requirements: (1) patentable subject matter; (2)
utility; (3) novelty; (4) nonobviousness; and (5) disclosure (35 U.S.C. §§ 101, 102, 103, 112;
Menell et al., 2017, ch. III). Two of these requirements have been particularly pertinent to
network industries: subject matter eligibility and nonobviousness.
i. Subject matter eligibility As noted above, the patent system has long afforded
protection for network and systems technologies, ranging from the telephone to wireless
communication and xerography. These technologies fit comfortably within the traditional
scope of patent protection. The patent system has, however, struggled to accommodate
software-related inventions. As illustrated above, APIs and other software technologies
are increasingly important in network industries.
Notwithstanding that the patent statute expressly authorizes patenting of processes
and machines (35 U.S.C. § 101), the availability of patent protection for software-related
inventions has been in flux since the beginning of the computer age. The issue emerged in
the 1960s as computer systems became more versatile, software languages developed, and
computer programming emerged from the shadow of electrical engineering. The Patent
Office struggled to fit software inventions within the traditional classification system and
struggled to keep up with the tremendous volume of prior art being generated. In 1965,
President Johnson appointed a commission to assess the overall efficacy of the patent
system (Executive Order No. 11,215, 30 Fed. Reg. 4661 (1965)). In recommending that
Congress exclude computer programs from patent eligibility, the Commission of govern-
ment officials, leading scientists, and representatives of industry (including IBM) noted
that ‘the creation of programs has undergone substantial and satisfactory growth in the
absence of patent protection’ and that ‘copyright protection for programs is presently
available’ (President’s Commission on the Patent System, 1967). But as discussed above,
copyright excluded protection for functional features of expressive works.
Congress did not act on this recommendation, and the eligibility of software-related
inventions fell to the Patent Office and the courts. Although granting a smattering of
software-related inventions in the mid to late 1960s, the Patent Office took a skeptical view
of software eligibility. This in part reflected concerns about the Patent Office’s ability to
examine this new and rapidly developing technological field.
The Supreme Court was soon brought into the fray. An inventor challenged the PTO’s
rejection of his claim to an algorithm that converted binary-coded decimal numerals
into pure binary numerals on subject matter grounds (Gottschalk v. Benson, 409 U.S.
63 (1972)). The Court held that ‘[p]henomena of nature, though just discovered, mental
processes, and abstract intellectual concepts are not patentable, as they are the basic tools
of scientific and technological work’ (409 U.S. at 67). The Court noted, however, it was not
categorically excluding software-related inventions from patent eligibility. Yet six years
later, the Court ruled that even newly discovered algorithms should be treated as in the

prior art, rendering software claims ineligible unless they contained some other inventive
concept (Parker v. Flook, 437 U.S. 584, 594 (1978)). The Supreme Court reversed course
in 1980, holding that software claims should be viewed as a whole and that the touchstone
for patentability of a process embodying a mathematical formula was whether there was
significant post-solution activity, that is, ‘transforming or reducing an article to a different
state or thing’ (Diamond v. Diehr, 450 U.S. 175, 188–9, 191–2 (1981)).
Over the ensuing 25 years, the Court of Appeals for the Federal Circuit loosened patent
eligibility limitations. Building on Diehr, the Federal Circuit chipped away at the post-
solution activity necessary to bring software-related claims within § 101 (In re Alappat, 33
F.3d 1526 (Fed. Cir. 1994) (holding that the display of data on a computer screen could
suffice)). In 1998, the Federal Circuit held that business methods were eligible for patent
protection so long as they produced a ‘useful, concrete, and tangible result’ (State Street
Bank v. Signature Financial Group, 149 F.3d 1368, 1373 (Fed. Cir. 1998) (quoting In re
Alappat)).
In the aftermath of the Federal Circuit’s State Street Bank decision, the PTO shifted its
position from skepticism about expansive patent eligibility to openness and even enthu-
siasm. Patents for software and business methods flooded the PTO. Entrepreneurs and
venture capitalists saw patenting as a valuable tool for developing (or at least claiming)
Internet businesses. The late 1990s witnessed unprecedented growth of start-up businesses
based on speculative initial public offerings secured, in part, on patent portfolios.
The bursting of the Internet (dot-com) stock bubble in 2000 produced a dramatic
shakeout. Bankruptcies and, subsequently, the auctioning and trading of Internet-related
patents, became widespread. Entities whose sole purpose was to assert these patents
emerged. Patent holding companies and non-practicing entities sought to monetize their
Internet patents, often purchased at bankruptcy auctions. Lawsuits by patent assertion
entities produced a tidal wave of patent validity challenges as well as calls by Silicon Valley
companies, policymakers, and scholars for policy reform.
These concerns led the Federal Circuit to reinvigorate patent eligibility limitations
(see, e.g., In re Nuijten, 500 F.3d 1346 (Fed. Cir. 2007) (holding that a watermarked elec-
tromagnetic signal does not fall into any of the four categories of patent-eligible subject
matter); In re Comiskey, 554 F.3d 967 (Fed. Cir. 2009) (affirming rejection of a business
method patent under § 101 as merely relying on mental steps)). In an en banc ruling, the
Federal Circuit synthesized the Supreme Court’s Benson, Flook, and Diehr precedents
into the ‘machine-or-transformation test’: a claimed process is patent-eligible under § 101
if it is tied to a particular machine or if it transforms a particular article into a different
state or thing (In re Bilski, 545 F.3d 943, 961 (Fed. Cir. 2008) (en banc)). Applying this
test, the Federal Circuit affirmed the Patent Office’s rejection of a claim for a method
for managing the consumption risk costs of a commodity. The Supreme Court upheld
the Federal Circuit’s decision, although it characterized the machine-or-transformation
test as a ‘useful and important clue, an investigative tool, for determining whether some
claimed inventions are processes under § 101,’ but too rigid a test of the Patent Act’s broad
statutory definition of ‘process’ (Bilski v. Kappos, 561 U.S. 593, 604 (2010); 35 U.S.C. §
100(b)). The Court declined to rule that business methods are categorically ineligible for
patent protection (Menell, 2011).
Two years later, the Supreme Court revived the Flook decision’s rule that for a claim
embodying a natural discovery or algorithm to be eligible for patentability, it must

contain a sufficiently inventive concept beyond the natural law or algorithm, even where
the patentee discovered the natural law or algorithm (Mayo Collaborative Services v.
Prometheus Laboratories, Inc., 566 U.S. 66 (2012); Alice Corp. v. CLS Bank International,
573 U.S. 208 (2014)). These decisions have dramatically shifted the patent eligibility
landscape, resulting in the invalidation of a vast swath of software-related claims and
eliminating patent protection for pure business methods. The decisions have also reduced
the availability of patent protection for software-based network technologies.
ii. Nonobviousness To ensure that patents are not granted to routine or conventional

applications of known principles, the Patent Act requires that ‘the differences between
the claimed invention and the prior art are such that the claimed invention as a whole
would have been obvious before the [invention was made] to a person having ordinary
skill in the art to which the claimed invention pertains’ (35 U.S.C. § 103). This requirement
has long been difficult to apply due to the difficulty of ignoring the fact of the claimed
invention. To avoid such hindsight bias, the Federal Circuit interpreted § 103 to require
that the prior art teach, suggest, or motivate ordinary skilled artisans to combine prior art
references to achieve the claimed invention. Absent such evidence, the claimed invention
was nonobvious (Teleflex, Inc. v. KSR Intern. Co., 119 Fed. Appx. 282 (Fed. Cir. 2005);
In re Dembiczak, 175 F.3d 994, 998 (Fed.Cir.1999); Application of Bergel, 292 F.2d 955,
956–7 (C.C.P.A. 1961) (predecessor court to the Federal Circuit)). While such suggestions
can be relatively common in scientific publications—through cross-references of other
publications—they are not readily found in more commercial and applied fields, such as
software engineering. Software products do not typically cross-reference other products.
As a result, many seemingly obvious inventions from the standpoint of common knowl-
edge were able to clear the Federal Circuit’s nonobviousness test.
As software patent litigation exploded following the burst of the Internet bubble in
2000, the Federal Circuit’s standard for determining whether an invention was sufficiently
inventive came under scrutiny. In KSR Int’l Co. v. Teleflex Inc., 550 U.S. 398 (2007), the
Supreme Court tightened the nonobviousness standard by holding that the teaching-
suggestion-motivation test was too rigid.
When there is a design need or market pressure to solve a problem and there are a finite number
of identified, predictable solutions, a person of ordinary skill has good reason to pursue the
known options within his or her technical grasp. If this leads to the anticipated success, it is
likely the product not of innovation but of ordinary skill and common sense. In that instance
the fact that a combination was obvious to try might show that it was obvious. (550 U.S. at 421)
The KSR decision raised the patentability bar, especially for software-related technologies
for which market factors and advances in collateral technologies are likely to drive new
products and processes.
2. Scope
The extent to which patents control network technologies depends upon the scope of the
patent claims. Pioneering patents can stake broad claims without fear of being anticipated
by prior art, whereas incremental inventions in crowded technology fields only garner
narrow protection. Moreover, pioneering inventors can often develop improvement
patents that expand their control and duration of protection. Xerox successfully followed

this strategy to monopolize the photocopying industry for several decades (Scherer, 1987,
at 1016–17; Bresnahan, 1985). The resulting ‘patent thicket’ delayed entry into the plain
paper copy industry.
Software patentees have used broad, vague functional claim language to obtain broad
coverage for their inventions (Menell and Meurer, 2013; Federal Trade Commission,
2011). By avoiding the statutory phrases ‘means’ or ‘step’ in their claims—which limit the
scope of their claims to the particular embodiments in the specification and ‘equivalents
thereof’ (35 U.S.C. § 112(f))—and instead using broad terms that lack structural limits
such as ‘module,’ patent drafters have sought to control all software solutions to particular
technological problems (Lemley, 2013). Such claims have caused substantial problems in
the Internet Age, and have resulted in a proliferation of demand letters, costly litigation,
and nuisance value settlements.
The courts and the PTO have sought to rein in these problems. The Supreme Court
invigorated the claim indefiniteness doctrine, enforcing the patent statute’s requirement
to ‘particularly point[] out and distinctly claim[] the subject matter’ sought to be patented
(35 U.S.C. § 112(b); Nautilus, Inc. v. Biosig Instruments, Inc., 134 S. Ct. 2120 (2014);
Menell and Meurer, 2013). The Federal Circuit has interpreted claims terms like ‘module’
and other vague terms (which it refers to as ‘nonce’ words) to invoke the limitations of
§ 112(f) (Williamson v. Citrix Online LLC, 792 F.3d 1339 (Fed. Cir. 2015) (en banc)). This
interpretation limits claim scope to the embodiments in the specification and equivalents
thereof. Further upstream, the Patent Office is pursuing administrative efforts to improve
claim clarity (General Accountability Office, 2016; Menell, 2013).
3. Licensing
Patent licensing plays a critical role in many network industries. Patents afford patent
owners the power to prevent others from making, using, offering to sell, selling, or
importing the patented invention in the United States during the term of the patent (35
U.S.C. § 271). They do not, however, ensure that patentees can practice their own patented
invention. The owner of a patent that improves on patented technologies controlled by
others would need a license from the upstream patent owner to make, use, or sell the
improvement. Licensing provides the key.
Many network technologies employ patented technologies. Several distinctive licensing
issues have developed to address network effects: (i) standard setting and commitments
to license patents on FRAND terms; (ii) insurance pools and license on transfer commit-
ments; and (iii) GPL viral license commitments. Over-reaching licensing provisions can
raise misuse and antitrust issues addressed in Section VI.
i. Standard setting and FRAND commitments SSOs seek to lessen the tension
between employing the best technological solutions in industry standards and ensuring
widespread access to standards by requiring members to disclose SEPs and license them
on FRAND terms (Contreras, 2019; National Research Council, 2013). Most SSOs,
however, have not expressly barred injunctive relief or set FRAND licensing schedules.
In 2015, the IEEE barred its members from holding patents covering IEEE standards
from seeking or threatening to seek injunctions or exclusion orders against potential
licensees who are willing to negotiate for licenses (IEEE, 2015).

ii. Insurance pools and license on transfer (LOT) commitments In response to

widespread assertion of patents by non-practicing entities following the bursting of the
Internet bubble in early 2000, several enterprises emerged to reduce patent risk (Rice,
2015, pp. 752–3). Since 2008, RPX (Rational Patent Exchange) Corporation has func-
tioned as a consortium of technology companies that acquires patents that pose potential
risks. RPX has promised not to assert patents in its portfolio.
As a further pre-commitment strategy to prevent patent holdup, a growing number of
technology companies have promised not to assert their patents under specified condi-
tions (Rice, 2015, pp. 747–52). Google has led an initiative whereby companies agree to
prevent their patents from ever being used by a non-practicing entity (NPE) against other
member companies through a license on transfer (LOT) pledge (Rice, 2015; Shultz and
Urban, 2012). The LOT network produces a network benefit. As more companies join the
pact, the freedom to be insulated from NPE patent assertion entities expands.
iii. GPL 3.0 As noted earlier, patents did not play a substantial role in the software
industry until the mid-1990s, after the GPL (1989) and the GPL 2 (1991) were established.
Although neither of these versions of the GPL expressly licensed patents, the FSF took
the position that the GPL 2 created an implied license (Meeker, 2015, p. 127).
GPL 3.0 took aim at this issue. Section 10 provides that the licensee
may not impose any further restrictions on the exercise of the rights granted or affirmed under
this License. For example, you may not impose a license fee, royalty, or other charge for exercise
of rights granted under this License, and you may not initiate litigation (including a cross-claim
or counterclaim in a lawsuit) alleging that any patent claim is infringed by making, using, selling,
offering for sale, or importing the Program or any portion of it.
Section 11 goes further: each ‘contributor’ to code governed by the GPL 3 grants a ‘a non-
exclusive, worldwide, royalty-free patent license under the contributor’s essential patent
claims, to make, use, sell, offer for sale, import and otherwise run, modify, and propagate
the contents of its contributor version.’ That provision defines a contributor’s ‘essential
patent claims’ to include ‘all patent claims owned or controlled by the contributor,
whether already acquired or hereafter acquired, that would be infringed by some manner,
permitted by this License, of making, using, or selling its contributor version.’ Section
11 does not extend to ‘claims that would be infringed only as a consequence of further
modification of the contributor version.’
Section 11 further provides that a licensee who is aware of a patent license governing
GPL 3.0 code must make the corresponding source code to run the object code and
modify the work publicly available or extend the patent license to downstream recipients.
Alternatively, the licensee must deprive itself of the benefit of the license. Section 11
further includes a non-discrimination provision ensuring that any patent licenses are
extended to all recipients of the GPL 3 work and works based on it.
These provisions pose several serious concerns to many commercial software develop-
ers (Meeker, 2015, pp. 129–30). For example, many patent litigation settlements provide
only limited, non-sublicensable, and possibly royalty-bearing rights that would not
comply with GPL 3.0 requirements. Thus, commercial enterprises have been reluctant to
embrace GPL 3.0. As of February 2017, GPL 3.0 was the fourth most widely adopted
open source license (8 percent of open source projects), behind the MIT License (a simple

permissive) (31 percent), GPL 2.0 (18 percent), and Apache 2.0 (15 percent) (Black Duck
Opensource Knowledge Base, 2017; see also Goldstein, 2018 (noting that use of permis-
sive open source licenses are on the rise; reporting that in 2018, 64 percent of open source
components have permissive licenses, an 8 percent rise over 2017).
4. Remedies
Patent remedies play a critical role in the control of network technologies that are subject
to patent assertions. The proliferation of software patents and litigation in the Internet
Age generated tremendous exposure for network industry companies, leading to calls for
statutory reform of patent remedies (Federal Trade Commission, 2011, 2003).
i. Injunctive relief The patent right—the right to exclude others from practicing the
patented technology— has historically been protected by injunctive relief. Courts tradi-
tionally viewed patent rights like other property interests and routinely protected them
through a ‘property’ rule—barring transgressors from trespassing or using the ‘property.’
Thus, for most of the history of patent law, courts awarded a permanent injunction as the
prospective infringement remedy absent extraordinary circumstances.
The embrace of software and business method patents during the dot-com bubble of
the mid to late 1990s gave way to concerns about injunctions threatening major technol-
ogy companies in the aftermath of the NASDAQ market crash in the early 2000s. Patents
that had been acquired to attract venture capital were auctioned off in bankruptcy sales
to patent monetization entities. The proliferation of demand letters and patent lawsuits
led scholars, technology companies, policymakers, and jurists to reconsider the traditional
view of patents as property interests that deserve near-automatic injunctive relief (Lee
and Melamed, 2016; Menell, 2007, 2011; Lemley and Shapiro, 2007a). The costs of
identifying patent holders, negotiating among potentially hundreds of patent holders,
and the disruption and delay of litigation created leverage for patent owners. The threat
of injunctive relief and high monetary damages enabled holders of dubious patents to
extract unwarranted and disproportionate value.
In a watershed decision, the Supreme Court ruled in eBay, Inc. v. MercExchange, LLC,
547 U.S. 388 (2006), that the award of injunctive relief in patent cases turns on balancing
of the traditional equitable factors associated with preliminary relief: (1) whether the
harm is irreparable; (2) adequacy of monetary damages to compensate for the harm;
(3) balance of hardships between the parties; and (4) the public interest. (Gergen et al.,
2012, at 208–9). The eBay decision has changed patent remedies dramatically. Seaman
(2016) finds that the overall rate of permanent injunctions being ordered as a remedy for
patent infringement has dropped from near 100 percent to 72.5 percent. The drop is most
significant in software cases (53 percent). Patent assertion entities (i.e., non-practicing
patent owners) obtained permanent injunctions in just 16 percent of their victories.
Courts take SSO FRAND commitments into account in evaluating requests for
injunctive relief under the eBay standard. Although many SSO policies do not expressly
address whether SEP owners can seek injunctive relief or exclusion orders, courts consider
FRAND commitments in weighing the irreparable harm prong of the eBay equitable
relief test (Apple Inc. v. Motorola, Inc., 757 F.3d 1286, 1331–32 (Fed. Cir. 2014) (noting
that absent unusual circumstances, such as an infringer refusing a FRAND royalty or
unreasonably delaying negotiations, it will be difficult for a patent owner subject to a

FRAND commitment to establish irreparable harm or that damages are not an adequate
remedy, and that even when an infringer has refused to accept any license offer, that does
not necessarily justify injunctive relief); cf. Apple Inc. v. Samsung Elecs. Co., 809 F.3d
633 (Fed. Cir. 2015) (emphasizing right to exclude and the importance of injunctions)).
The eBay decision does not, however, leave the patent owner without a prospective
remedy. The court will fashion a prospective monetary damage measure, such as a running
royalty or a permanent damage amount—essentially a compulsory license. The eBay deci-
sion has led to a rise in patent enforcement filings at the International Trade Commission,
which enforces infringement findings with exclusion orders barring importation of
infringing articles (Chien and Lemley, 2012).
ii. Monetary relief The Patent Act authorizes the award of ‘damages adequate to
compensate for the infringement but in no event less than a reasonable royalty for the use
made of the invention by the infringer’ (35 U.S.C. § 284). Thus, patentees can recover lost
profits or a reasonable royalty resulting from infringing activity. The Patent Act further
authorizes judges to increase damages awards up to three times the compensatory level
where the infringer has acted willfully or recklessly (35 U.S.C. § 284; Halo Electronics, Inc.
v. Pulse Electronics, Inc., 136 S. Ct. 1923 (2016)). Policymakers and scholars see the goal
of patent damages to restore the parties to the position they would have achieved had they
negotiated a patent license before the infringement occurred (Lee and Melamed, 2016).
Patent law has long struggled to deal with apportioning patent value when a patent
covers only one component of a larger product or system (Cincinnati Car Co. v. New
York Rapid Transit Corp., 66 F.2d 592, 593 (2d Cir. 1933) (Learned Hand, J.) (observing
that the allocation of profits among multiple components ‘is in its nature unanswerable’)).
The problem has become particularly acute in platform technologies involving multiple
components and patented technologies. The serial nature of patent litigation, the economic
complexity of multi-component products, and court-imposed time limits on the presenta-
tion of evidence make it difficult for juries to apportion value among multiple components
and factors driving market demand for infringing products (Graham et al., 2017).
In theory, a wide range of royalty bases can be used with appropriately calibrated
royalty rates to account for the myriad factors affecting consumer demand. In practice,
however, the open-ended nature of the inquiry (Georgia-Pacific Corp. v. U.S. Plywood
Corp., 318 F. Supp. 1116 (S.D.N.Y. 1970) (identifying 15 factors)) can lead to a very large
royalty range across comparable cases. The Federal Circuit has sought to rationalize
awards by using the smallest saleable patent-practicing unit (SSPPU), as opposed to the
entire market value of the product or system, as the royalty base (LaserDynamics Inc. v.
Quanta Computer, Inc., 694 F.3d 51 (Fed. Cir. 2012); Cornell Univ. v. Hewlett-Packard
Co., 609 F. Supp. 2d 279 (N.D.N.Y. 2009) (Rader, J., sitting by designation); Uniloc
USA, Inc. v. Microsoft Corp., 632 F.3d 1292, 1320 (Fed. Cir. 2011); Lucent Techs., Inc.
v. Gateway, Inc., 580 F.3d 1301, 1336 (Fed. Cir. 2009); Ericsson, Inc. v. D-Link Sys., Inc.,
773 F.3d 1201, 1226 (Fed. Cir. 2014) (citing VirnetX, Inc. v. Cisco Sys., Inc., 767 F.3d
1308 (Fed. Cir. 2014) (‘[W]here multicomponent products are involved, the governing rule
is that the ultimate combination of royalty base and royalty rate must reflect the value
attributable to the infringing features of the product, and no more’))).
As noted above, SSOs have sought to alleviate the tension between technological pro-
gress and widespread access to standards by requiring members to disclose SEPs during

the standard setting process and license them to standards implementers on FRAND
terms (National Research Council, 2013). Nonetheless, the valuation of SEPs is difficult,
especially when industry standards encompass multiple technologies and hundreds of
patents. The challenge lies in separating the value of the particular technologies and
patents from the often tremendous value from standardization, which is attributable to
network effects. Once consumers adopt a product, they become locked in to the standard
to varying degrees. This can provide patentees with tremendous leverage. Courts have sur-
mounted this challenge by interpreting the principal goal of standard setting agreements
to be widespread adoption of the standard and barring FRAND licensors from capturing
the coordination and network value of the standard (CSIRO v. Cisco Sys., Inc., 809 F.3d
1295 (Fed. Cir. 2015); Ericsson, Inc. v. D-Link Sys., Inc., 773 F.3d 1201, 1229–35 (Fed.
Cir. 2014); Microsoft Corp. v. Motorola, Inc., No. C10-1823JLR, 2013 U.S. Dist. LEXIS
60233, 2013 WL 2111217 (W.D. Wash. April 25, 2013); In re Innovatio IP Ventures LLC
Patent Litig., No. 11 C 9308, 2013 U.S. Dist. LEXIS 144061, 2013 WL 5593609 (N.D. Ill.
October 3, 2013)).
5. Design patents

Design patents—which afford 15 years of protection for ‘new, original and ornamental
designs for an article of manufacture’ (35 U.S.C. § 171)—have come into play in some
network technology markets. As with copyright and trademark protection, design patents
do not extend to the functionality of useful articles (Buccafusco and Lemley, 2016; Du
Mont and Janis, 2012). Only utility patent protection can protect such elements.
Separating ornamental from functional features has proven difficult. The Federal
Circuit will invalidate a design patent only if the claimed design is dictated solely by
the function of the article of manufacture (Best Lock Corp. v. Ilco Unican Corp., 94
F.3d 1563, 1566 (Fed. Cir. 1996)). Some decisions have applied a looser balancing test:
asking whether a design is ‘primarily functional’ (L.A. Gear, Inc. v. Thom McAn Shoe
Co., 988 F.2d 1117, 1123 (Fed. Cir. 1993) (‘[T]he utility of each of the various elements
that comprise the design is not the relevant inquiry with respect to a design patent. In
determining whether a design is primarily functional or primarily ornamental the claimed
design is viewed in its entirety, for the ultimate question is not the functional or decorative
aspect of each separate feature, but the overall appearance of the article, in determining
whether the claimed design is dictated by the utilitarian purpose of the article’); Lee v.
Dayton-Hudson Corp., 838 F.2d 1186, 1188 (Fed. Cir. 1988)).
The difficulty lies in the fact that functionality is often intertwined with ornamentality,
especially in minimalist designs that merge form with function. Furthermore, compila-
tions of design features can themselves be functional. Some Federal Circuit decisions
address this challenge by dissecting the claimed design through a process that aligns
with copyright law’s treatment of the idea-expression dichotomy (see Section V(B)(2)
(i)). Thus, if a claimed design contains ‘both functional and ornamental features, the
patentee must show that the perceived similarity is based on the ornamental features of
the design’ (OddzOn Prods., Inc. v. Just Toys, Inc., 122 F.3d 1396, 1405 (Fed. Cir. 1997).
The courts ‘factor[] out the functional aspects of [the claimed design] as part of its claim
construction’ (122 F.3d at 1405; Richardson v. Stanley Works, Inc., 597 F.3d 1288, 1293
(Fed. Cir. 2010)). This approach, however, is in tension with the Federal Circuit’s holding
that claimed designs should be evaluated as a whole (Egyptian Goddess, Inc. v. Swisa,

Inc., 543 F.3d 665, 679 (Fed. Cir. 2008) (en banc) (rejecting focusing on a design’s ‘point-
of-novelty’; Crocs, Inc. v. Int’l Trade Comm’n, 598 F.3d 1294, 1302–03 (Fed. Cir. 2010)).
Some decisions have suggested that courts can surmount the separability challenge by
considering
whether the protected design represents the best design; whether alternative designs would
adversely affect the utility of the specified article; whether there are any concomitant utility
patents; whether the advertising touts particular features of the design as having specific utility;
and whether there are any elements in the design or an overall appearance clearly not dictated
by function. (Berry Sterling Corp. v. Pescor Plastics, 122 F.3d 1452, 1456 (Fed. Cir. 1997); see
also PHG Techs., LLC v. St. John Cos., 469 F.3d 1361 (Fed. Cir. 2006))
This standard parallels an earlier formulation of trademark law’s functionality doctrine

(In re Morton-Norwich Prods., Inc., 671 F.2d 1332, 1340-41 (C.C.P.A. 1982); see also
Amini Innovation Corp. v. Anthony Cal., Inc., 439 F.3d 1365, 1371 (Fed. Cir. 2006)
(stating that an ‘aspect’ of a patented design is functional ‘if it is essential to the use
or purpose of the article or if it affects the cost or quality of the article,’ a trademark
functionality standard articulated in Inwood Labs., Inc. v. Ives Labs., Inc., 456 U.S.
844, 850, n.10 (1982), discussed in Section V(C)). This approach reflects a concern with
design patents preempting competition. A design is functional if there are no alternative
designs that accomplish a function equally well (Rosco, Inc. v. Mirror Lite Co., 304 F.3d
1373, 1378 (Fed. Cir. 2002) (reasoning that ‘if other designs could produce the same or
similar functional capabilities, the design of the article in question is likely ornamental,
not functional’); Seiko Epson Corp. v. Nu-Kote Int’l, Inc., 190 F.3d 1360, 1368 (Fed. Cir.
1999) (explaining that ‘the design must not be governed solely by function, i.e., that this is
not the only possible form of the article that could perform its function’); L.A. Gear, Inc.
v. Thom McAn Shoe Co., 988 F.2d 1117, 1123 (Fed. Cir. 1993) (‘When there are several
ways to achieve the function of an article of manufacture, the design of the article is more
likely to serve a primarily ornamental purpose’)).
The Federal Circuit applied these principles in a case involving interoperability. In Best
Lock Corp. v. Ilco Unican Corp., 94 F.3d 1563 (Fed. Cir. 1996), Best Lock claimed an
unusual profile for a key blade blank, the form used for manufacturing (cutting) keys. Ilco
distributed key blanks with that key blade shape. The Federal Circuit found that function
alone dictated the key blade design because ‘no alternative blank key blade would fit the
corresponding lock’ (94 F.3d at 1566).
The integration of form and function in many product markets has brought design
patents into play in some network technology markets. Most notably, Apple successfully
asserted design patents covering the rounded rectangular shape of mobile communica-
tions devices against Samsung (Apple Inc. v. Samsung Elecs. Co., 786 F.3d 983 (Fed. Cir.
2015)) (holding that iPhone and iPad designs were functional (and hence unprotectable
under trademark law) but not functional under design patent law). Technology companies
and designers have also obtained design patents on virtual designs, patents that cover the
designs of graphical user interfaces for smartphones, tablets, and other products, as well
as the designs of icons or other artifacts of various virtual environments (Du Mont and
Janis, 2012).
The strong monetary remedies available for design patents further encourages seeking
design patents to protect features of network technologies. The Patent Act provides

for recovery of the ‘total profit’ on the sale of ‘any article of manufacture to which [a
protected design] has been applied’ (35 U.S.C. § 289). Although the Supreme Court held
that the term ‘article of manufacture’ encompasses both a product and a component of
that product (Samsung Elecs. v. Apple Inc., 137 S. Ct. 429 (2016)), the apportionment of
damages in design patent cases is uncertain.
VI. INTERPLAY OF INTELLECTUAL PROPERTY

PROTECTION AND COMPETITION POLICY IN NETWORK
INDUSTRIES
The Sherman Antitrust Act prohibits contracts in restraint of trade and monopolization
or attempts to monopolize markets. The courts have long recognized that patent and
copyright protections—government-authorized rights to exclude others from using
protected technologies and copying works of authorship—as exceptions to antitrust
law. Yet, patent and copyright protections can concentrate economic power in ways that
undermine competition. This is especially true in network technology markets, where
positive feedback effects often lead to strong and durable monopolies. At the same time,
high concentration can promote desirable network effects.
These considerations ameliorate and complicate the interplay of intellectual property
protection and competition policy. Section A explores limitations on improper leveraging
of intellectual property rights that arise in private enforcement of intellectual property
and contracts. Section B examines public enforcement of antitrust law and competition
policy in network markets.
A. Private Enforcement
Courts have recognized limits on the exercise of patent and copyright protection that
apply with special force in network industries. As we have already seen, various internal
intellectual property doctrines—such as copyright’s fair use doctrine and the use of
equitable balancing in dispensing remedies—bring competition policy concerns into
intellectual property law. In addition, courts have developed equitable and contract-based
defenses to prevent anti-competitive abuses of intellectual property rights.
1. Misuse doctrines

Drawing on tort law’s ‘unclean hands’ doctrine, courts developed the patent misuse
doctrines as a common law equitable defense to an infringement claim (Bohannan, 2011).
Unlike the purely equitable defense of ‘unclean hands,’ the misuse doctrines apply to
suits for damages as well as equitable relief. The misuse doctrine bars patent owners from
expanding the scope or term of the intellectual property right through licensing restric-
tions. The Supreme Court prevented Thomas Edison from leveraging a patent on motion
picture projectors to control what films could be exhibited using that projector (Motion
Picture Patents Co. v. Universal Film Mfg. Co., 243 U.S. 502 (1917) (refusing to enforce a
licensing provision restricting use of the machine to motion pictures licensed by Edison’s
film company); see also Carbice Corp. v. American Patents Development Corp., 283 U.S.
27 (1931); Morton Salt Co. v. G.S. Suppiger Co., 314 U.S. 488 (1942)). The doctrine bars

enforcement of the patent until the anti-competitive effects of the restriction have been
purged.
The patent misuse doctrine applies whether or not an antitrust violation has been estab-
lished. The expansion of the patent misuse doctrine in the 1940s led Congress to exclude
contributory patent infringement claims from the ambit of patent misuse in the 1952
Patent Act. (35 U.S.C. § 271(d)(1)–(3)). Nonetheless, the uncertain scope and severe remedy
of patent misuse continued to generate criticism, especially as economists and courts
came to question categorical antitrust prohibitions in favor of rule of reason balancing
(Hovenkamp, 2015, 2017; Bohannan, 2011; Lemley, 1990; but see Feldman, 2003; Merges,
1988 (noting that ‘the often very limited (or “thin”) markets for patented technology make
it difficult to apply antitrust law’s consumer-demand definition of the relevant market’)).
Congress amended the Patent Act in 1988 to codify two additional patent misuse
limitations (Patent Misuse Reform Act of 1988, Pub. L. No. 100-703, 102 Stat. 4674 (H.R.
4972)). Congress insulated refusals to license any rights to a patent from charges of patent
misuse (35 U.S.C. § 271(d)(4)). This provision, however, can effectively be side-stepped
through contract, such as a FRAND commitment. In that circumstance, the third-party
beneficiary of the SSO agreement has a breach of contract action for failure to license
SEPs on FRAND terms (see Section VI(A)(4)). Furthermore, Congress barred applica-
tion of the patent misuse doctrine to tying arrangements unless the patentee has market
power in the relevant market for the patent or has patented tying a product on which the
license or sale is conditioned (35 U.S.C. § 271(d)(5)), thereby bringing patent misuse more
closely in line with antitrust liability (Hovenkamp, 2017).
The courts have struggled to disentangle patent misuse doctrine from antitrust analysis
(Bohannan, 2011; Feldman, 2003; Mueller, 2002; Pitofsky, 2001). Although the two fields
share common concerns, the misuse doctrine has sought to promote intellectual property
policies of encouraging innovation, freedom to operate outside of intellectual property
protections, and access to the public domain even when objectionable practices do not
violate antitrust law (Bohannan, 2011).
In a case echoing the Motion Picture Patents case, a music copyright licensor sought
to require theaters to obtain a performance license before they even knew what music
would be incorporated into the films they would show (M. Witmark & Sons v. Jensen, 80
F. Supp. 843 (D. Minn. 1948), appeal dismissed sub nom., M. Witmark & Sons v. Berger
Amusement Co., 177 F.2d 515 (8th Cir. 1949)). The district court found that such a license
agreement improperly asserted control over all films and hence constituted copyright
misuse. As a result, the court barred enforcement against the theater owner.
As the emergence of computer software brought copyright more directly into play in
innovation markets, the copyright misuse doctrine has come into wider use. In Lasercomb
America, Inc. v. Reynolds (911 F.2d 970 (4th Cir. 1990)), a software copyright licensor
prohibited licensees from ‘writing, developing, producing or selling computer assisted die
making software, directly or indirectly without Lasercomb’s prior written consent’ for a
term of 99 years. Drawing on the principles underlying intellectual property protection as
well as patent misuse jurisprudence, the court determined that copyright misuse is a valid
defense and barred Lasercomb’s infringement action. In this case, the licensor sought to
foreclose competition in computer software innovation.
The copyright misuse doctrine has since been raised in a variety of settings, including
tying arrangements, anti-competitive clauses in licensing agreements, mandatory blanket

licenses, and refusals to license, but it remains murky (Herrell, 2011; Fellmeth, 1998).
Several cases in which copyright misuse has been found to be viable involve network
effects. For example, the Ninth Circuit found copyright misuse in Practice Mgmt. Info.
Corp. v. American Med. Assoc., 121 F.3d 516, amended by, 133 F.3d 1140 (9th Cir. 1997).
The court held that the licensing terms of the AMA’s Physician’s Current Procedural
Terminology to the Health Care Financing Administration gave the AMA a substantial
and unfair advantage over its competitors and hence constituted copyright misuse. The
Fifth Circuit in DSC Communications Corp. v. DGI Technologies, Inc. held that the
copyright misuse doctrine might be viable to defend assertions of copyright protection
over interoperable features of computer software (681 F.3d 597 (5th Cir. 1996)). The
Seventh Circuit in Assessment Techs., LLC v. WIREdata, Inc. held that licensing restric-
tions on a tax assessment database to control access to public domain data inputted by
public tax assessors could constitute copyright misuse (350 F.3d 640, 647 (7th Cir. 2003)).
As with the patent misuse doctrine, the interplay of copyright misuse doctrine and
antitrust liability remains unclear. Courts applying the copyright misuse doctrine gener-
ally evaluate whether the conduct thwarts the underlying policies of copyright law. Some
courts, however, mistakenly view the doctrine as co-extensive with antitrust law (Fellmeth,
1998, pp. 22–3).
2. The principle of exhaustion

With some resemblance to misuse, the long-standing common law doctrine of exhaustion
(also known as first-sale) preserves the public interest in free competition by limiting the
ability of intellectual property (IP) owners to control secondary markets for patented
products and copyrights works (17 U.S.C. § 109(a), Adams v. Burke, 84 U.S. 453 (1873)).
The idea is straightforward: the first authorized sale of the patented or copyrighted
product exhausts the monopolistic power given to the owner as a reward for his efforts
and contribution to society.
Following the first-sale, the owner can no longer control the manner in which the
product is sold or used in secondary markets, either by downstream purchasers or
subsequent sellers. This limitation on monopoly, which traces back to the common law’s
general hostility toward restraints on alienation, fosters competition in secondary markets
for innovative and creative works.
Recently, in a much-debated decision involving Lexmark cartridges, the Supreme
Court emphasized the destructive anti-competitive effects post-sale restrictions (with
servitude-like features) that ‘run with’ the product have on free commerce (Impression
Prods. v. Lexmark Int’l, Inc., 137 S. Ct. 1523, 1532 (2017); Van Houweling, 2008, 2019).
Refusing to allow patentees to ‘[sputter] the smooth flow of commerce,’ the Court held
that the principle of exhaustion prevents the enforcement of contractual post-sale
restrictions through patent law. The patentee can impose contractual restrictions on
secondary markets, but cannot use patent law to control how the product is being used
or sold downstream after the point of the first sale. The Lexmark decision reaffirmed
the vital role IP limiting doctrines such as preemption, exhaustion, and misuse play in
limiting IP owners’ ability to hinder downstream innovation thorough over-reaching,
often boilerplate, contractual language (Elazari Bar-On, 2019; Perzanowski and Schultz,
2016, chs 9–10).

3. Ambush of standard setting processes

As highlighted in Section V(D)(3), SSOs play a critical role in addressing the market
failures surrounding network technologies. These organizations require companies that
participate in standard setting processes to disclose relevant information about actual and
potential patents implicated by draft standards and commit to license such technologies
on FRAND terms (Patterson, 2012, pp. 513–21). Given the dynamic nature of technologi-
cal progress, such conditions can be open to interpretation.
The two most prominent cases alleging that SSO participants engaged in deceptive
practices—one involving Rambus and the other involving Qualcomm—reached different
conclusions. The Rambus litigation grew out of standards development for dynamic
random-access memory (DRAM) chips, a memory technology that was widely adopted
throughout the computer industry. In 1990, Rambus sought patents on its architecture for
such chips. Around that time, the Joint Electron Device Engineering Council (JEDEC)
organized an open standard setting process with a broad range of industry participants
including Rambus. As would be revealed in later litigation, Rambus used information
gained at the meetings to amend its patent applications so that the standards would read
on its patents. Rambus concealed these efforts and subsequently withdrew from JEDEC.
After the JEDEC standards gained widespread acceptance in products, Rambus began
making royalty demands from implementers and, beginning in 2000, brought a series of
enforcement actions. The jury in one of the key cases found that Rambus committed fraud
and breached JEDEC obligations by failing to disclose its patents. The Federal Circuit
reversed in a divided opinion, with the majority finding that the JEDEC policy statements
were too vague to support a fraud finding (Rambus Inc. v. Infineon Technologies AG, 318
F.3d 1081, 1098 (Fed. Cir. 2003)).
The Qualcomm litigation grew out of the development of the Joint Video Team (JVT)
standard for video compression technology. Qualcomm, a pioneer in semiconductor
design for mobile communications devices and various other technologies, participated
in the JVT standard setting process. It later sought to enforce several patents applicable
to that standard against Broadcom. Broadcom successfully defended on the ground that
Qualcomm had waived its rights to enforce the patent as a result of its failure to disclose
the patents as part of the standard setting process (Qualcomm Inc. v. Broadcom Corp.,
548 F.3d 1004 (Fed. Cir. 2008)). The court barred Qualcomm from enforcing the patents
at issue against any products implementing the pertinent JVT standard.
The different results in these litigations reflect several factors. SSO policies in the early
1990s varied in how clearly they set forth disclosure requirements. Furthermore, defend-
ants in these cases faced a variety of complex evidentiary requirements and heightened
pleading standards to equitable defenses such as laches, waiver, actual or implied license,
equitable estoppel, and fraud (Royall et al., 2009). The Rambus controversy led to public
enforcement actions in the United States and Europe that are discussed in Section VI(B)
(2).
4. Breach of contract for failure to license SEPs on FRAND terms

As explored earlier (Section V(D)(3)–(4)), SSOs typically require companies participating
in standard setting processes to commit to license SEPs on FRAND terms. The litigation
between Microsoft and Motorola established key principles regarding the determination
of FRAND licensing terms and provided for the award of contract damages for breach of

the FRAND commitment. Motorola participated in standard setting processes governed

by the IEEE and the International Telecommunication Union (ITU) establishing Wi-Fi
(802.11) and video compression (H.264) standards. As part of its participation in these
processes, Motorola agreed to license its patents that are essential to those standards on
reasonable and non-discriminatory (RAND)4 terms.
The controversy began in October 2010 when Microsoft filed actions in the U.S.
International Trade Commission and the Western District of Washington alleging that
Motorola was infringing several Microsoft smartphone patents. Those filings immediately
led to settlement negotiations involving cross-licensing of patents between the two com-
panies. Later that month, Motorola sent letters to Microsoft requesting royalties equal to
2.25 percent of Microsoft’s sales revenues from Windows and Xbox products incorporat-
ing the standards. Microsoft declined and immediately filed suit in the Western District of
Washington alleging that Motorola’s offer breached its RAND commitment. Microsoft
asserted that it was a third-party beneficiary of the SSO agreements. Thereupon Motorola
filed patent enforcement suits with the ITC, seeking an exclusion order against importing
Microsoft’s Xbox products into the United States, and with a German court, seeking an
injunction against sales of Microsoft’s H.264-compliant products. The German action
threatened all of Microsoft’s Windows and Xbox European sales because its distribution
center was located in Germany. As a result, Microsoft immediately relocated its distribu-
tion center to the Netherlands at substantial cost.
Judge Robart adapted the Georgia-Pacific reasonable royalty framework to the
FRAND context (Microsoft Corp. v. Motorola, Inc., 2013 WL 2111217 (W.D. Wash.
2013)). In so doing, he set forth the following principles:
(1) ‘A RAND royalty should be set at a level consistent with the SSOs’ goal of promoting
widespread adoption of their standards’;
(2) a proper methodology should ‘recognize and seek to mitigate the risk of patent
hold-up that RAND commitments are intended to avoid’;
(3) ‘a proper methodology for determining a RAND royalty should address the risk of
royalty stacking by considering the aggregate royalties that would apply if other SEP
holders made royalty demands of the implementer’;
(4) ‘At the same time, a RAND royalty should be set with the understanding that SSOs
include technology intended to create valuable standards,’ which requires that the
RAND commitment [] guarantee that holders of valuable intellectual property will
receive reasonable royalties on that property’; and
(5) ‘From an economic perspective, a RAND commitment should be interpreted to
limit a patent holder to a reasonable royalty on the economic value of its patented
technology itself, apart from the value associated with incorporation of the patented
technology into the standard’ (2013 WL 2111217 at 12; Page, 2014).
Applying these economic guideposts, Judge Robart concluded that the reasonable royalty
should be approximately 1/100th of Motorola’s 2.25 percent license offer. Motorola’s
4
FRAND and RAND are used interchangeably in the technology industries. FRAND is the
more common usage.

patents constituted less than 10 percent of the WiFi 802.11 pool and none were shown
to be of special importance. Judge Robart noted that if each of 92 companies that
owned SEPs for the 802.11 and H.264 standards demanded a royalty rate comparable to
Motorola’s offer, the sum of the royalties would exceed the selling price of the Xbox (2013
WL 2111217 at 52, 72–73). In a later proceeding, a jury awarded Microsoft $14.52 million
($11.49 million for relocating its distribution center and $3.03 million in attorneys’ fees
and litigation costs) for Motorola’s breach of its contractual commitment to license its
SEPs on RAND terms.
The Ninth Circuit upheld these determinations, expressly recognizing the key role of
RAND commitments in ‘mitigating the risk that a SEP holder will extract more than the
fair value of its patented technology . . . Under these agreements, an SEP holder cannot
refuse a license to a manufacturer who commits to paying the RAND rate’ (Microsoft
Corp. v. Motorola, Inc., 795 F.3d 1024, 1031 (9th Cir. 2015) (affirming Judge Robart’s
decision)).
5. Private antitrust liability

Section 4 of the Clayton Antitrust Act of 1914 authorizes recovery of damages by ‘any
person injured in his business or property by reason of anything forbidden in the antitrust
laws,’ including the Sherman Act (15 U.S.C. § 4 (1914)). Companies have used this private
right of action to combat anti-competitive practices in network industries. Three issues
are particularly relevant to network technology markets: (i) refusal to license patented
technologies; (ii) patent thickets; and (iii) leveraging of monopoly power.
i. Refusals to license patented technologies and copyright-protected works As noted

above, the Patent Act grants patentees the exclusive right to use patented technologies.
Congress reinforced that power by expressly providing that a refusal to license a patent
cannot be the basis for a patent misuse defense. The courts are divided over whether a
refusal to license patented technology can constitute an antitrust violation. The so-called
‘essential facilities’ doctrine, which holds that ‘an owner of a crucial input cannot deny
access if a firm seeking access cannot practicably obtain the input elsewhere’ (Ratner,
1988, p. 330; Otter Tail Power Co. v. United States, 410 U.S. 366 (1973)), has lost favor
among commentators (Areeda, 1990; Ratner, 1988) and the courts (see Eastman Kodak
Co. v. Image Technical Services, Inc., 504 U.S. 451 (1992) (declining to apply essential
facilities doctrine); Aspen Skiing Co. v. Aspen Highlands Skiing Corp., 472 U.S. 585
(1985) (same)). The courts focus on whether the defendant has a legitimate business
justification for its conduct (Baker, 1999). The key modern cases involve the control of
service and replacement part aftermarkets.
In the late 1980s, a variety of companies that had entered the market to service Kodak
photocopiers found themselves cut off from replacement parts. Kodak ended its practice
of licensing and selling replacement parts to competing service companies and required
that its original equipment manufacturers not sell parts to independent service operators.
The independent service organizations brought suit, claiming that Kodak unlawfully
tied the sale of service for Kodak machines with the sale of parts in violation of § 1 of
the Sherman Act, and monopolized or attempted to monopolize the sale of service for
Kodak machines in violation of § 2 of the Sherman Act. Kodak defended in part on its
intellectual property rights.

While recognizing that patent and copyright owners have exclusive rights to their pro-
tected works, the Ninth Circuit nonetheless held that these laws only afford the intellectual
property owner a rebuttable ‘presumptively valid business justification’ for consumer
harm (Image Tech. Servs., Inc. v. Eastman Kodak Co., 125 F.3d 1195, 1218 (9th Cir. 1997)
(quoting Data General Corp. v. Grumman Systems Support Corp., 436 F.3d 1147, 1187
(1st Cir. 1994))). The court upheld a jury verdict finding Kodak liable for monopolizing
or attempting to monopolize the service aftermarket for Kodak copiers.
In an analogous case involving Xerox’s refusal to sell patented parts and copyrighted
manuals and to license copyrighted software, the Federal Circuit declined to follow the
Ninth Circuit’s analysis (In re Independent Service Organizations Antitrust Litigation,
203 F.3d 1322 (Fed. Cir. 2000); see also Intergraph Corp. v. Intel Corp., 195 F.3d 1346
(Fed. Cir. 1999)). The Federal Circuit limited its focus to whether Xerox’s refusal to sell its
patented parts exceeded the scope of the patent grant. Finding that it did not, ‘Xerox was
under no obligation to sell or license its patented parts and did not violate the antitrust
laws by refusing to do so’ (203 F.3d at 1328). The court ruled that so long as a patent
infringement suit would not have been objectively baseless, the patentee’s motivations
for asserting its statutory right to exclude are immaterial. Similarly, the Federal Circuit
further held that so long as Xerox’s copyrights were not ‘obtained by unlawful means or
were used to gain monopoly power beyond the statutory grant,’ then ‘Xerox’s refusal to
sell or license its copyrighted works was squarely within the rights granted by Congress
to the copyright holder and did not constitute a violation of the antitrust laws’ (203 F.3d
at 1329).
ii. Patent thickets In a related vein, competitors have sought to challenge the accumu-
lation of a broad portfolio of patents on antitrust grounds. Accumulation and pooling
of patents can broaden the effective scope and reduce the uncertainty surrounding inven-
tions, thereby enhancing appropriability (Parchomovsky and Wagner, 2005, pp. 32–41).
Nonetheless, strong and broad patent portfolios can discourage innovation and entry by
potential competitors (Hall et al., 2016; Hovenkamp, 2012, p. 1130 (discussing the costs
of defending against many patents of ambiguous scope); Rubinfeld and Maness, 2005).
The leading case involved Xerox Corporation, which built a portfolio of over 1,000
patents relating to its plain paper copying technology. Xerox only used 35 to 40 percent
of those patents in actual Xerox products (Sobel, 1984), relying on the balance to erect
a defensive thicket around its photocopier technology (Saunders, 2002). Xerox refused
to grant licenses for plain paper copying, although it did grant some licenses for other
fields, including coated paper copiers. SCM Corporation, which had licensed some of
Xerox’s patents for coated paper copies, filed an antitrust claim against Xerox alleging
that ‘Xerox’s acquisition of its patents and subsequent exercise of the exclusionary power
in them violated the antitrust laws and injured SCM’ (SCM Corp. v. Xerox Corp., 645 F.2d
1195, 1203 (2d Cir. 1981)). SCM asserted that Xerox’s patent accumulation strategy was
intended to forestall competition, as reflected in its failure to use many of its patents. SCM
claimed that ‘Xerox’s patents were so numerous and complex that they created a “thicket”
that prevented [SCM from] designing around the patents.’ The Second Circuit acknowl-
edged that ‘tension between the objectives of preserving economic incentives to enhance
competition while at the same time trying to contain the power a successful competitor
acquires is heightened tremendously when the patent laws come into play,’ emphasizing

that the Xerox case ‘demonstrate[s that] the acquisition of a patent can create the potential
for tremendous market power’ (645 F.2d at 1205). Nonetheless, the court ultimately ruled
that ‘where a patent has been lawfully acquired, subsequent conduct permissible under
the patent laws cannot trigger any liability under the antitrust laws’ (645 F.2d at 1206).
The anti-competitive concerns relating to patent thickets are exacerbated by the
ambiguity of many software patent claims (Bessen, 2004). Cross-licensing and patent
pools can, however, alleviate concerns about patent thickets (Barnett, 2014, 2015; Gilbert,
2004; Shapiro, 2001; Merges, 2001). Economists generally believe that the inclusion of
complementary and potentially blocking patents in a patent pool promotes competition
by reducing the transaction costs and promoting licensing (Shapiro, 2001, p. 144).
iii. Improper leveraging of market power In the mid-1990s, Microsoft Corporation

held a dominant position in the desktop software marketplace just as the Internet emerged
as an economic platform. Sun Microsystems’s Java programming language for websites
was rapidly gaining salience as a technology for easily transforming static webpages into
engaging, animated, interactive websites. After failing to develop its own web develop-
ment package, Microsoft entered into a Technology License and Distribution Agreement
(TLDA) with Sun that allowed Microsoft to use, modify, and adapt Java technology in
developing MS Internet Explorer 4.0 and other software products. To safeguard Sun’s
WORA interoperability principle, the TLDA required that Microsoft adhere to Java’s
standardized application environment and compliance tests.
Microsoft’s deployment of its own version of Java, compatible only with other Microsoft
products in violation of the WORA principle, threatened Sun’s Java development strategy.
In October 1997, Sun sued Microsoft for breach of contract, trademark infringement,
copyright infringement, false advertising, and unfair competition (Markoff, 1997). In
early 2002, Microsoft agreed to pay Sun $20 million and was permanently prohibited from
using ‘Java compatible’ trademarks on its products (Shankland, 2002a). The following
year, Sun brought an antitrust and patent infringement action against Microsoft resulting
in an award of over $1 billion (Pruitt and Roberts, 2004; Shankland, 2002b).
B. Public Enforcement
Federal and state antitrust authorities have long played substantial roles in policing
network market competition. The US Department of Justice’s filing of an antitrust action
against IBM in 1969 reshaped the competitive landscape of the computer hardware indus-
try and paved the way for a vibrant software industry. At the time, IBM bundled software
and services into the cost of leasing use of its hardware, making it difficult for competitors
to charge for software development and products. Immediately following the filing of
the enforcement action, IBM unbundled software and services from its hardware sales
thereby opening up markets for software products (Grad, 2002). Although the antitrust
case dragged on for more than a decade and was ultimately dropped, IBM’s unbundling
decision in conjunction with the emergence of mini- and microcomputers markets
revolutionized the computer industry. Similarly, the US Department of Justice’s filing of
antitrust litigation against AT&T in 1974 led to the breakup of the largest corporation in
the United States nearly a decade later and hastened the modern competitive and highly
innovative telecommunications marketplace.

Advances in the information, financial, and communications technologies have vastly

increased the significance of network markets as well as the government’s role in regulat-
ing these markets. As highlighted in Sections II and III, network technologies are prone
to high concentration levels that can enhance consumer welfare through network effects.
Therefore, antitrust authorities have had to shift their focus away from market concentra-
tion toward anti-competitive tactics such as leveraging market power into new markets
and stifling innovation. This section summarizes the major contours of this shift. Section
1 discusses the evolution of Department of Justice and Federal Trade Commission guide-
lines for intellectual property licensing. Section 2 discusses significant network market
enforcement actions and the challenges of crafting remedies.
1. Intellectual property licensing guidelines

Beginning in the 1930s, antitrust regulators took a skeptical view of intellectual property
(Briggs, 2009, pp. 67–8). By the early 1970s, these concerns reached their apex in the US
Department of Justice Antitrust Division’s ‘Nine No-No’s’:
(1) It is unlawful to require a licensee to purchase unpatented materials from the licensor;
(2) It is unlawful for a patentee to require a licensee to assign to the patentee any patent which
may be issued to the licensee after the licensing arrangement is executed;
(3) It is unlawful to attempt to restrict a purchaser of a patented product in the resale of that
product;
(4) A patentee may not restrict his licensee’s freedom to deal in the products or services not
within the scope of the patent;
(5) It is unlawful for a patentee to agree with his licensee that he will not, without the licensee’s
consent, grant further licenses to any other person;
(6) Mandatory package licensing is an unlawful extension of the patent grant;
(7) It is unlawful for a patentee to insist, as a condition of the license, that his licensee pay
royalties in an amount not reasonably related to the licensee’s sales of products covered by
the patent—for example, royalties on the total sales of products of the general type covered
by the licensed patent;
(8) It is unlawful for the owner of a process patent to place restrictions on his licensee’s sales
of products made by the use of the patented process; and
(9) It is unlawful for a patentee to require a licensee to adhere to any specified or minimum
price with respect to the licensee’s sale of the licensed products. (Wilson, 1970)
Furthermore, even if a patent-related restraint was not per se unlawful under one of the
Nine No-No’s, the Department of Justice would still consider bringing an enforcement
action if the particular provision was not necessary to the patentee’s exploitation of
its lawful monopoly and there were less restrictive alternatives to the restrictions that
were more likely to foster competition (Briggs, 2009; Wilson, 1970). These enforcement
principles focused on attempts by patent holders to extend their patent monopolies to
unpatented supplies, to gain control over improvements of their innovations, to determine
prices for resale of their patented products, or to engage in market allocations.
With the growing importance of intellectual property assets in the 1970s and 1980s and
the dawning of the digital age, economists came to see unconstrained patent licensing as
an innovation driver (Gilbert and Shapiro, 1997, p. 286). In 1988, the Antitrust Division
shifted from absolute (per se) opposition to licensing restrictions to a ‘rule of reason’
approach to patent licensing that balanced the pro-competitive effects of licensing against
potential anti-competitive effects in related markets (U.S. Dep’t. Justice, 1988). In 1995,
the Department of Justice and the Federal Trade Commission (FTC) expanded upon

the 1988 guidelines in crafting the ‘Antitrust Guidelines for the Licensing of Intellectual
Property’ (U.S. Dep’t. Justice and Federal Trade Commission, 1995; U.S. Dep’t. Justice
and Federal Trade Commission, 2007). These guidelines expressly recognized the
generally pro-competitive nature of licensing arrangements, rejected the presumption
that intellectual property necessarily creates market power in the antitrust context,
and endorsed applying the same general antitrust approach to the analysis of conduct
involving intellectual property that the agencies apply to conduct involving other forms
of tangible or intangible property.
With the growing role of patents in network industries, the Department of Justice and
the FTC increasingly recognized the importance of licensing standards-essential patents
on FRAND terms (U.S. Dep’t. Justice and PTO, 2013; Federal Trade Commission, 2012a;
Federal Trade Commission, 2011; U.S. Dept. of Justice and Federal Trade Commission,
2007). Accordingly, the Department of Justice and the FTC have been far more receptive
to patent pools (see, e.g., U.S. Dep’t. Justice, 2006; U.S. Dep’t. Justice, 2007; U.S. Dep’t.
Justice, 2015; In re Negotiated Data Solutions LLC (N-Data), No. C-4234, 2008 WL
4407246 (F.T.C.); In re Motorola Mobility LLC, F.T.C. File No. C-4410 (2013)). As
Gilbert (2010) explains:
Competition policy toward patent pools has focused on the prevention of anticompetitive
practices by patent pool members—individually or collectively through the licensing policies of
the pool—and has generally paid little attention to the question of how to encourage the forma-
tion and stability of patent pools that benefit consumers. While patent pools have substantial
procompetitive benefits when the manufacture or use of products may infringe multiple patents,
powerful economic forces prevent beneficial patent pools from forming or limit the patents in
the pool to only a fraction of the patents that cover the products.
Competition policy should recognize the fragility of patent pools and ensure that patent pool
members acting collectively have the same latitude to determine royalties and licensing terms
as a single licensor, provided that the pool does not harm lawful competition that would have
occurred in the absence of the pool’s licenses. In determining which types of patents should
be allowed in a pool, competition policy should recognize that a patent pool confers potential
benefits if it includes two or more valid complementary patents, and need not harm competition
if it has at least one valid patent that is essential to make, sell, or use a product. Inclusion of
inessential patents raises potential concerns about foreclosure of alternative technologies and
higher royalties for some licenses than would have occurred if these patents were excluded from
the pool. However, these concerns should be balanced against the costs of excluding potentially
essential patents from the pool.
Disappointingly, the US Dep’t of Justice and the FTC’s updated intellectual property
guidelines (U.S. Dep’t. Justice and Federal Trade Commission, 2017) (largely reaffirm-
ing and modestly updating the 1995 guidelines)) omit mention of SEPs and FRAND
(Comments of Law and Business Scholars Submitted to the U.S. Department of Justice
and Federal Trade Commission Regarding a Proposed Update to the Antitrust Guidelines
for the Licensing of Intellectual Property, 2016).
2. Significant network market enforcement actions

Notwithstanding the Department of Justice’s and FTC’s loosening of licensing restric-
tions, antitrust authorities have pursued several notable enforcement actions in network
industries over the past two decades.
In 1996, the FTC alleged that Dell Computer Corporation had violated the Federal

Trade Commission Act by failing to disclose its patent rights during the Video Electronics
Standards Association standard setting process and then threatening to enforce those
rights against others involved in that process (Matter of Dell Computer Corp., 121
F.T.C. 616, 1996 WL 33412055 (May 20, 1996)). The resulting consent decree barred Dell
from enforcing its patent against computer manufacturers incorporating the pertinent
standard.
In 1998, the FTC issued a complaint against Intel Corporation alleging that Intel had
sought to maintain its dominance in the microprocessor marketplace by denying essential
technical information and product samples of new microprocessors to companies that,
because of intellectual property disputes, had initiated or threatened to initiate litigation
against Intel or Intel’s customers (In re Intel Corp., No. 9288, 1999 F.T.C. LEXIS 145
(Aug. 3, 1999); Pitofsky, 2001). The resulting consent decree recognized Intel’s right
to withhold licenses of its product or information, but limited Intel’s ability to retract
licenses when customers sought to vindicate its intellectual property rights.
Most significantly, the Department of Justice and 18 states brought actions against
Microsoft in 1998 alleging that Microsoft’s bundling of its browser (Internet Explorer)
with its Windows operating system along with restrictive licensing agreements with
original equipment manufacturers (such as pricing use of its operating system based
on a per processor basis) violated antitrust law (United States v. Microsoft Corp., No.
98-1232 (D.D.C. May 18, 1998); New York v. Microsoft Corp., No. 98-1233 (D.D.C. May
18, 1998)). The government specifically targeted Microsoft’s efforts to exclude Netscape
from the browser market and to suppress Sun’s Java web programming platform. The
federal government settled its claims with Microsoft in 2001. The consent decree required
Microsoft to share its application programming interfaces with third-party companies
and established a process for supervising compliance with the agreement over a five year
period. Nine states proceeded to trial and ultimately implemented somewhat greater
oversight over Microsoft’s activities.
Crafting a remedy proved especially difficult due to the strong consumer benefits
attributable to Microsoft’s widely adopted and highly integrated computing platform (see
United States v. Microsoft Corp., 253 F.3d 34, 102 (D.C. Cir. 2001) (reversing the district
court order that would have broken Microsoft up because it failed to address Microsoft’s
contention that such an order would ‘lower [] rates of innovation and disrupt [] the
evolution of Windows as a software development platform’)). Breaking up the company
would certainly have caused substantial consumer harm. In the end, the rapid emergence
of the Internet Age and mobile computing—along with the ascendance of a new set of
competitors such as Google and Facebook, as well as the resurgence of Apple—eroded
Microsoft’s dominance.
In 2012, the FTC required that Robert Bosch GmbH sell SPX Service Solutions, a
business that makes equipment used to recharge vehicle air conditioning systems, grant
licenses to key patents needed to compete in the market for such equipment on the ground
that SPX harmed competition by reneging on a commitment to license SEPs on FRAND
terms. The FTC declared that ‘[p]atent holders that seek injunctive relief against willing
licensees of their FRAND-encumbered SEPs should understand that in appropriate cases
the Commission can and will challenge this conduct as an unfair method of competition
under Section 5 of the FTC Act’ (Federal Trade Commission, 2012b).
The Department of Justice has conditioned its approval of acquisitions of

s ubstantial patent portfolios by firms with substantial market presence on the commit-

ments to license SEPs on FRAND terms. Prominent examples include: (1) Google’s
acquisition of Motorola Mobility’s portfolio of 17,000 patents and 6,800 patent
applications; (2) Apple’s acquisition of the nearly 900 patents originally held by Novell
and purchased in 2010 by a coalition including Apple, EMC, Microsoft, and Oracle;
and (3) acquisition by the ‘Rockstar’ group (made up of Apple, Microsoft, and RIM)
of the 6,000 patents and applications made available in the Nortel bankruptcy auction
(Carrier, 2014).
VII. ASSESSMENT OF INTELLECTUAL PROPERTY

PROTECTION AND COMPETITION POLICY FOR
NETWORK TECHNOLOGIES
Drawing on the evolution of intellectual property protection and competition policy
explored in Sections V and VI, this section assesses how the various legal, market,
and policy institutions have adapted to the emergence of network technologies in the
Information Age. Section A discusses institutional and policy economy considerations.
Section B then assesses the performance of legal and policy institutions against the
normative principles highlighted in Section IV.
A. Institutional Considerations
Intellectual property is not a single, monolithic protective system but rather a complex,
overlapping set of protections. Inventors and platform developers can utilize various
modes of protecting their innovative endeavors. In addition, they can coordinate with
other entrepreneurs to promote and leverage network effects, subject to antitrust
constraints.
Protectionist entrepreneurs will naturally exploit the weakest link within the intellectual
property chain to gain market advantage. As a result, the efficacy of the intellectual
property system depends critically upon intellectual property gatekeepers—judges, patent
examiners, and antitrust enforcers—to ensure that the system coheres.
Therefore, the intellectual property system can be strained and fail to promote balanced
protection where critical gatekeepers lack adequate understanding of the overall system
or the technologies and economics at issue. The structure of the federal courts creates two
opposing vulnerabilities. On the one hand, the regional circuit courts—which handle most
copyright and trademark disputes—lack specialization and technological training. They
can struggle to understand the complexities of computer software and other technical
subject matter in the network technology fields. On the other hand, the Federal Circuit—
which handles all patent appeals and some copyright and trademark appeals—is special-
ized, which can skew their perspective. As numerous scholars have explored, specialty
courts, such as the Federal Circuit, are prone to tunnel vision and political capture which
could lead to more protectionist interpretations of intellectual property law (Dreyfuss,
1989, p. 26; Allison and Lemley, 1998, p. 251; Merges, 2000, p. 2224; Landes and Posner,
2003, pp. 334–53; Landes and Posner, 2004, p. 128).
As the Open Handset Alliance and the open source movement have demonstrated, free

market institutions can check overbroad intellectual property protection. Entrepreneurs

can contract around intellectual property systems in creative ways (Merges, 2004).
B. Measuring Progress Based on the Normative Principles
The past half century, spanning the birth and ascendancy of the Information Age, has
included dramatic evolution of intellectual property protection for network technolo-
gies. The process has not always been smooth, but has generally been inclined toward
more efficient and effective rules and institutions. Nonetheless, the complexity of the
intellectual property system and the dynamism of network technologies has produced
persistent pathologies. Fortunately, the flexibility afforded by free market competition
and new technologies (such as cloud-based computing) have been valuable antidotes
and alternatives to unwarranted intellectual property protection and the accompanying
market power. The evolutionary process continues to unfold and adherence to the key
normative principles will benefit from the lessons of the past and ongoing vigilance.
1. Parsimony principle

The parsimony principle aims to promote realization of network benefits by denying
intellectual property protection for functional attributes of network technologies absent
significant technological advance. This principle comes into conflict with the motivation
of some platform developers to control platform development and profit from network
effects.
Thus, leading platform technology companies advocate robust intellectual property
protection for network features of computer software and other technologies through
copyright, trademark, and design patent law. These legal regimes do not require assess-
ment of novelty or nonobviousness. In an effort to garner long-lived copyright protection
for interface and other software components, they have characterized software code as
‘high-tech poetry’ and analogized computer programs to epic poems and great literature
(Clapes et al., 1987, p. 1500, 1584; see also Clapes, 1994).
Some general jurisdiction judges, with little technical background, were initially
receptive to such arguments. They perceived the textual form of software code as more
analogous to more conventional literary works than the gears and levers of machines and
were less attuned to the broader intellectual property landscape channeling protection for
functional features to the utility patent system. Dicta in the Apple v. Franklin decision
opined that ‘total compatibility with independently developed application programs . . .
is a commercial and competitive objective which does not enter into the somewhat meta-
physical issue of whether particular ideas and expressions have merged’ (Apple Computer,
Inc. v. Franklin Computer Corp., 714 F.2d 1253 (3d Cir. 1983)). In Whelan Associates,
Inc. v. Jaslow Dental Laboratory, Inc., 797 F.2d 1222 (3d Cir. 1986), the Third Circuit’s
conflation of merger analysis and the idea-expression dichotomy implicitly allowed
copyright protection of procedures, processes, systems, and methods of operation that
are expressly excluded under § 102(b).
Fortunately, a series of cases in the early to mid-1990s better appreciated the distinction
between functionality and creative expression (see Computer Assocs. Int’l v. Altai, Inc.,
982 F.2d 693 (2d Cir. 1992); Sega Enterprises Ltd. v. Accolade, Inc., 977 F.2d 1510 (9th
Cir. 1993); Apple Computer, Inc. v. Microsoft Corp., 799 F. Supp. 1006 (N.D. Cal 1992),

aff’d in part, rev’d in part, 35 F.2d 1435 (9th Cir. 1994); Lotus Dev. Corp. v. Borland
Int’l, Inc., 49 F.3d 807, 813 (1st Cir. 1995), aff’d by equally divided Court, 516 U.S. 233
(1996)). As a result, while programming a computer can unquestionably be considered
‘creative’ in a general sense, limiting doctrines ensure that the functional aspects are
unprotectable under copyright law. The design of an efficient mechanical machine like-
wise can be creative, but such devices are not eligible for copyright protection unless the
aesthetic features can be separated from the functional attributes under the useful article
doctrine (17 U.S.C. § 101 (definition of ‘Pictorial, graphic, and sculptural works’ excludes
functional features)). Lines of code are the gears and levers of digital machines. The fact
that computer software, like a sculptural work, is eligible for copyright protection does
not authorize protection for functional features (17 U.S.C. § 102(b)).
Several major technological advances beginning in the mid-1990s de-emphasized the
role of copyright protection for computer software. The emergence of the Internet as a
low-cost, highly scalable distribution ecosystem in the mid to late 1990s vastly expanded
the potential for indirect appropriability (e.g., through keyword advertising), and shifted
software developers toward open source development. Advances in mobile, Internet-
connected digital devices in the early-2000 period paved the way for using software to
promote sales of hardware and vastly expanded software distribution through app stores.
The new app economy opened a vast array of non-copyright-based business models,
such as new forms of advertising ranging (e.g., Yelp). The emergence of cloud-based
computing (Software as a Service) reinvigorated digital rights management. These shifts,
in combination with the norms that took hold following the Lotus v. Borland litigation,
produced a period of relative peace with regard to copyright protection of network
features of computer software (Profitt, 2011 (observing that ‘[h]istorically, APIs have
been regarded as not falling under copyright—the reasoning being that APIs are not
creative implementations but rather statements of fact,’ but also noting the issue had been
clouded by the distinction of ‘open’ and ‘closed’); Menell, 1998). The parsimony principle
prevailed.
That peace was shattered in 2010 with Oracle’s filing of a copyright (and patent)
infringement lawsuit against Google alleging that the Android operating system infringed
copyright protection for the declarations (function names and definitions) in the Java
APIs. Drawing on the strategy of the first wave of API copyright litigation, Oracle
analogized the labels and code used in the Java APIs to the chapter titles, character names,
and plot elements of Harry Potter novels (Opening Brief and Addendum of Plaintiff-
Appellant, Oracle America, Inc. v. Google, Inc., 2013, pp. 12–13). Based on a questionable
interpretation of Ninth Circuit precedent (Menell, 2018), the Federal Circuit ruled that
the structure, sequence, and organization of the 37 Java APIs were copyrightable and
remanded the fair use issue for retrial (Oracle America, Inc. v. Google, Inc., 750 F.3d 1339
(Fed. Cir. 2014)).
Apple’s garnering of design patent protection for the rounded, rectangle shape of
its iPhone and iPod devices and visual icons also undercut the parsimony principle
(U.S. Design Patent Nos. D618,677, D593,087, and D604,305; Apple Inc. v. Samsung
Elecs. Co., 786 F.3d 983 (Fed. Cir. 2015) (affirming decision that design patents were
valid and had been infringed)). These functional elements garnered substantial protec-
tion without any showing that they constituted novel and nonobvious technological
advances.

These decisions directly undermine the parsimony principle. As a result of the Oracle
v. Google decision, the safe harbor of clean room implementation of functional specifica-
tions is no longer safe. The Oracle v. Google precedent creates the potential for software
developers to assert long-lived copyright protection over interface specifications without
meeting a substantial threshold of technological advance.
Thus, the Oracle v. Google decision warns innovators to steer clear of proprietary
software in developing platforms and extensions. Future developers will be careful to
avoid using APIs that are vulnerable to copyright assertion. This will reduce the flexibility
to join or interoperate with platforms that are not open, but will encourage greater use of
open platforms, collaboration, and ex ante resolution of legal rights. Thus, even though
the parsimony principle has been undermined, the flexibility to work around copyright
protection through open source and collaborative solutions limits its adverse effects.
2. Proportionality principle

The proportionality principle is the flip side of the parsimony principle coin. Balanced
protection for true technological advances in network technologies might be needed
to overcome the excess inertia generated by network bandwagons. Patent law provides
protection for novel, nonobvious, and adequately disclosed advances in computer sys-
tems, processes, and interface design, and other network technologies. Unlike copyright,
trademark, or design patent law, utility patent protection protects the functional aspects
for network technologies. In theory, therefore, patent protection can provide meaningful
protection for overcoming excess inertia. Its efficacy, however, depends on whether it
provides the right balance.
In practice, patent protection for interface design and other network technologies has
been decidedly mixed. The standards for patent protection might be too low or too high
and the duration of protection might be too short or long to provide the optimal incentive.
Moreover, unlike lock-out code, the scope of patent protection does not necessarily align
with network features. Furthermore, the costs of pursuing and enforcing patents can
distort incentives.
Patent protection of computer software, a principal source of network effects, has
experienced a roller coaster over the past four decades. The Patent Office resisted patent
protection for computer software in the 1960s and only grudgingly afforded such protec-
tion in the 1970s and 1980s (Moskowitz, 1982, pp. 281–2, 309–11). The Supreme Court
struggled to resolve the eligibility of patent protection for computer software in the
1970s (Gottschalk v. Benson, 409 U.S. 63 (1972); Parker v. Flook, 437 U.S. 584 (1978)),
but ultimately cautiously held that computer programs were eligible in 1981 (Diamond
v. Diehr, 450 U.S. 175 (1981)). Nonetheless, software companies were reluctant to pursue
such protection, preferring technical protection measures and copyright protection.
Several factors shifted the software industry toward patent acquisition in the early
1990s. Fading hardware companies turned to patent licensing and enforcement campaigns
(Phelps and Kline, 2010; Jaffe and Lerner, 2004, pp. 14–15; Rivette and Kline, 1999).
In addition, some smaller software companies succeeded in enforcing software patents
against larger software companies (see, e.g., Stac Elec. v. Microsoft Corp., No. 93-0413
(S.D. Cal. 1994), appeal dismissed per stipulation, 38 F.3d 1222 (Fed. Cir. 1994)). These
developments prompted software companies to pursue defensive patenting (Federal
Trade Commission, 2011, pp. 43–5, 56 (discussing defensive patenting)). Furthermore,

the Federal Circuit liberalized the standards for protecting computers software (see In re
Alappat, 33 F.3d 1526 (Fed. Cir. 1994); State Street Bank and Trust Company v. Signature
Financial Group, Inc., 149 F.3d 1368 (Fed. Cir. 1998)), just as the Internet (dot-com) era
was taking off. This led to a software patenting gold rush in which start-up companies
sought patents as signals for raising venture capital and established companies stockpiled
patents for defensive purposes.
As discussed in Section V(D), the bursting of the dot-com bubble in 2000 resulted in
many software patents falling into the hands of patent aggregators, such as Intellectual
Ventures, which produced an unprecedented wave of costly and disruptive patent asser-
tion activity. The low quality and amorphous scope of many of these patents imposed
tremendous costs on the software industry and complicated entry into many network
technology markets. In addition, new network technologies, such as smart phones,
developed in a patent thicket ecosystem.
The effects of patent aggregation and assertion were somewhat alleviated by standard
setting organizations requiring FRAND cross-licensing, the emergence of defensive
buying funds, such as RPX and Allied Security Trust, and patent pledges (Schultz
and Urban, 2012). Moreover, the Supreme Court substantially reduced the risk of
injunctive relief (eBay, Inc. v. MercExchange, LLC, 547 U.S. 388 (2006)), tightened the
nonobviousness standard (KSR Int’l Co. v. Teleflex Inc., 550 U.S. 398 (2007)), promoted
clearer patent boundaries (Nautilus, Inc. v. Biosig Instruments, Inc., 134 S. Ct. 2120
(2014)), and restricted patent eligibility (Bilski v. Kappos, 561 U.S. 593, 604 (2010);
Mayo Collaborative Services v. Prometheus Laboratories, Inc., 566 U.S. 66 (2012); Alice
Corp. v. CLS Bank International, 573 U.S. 208 (2014)). Congress passed legislation
streamlining administrative patent review (America Invents Act, Pub. L. No. 112-29,
125 Stat. 284 (2011)).
Nonetheless, patent protection for network technologies has proven to be a complex
and costly tool for achieving proportional appropriability for network technology innova-
tions. The system has, however, become more balanced and predictable, with improved
screening of patent applications, more timely and cost-effective means for invalidating
dubious patents through inter partes review at the Patent Trial and Appeal Board, and
improved coordination through standard setting and FRAND licensing.
3. Deterrence principle

The deterrence principle stems from, and interacts with, the proportionality principle.
Network effects often lead to high market concentration levels, which bring market power
with them. The deterrence principle seeks to stunt abuse of such power while promoting
network benefits. One of the main antidotes to market dominance by a single platform
sponsor is collaboration through standard setting organizations and licensing agreements,
such as FRAND commitments. While such private solutions can promote innovation and
downstream competition, they create the potential for anti-competitive behavior.
The past several decades have witnessed substantial evolution of antitrust doctrines
and enforcement policies toward a balanced innovation and competitive ecosystem.
Antitrust enforcers have come to appreciate the economic benefits of high concentra-
tion in network technology markets while also focusing on abusive practices, such as
failure to disclose essential patents to standard setting organizations. Standard setting
organizations have developed more sophisticated disclosure requirements. In addition,

courts have broadened their assessment of antitrust, contract, and patent remedies in
view of network effects.
The dynamism of network technologies and markets, however, will continue to
challenge enforcers, policymakers, and courts. As reflected in the Sun v. Microsoft and
Oracle v. Google litigation, there is a subtle line between promoting interoperability and
encouraging innovative forking of established standards (cf. Farrell, 2007).
VIII. FUTURE RESEARCH DIRECTIONS
Following Moore’s and Metcalfe’s ‘Laws,’ network technologies are growing at exponen-
tial rates. Digital technologies increasingly drive economic growth. Due in substantial
part to the Internet and advances in digital technology, network effects are rapidly
diffusing across the economic landscape. Consequently, the interplay of network
technologies and intellectual property will continue to evolve rapidly in the coming
years and decades.
The opportunities for further research in this field are nearly limitless. Network effects
are increasingly important across a growing swath of industries: consumer and industrial
products (Internet of Things), energy (smartgrid, autonomous driving, renewable energy),
bioinformatics, machine learning, social media, advertising, content creation, and science
(database development). The interactions with the range of economic modes (such as
contract, business associations, and multi-sided markets), as well as other areas of law
(such as privacy and civil liberties) provide a wealth of important research opportunities
to explore.
REFERENCES
Aitken, Hugh G.J. 1985. The Continuous Wave: Technology and American Radio, 1900–1932. Princeton, NJ:
Princeton University Press.
Allison, John R., and Mark A. Lemley. 1998. ‘Empirical Evidence on the Validity of Litigated Patents,’ 26
AIPLA Quarterly Journal 185–275.
Areeda, Phillip. 1990. ‘Essential Facilities: An Epithet in Need of Limiting Principles,’ 58 Antitrust Law Journal
841–53.
Baker, Jonathan B. 1999. ‘Promoting Innovation Competition through the Aspen/Kodak Rule,’ 7 George Mason
Band, Jonathan. 2017. Interfaces on Trial 3.0: Oracle America v. Google and Beyond. Accessed March 21, 2019
at https://papers.ssrn.com/sol3/papers.cfm?abstract_id=2876853.
Band, Jonathan, and Masanobu Katoh. 2011. Interfaces on Trial 2.0. Cambridge, MA: The MIT Press.
Barnett, Jonathan M. 2014. ‘From Patent Thickets to Patent Networks: The Legal Infrastructure of Digital
Economy,’ 55 Jurimetrics 1–53.
Barnett, Jonathan M. 2015. ‘The Anti-Commons Revisited,’ 29 Harvard Journal of Law & Technology 127–203.
Barnett, Jonathan M. 2017. ‘Has the Academy Led Patent Law Astray?,’ 32 Berkeley Technology Law Journal
1313–80.
Baseman, Kenneth C., Frederick R. Warren-Boulton, and Glenn A. Woroch. 1995. ‘Microsoft Plays Hardball:
The Use of Exclusionary Pricing and Technical Incompatibility to Maintain Monopoly Power in Markets for
Operating System Software,’ 40 Antitrust Bulletin 265–315.
Benkler, Yochai. 2002. ‘Coase’s Penguin, or, Linux and The Nature of the Firm,’ 112 The Yale Law Journal
369–446.
Benkler, Yochai. 2004. ‘Sharing Nicely: On Shareable Goods and the Emergence of Sharing as a Modality of
Economic Production,’ 114 The Yale Law Journal 273–358.

Berners-Lee, Tim. 2000. Weaving the Web: The Original Design and Ultimate Destiny of the World Wide Web.
New York, NY: HarperCollins Publishers Inc.
Besen, Stanley M., and Joseph Farrell. 1994. ‘Choosing How to Compete: Strategies and Tactics in
Standardization,’ 8 The Journal of Economic Perspectives 117–31.
Bessen, James. 2004. ‘Patent Thickets: Strategic Patenting of Complex Technologies,’ Research on Innovation
Working Paper Series, No. 0401.
Black Duck Opensource KnowledgeBase. 2017. ‘Top Open Source Licenses,’ accessed March 21, 2019 at https://
www.blackducksoftware.com/top-open-source-licenses.
Bohannan, Christina. 2011. ‘IP Misuse as Foreclosure,’ 96 Iowa Law Review 475–528.
Borwein, Jonathan, and David H. Bailey. 2015. ‘Moore’s Law is 50 Years Old but Will it Continue?,’ The
Conversation, accessed March 21, 2019 at http://theconversation.com/moores-law-is-50-years-old-but-will-
it-continue-44511.
Bresnahan, Timothy F. 1985. ‘Post-Entry Competition in the Plain Paper Copier Market,’ 75 The American
Briggs, John DeQ. 2009. ‘Intellectual Property and Antitrust: Two Scorpions in a Bottle,’ 10 The Sedona
Conference Journal 65–93.
Briscoe, Bob, Andrew Odlyzko, and Benjamin Tilly. 2006. ‘Metcalfe’s Law Is Wrong: Communications
Networks Increase in Value as They Add Members—But by How Much?,’ 43 IEEE Spectrum 34–9.
Bruce, Robert V. 1990. Bell: Alexander Graham Bell and the Conquest of Solitude. Ithaca, NY: Cornell University
Press.
Buccafusco, Christopher, and Mark A. Lemley. 2016. ‘Functionality Screens,’ 103 Virginia Law Review
1293–377.
Campbell-Kelly, Martin, Daniel Garcia-Swartz, Richard Lam, and Yilei Yang. 2015. ‘Economic and Business
Perspectives on Smartphones as Multi-Sided Platforms,’ 39 Telecommunications Policy 717–34.
Carrier, Michael A. 2014. ‘What You Need to Know about Standard Essential Patents,’ 8 CPI Antitrust Chronicle
1–10.
Carver, Brian W. 2005. ‘Share and Share Alike: Understanding and Enforcing Open Source and Free Software
Licenses,’ 20 Berkeley Technology Law Journal 443–81.
Ceruzzi, Paul E. 2003. A History of Modern Computing. Cambridge, MA: The MIT Press. 2nd ed.
Chien, Colleen V., and Mark A. Lemley. 2012. ‘Patent Holdup, the ITC, and the Public Interest,’ 98 Cornell
Law Review 1–46.
Chisum, Donald S., Rochelle Cooper Dreyfuss, Paul Goldstein, Robert A. Gorman, Dennis S. Karjala, Edmund
W. Kitch, Peter S. Menell, Leo J. Raskind, Jerome H. Reichman, and Pamela Samuelson. 1989. ‘LaST Frontier
Conference Report on Copyright Protection of Computer Software,’ 30 Jurimetrics Journal 15–33.
Clapes, Anthony L. 1994. ‘Confessions of an Amicus Curiae: Technophobia, Law and Creativity in the Digital
Arts,’ 19 University of Dayton Law Review 903–74.
Clapes, Anthony L., Patrick Lynch, and Mark R. Steinberg. 1987. ‘Silicon Epics and Binary Bards: Determining
the Proper Scope of Copyright Protection for Computer Programs,’ 34 UCLA Law Review 1493–546.
Contreras, Jorge L., and Richard J. Gilbert. 2015. ‘A Unified Framework for RAND and Other Reasonable
Royalties,’ 30 Berkeley Technology Law Journal 1451–504.
Determann, Lothar. 2006. ‘Dangerous Liaisons: Software Combinations as Derivative Works? Distribution,
Installation, and Execution of Linked Programs under Copyright Law, Commercial Licenses, and the GPL,’
21 Berkeley Technology Law Journal 1421–98.
Dreyfuss, Rochelle C. 1989. ‘The Federal Circuit: A Case Study in Specialized Courts,’ 64 New York University
Law Review 1–77.
Du Mont, Jason J., and Mark D. Janis. 2012. ‘Functionality in Design Protection Systems,’ 19 Journal of
Intellectual Property Law 261–303.
Dybvig, Philip H., and Chester S. Spatt. 1983. ‘Adoption Externalities as Public Goods,’ 20 Journal of Public
Economics 231–47.
Elazari Bar-On, Amit. 2019. ‘Unconscionability 2.0 and the IP Boilerplate: A Revised Doctrine of
Unconscionability for the Information Age.’ Forthcoming 2019.
Elhauge, Einer. 2008. ‘Do Patent Holdup and Royalty Stacking Lead to Systematically Excessive Royalties?,’ 4
Journal of Competition Law and Economics 535–70.
Englund, John. 1990. ‘Idea, Process, or Protected Expression? Determining the Scope of Copyright Protection
of the Structure of Computer Programs,’ 88 Michigan Law Review 866–909.
Evans, David S. 2003. ‘The Antitrust Economics of Multi-Sided Platform Markets,’ 20 Yale Journal on
Regulation 325–81.

Farrell, Joseph, and Paul Klemperer. 2007. ‘Coordination and Lock-In: Competition with Switching Costs and
Network Effects,’ in Mark Armstrong and Robert H. Porter, eds., Handbook of Industrial Organization, vol.
3. Amsterdam: Elsevier B.V.
Farrell, Joseph, and Garth Saloner. 1985. ‘Standardization, Compatibility, and Innovation,’ 16 The RAND
Farrell, Joseph, and Garth Saloner. 1986. ‘Installed Base and Compatibility: Innovation, Product Preannounce
ments, and Predation,’ 76 The American Law and Economics Review 940–55.
Farrell, Joseph, and Garth Saloner. 1998. ‘Coordination Through Committees and Markets,’ 19 The RAND
Farrell, Joseph, and Timothy Simcoe. 2012. ‘Choosing the Rules for Consensus Standardization,’ 43 The RAND
Feldman, Robin C. 2003. ‘The Insufficiency of Antitrust Analysis for Patent Misuse,’ 55 Hastings Law Journal
399–449.
Fellmeth, Aaron X. 1998. ‘Copyright Misuse and the Limits of the Intellectual Property Monopoly,’ 6 Journal
of Intellectual Property Law 1–40.
Final Report of the National Commission on New Technology Uses of Copyrighted Works. CONTU. 1979.
Accessed March 21, 2019 at http://digital-law-online.info/CONTU/PDF/index.html.
Floren, Paul. 1997. ‘Sun’s Java: Can It Burn Microsoft?,’ The New York Times (Jan. 20, 1997), accessed March
21, 2019 at http://www.nytimes.com/1997/01/20/business/worldbusiness/20iht-java.t.html.
Galetovic, Alexander, and Kirti Gupta. 2016. ‘Royalty Stacking and Standard Essential Patents: Theory and
Evidence from the World Mobile Wireless Industry,’ Hoover Institute IP2 Working Paper Series, No. 15012,
accessed March 21, 2019 at https://hooverip2.org/working-paper/wp15012/.
Gallini, N.T. 1992. ‘Patent Policy and Costly Imitation,’ 44 Journal of Industrial Economics 52–63.
Gandal, Neil. 1994. ‘Hedonic Price Indexes for Spreadsheets and an Empirical Test for Network Externalities,’
25 The RAND Journal of Economics 160–70.
Gandal, Neil, Nataly Gantman, and David Genesove. 2007. ‘Intellectual Property and Standardization
Committee Participation in the US Modem Industry,’ in Shane Greenstein and Victor Stango, eds., Standards
and Public Policy, vol. 1. Cambridge: Cambridge University Press.
Gates, Bill. 1976. ‘An Open Letter to Hobbyists,’ 2 Homebrew Computer Club Newsletter 2.
Geradin, Damien, and Miguel Rato. 2007. ‘Can Standard-Setting Lead to Exploitative Abuse? A Dissonant
View on Patent Hold-Up, Royalty-Stacking and the Meaning of FRAND,’ 3 European Competition Journal
101–61.
Geradin, Damien, Anne Layne-Farrar, and A. Jorge Padilla. 2008. ‘The Complements Problem within Standard
Setting: Assessing the Evidence of Royalty Stacking,’ 14 Boston University Journal of Science & Technology
Law 144–76.
Gergen, Mark P., John M. Golden, and Henry E. Smith. 2012. ‘The Supreme Court’s Accidental Revolution?
The Test for Permanent Injunctions,’ 112 Columbia Law Review 203–49.
Law Review 3–49.
Gilbert, Richard J. 2010. ‘Ties that Bind: Policies to Promote (Good) Patent Pools,’ 77 Antitrust Law Journal
1–47.
Gilbert, Richard J., and Carl Shapiro. 1997. Antitrust Issues in the Licensing of Intellectual Property: The Nine
No-No’s Meet the Nineties. Brookings Papers: Micro.
GNU General Public License 3.0. 2007. Accessed March 21, 2019 at https://www.gnu.org/licenses/gpl-3.0.en.
html.
Golden, John M. 2007. ‘“Patent Trolls” and Patent Remedies,’ 85 Texas Law Review 2111–61.
Goldstein, Ayala. 2018. ‘Top 10 Open Source Licenses in 2018: Trends and Predictions,’ White Source (Dec. 3,
2018). Accessed May 4, 2019 at https://resources.whitesourcesoftware.com/blog-whitesource/top-open-source
-licenses-trends-and-predictions.
Goldstein, Paul. 2000. International Copyright: Principles, Law, and Practice. New York: Oxford University
Press.
Grad, Burton. 2002. ‘A Personal Recollection: IBM’s Unbundling of Software and Services,’ 24 IEEE Annals
of the History of Computing 64–71.
Law & Technology Patent Damages Workshop,’ 25 Texas Intellectual Property Law Journal 113–40.
Hall, Bronwyn H., Christian Helmers, and Georg von Graevenitz. 2016. ‘Technology Entry in the Presence of
Patent Thickets,’ Institute for Fiscal Studies Working Paper Series, No. 16-02.
Herrell, Jonas P. 2011. ‘The Copyright Misuse Doctrine’s Role in Open and Closed Technology Platforms,’ 26
Berkeley Technology Law Journal 441–89.
Hovenkamp, Herbert. 2012. ‘Antitrust and the Movement of Technology,’ 19 George Mason Law Review
1119–145.

Hovenkamp, Herbert. 2015. ‘Antitrust and the Patent System: A Reexamination,’ 76 Ohio State Law Journal
467–564.
Hovenkamp, Herbert. 2017. ‘The Rule of Reason,’ 1778 Faculty Scholarship at the University of Pennsylvania Law
1–64.
Kaplan, Jerry. 1996. Startup: A Silicon Valley Adventure. London: Penguin Books.
Karjala, Dennis S. 1987. ‘Copyright, Computer Software, and the New Protectionism,’ 28 Jurimetrics 33–96.
Katz, Michael, and Carl Shapiro. 1985. ‘Network Externalities, Competition, and Compatibility,’ 75 American
Katz, Michael, and Carl Shapiro. 1986. ‘Technology Adoption in the Presence of Network Externalities,’ 94
Journal of Political Economy 822–41.
Katz, Michael, and Carl Shapiro. 1994. ‘Systems Competition and Network Effects,’ 8 Journal of Economic
Perspectives 93–115.
Landes, William M., and Richard A. Posner. 2004. ‘An Empirical Analysis of the Patent Court,’ 71 University
Langlois, Richard N. 2001. ‘Technological Standards, Innovation, and Essential Facilities: Toward a Schumpeterian
Post-Chicago Approach,’ in Jerry Ellig, ed., Dynamic Competition and Public Policy: Technology, Innovation,
and Antitrust Issues. Cambridge: Cambridge University Press.
Lee, William F., and Douglas A. Melamed. 2016. ‘Breaking the Vicious Cycle of Patent Damages,’ 101 Cornell
Lemley, Mark A. 1990. ‘The Economic Irrationality of the Patent Misuse Doctrine,’ 78 Stanford Law Review
1599–632.
Lemley, Mark A. 2002. ‘Intellectual Property Rights and Standard-Setting Organizations,’ 90 California Law
Review 1889–973.
Lemley, Mark A. 2013. ‘Software Patents and the Return of Functional Claiming,’ 2013 Wisconsin Law Review
905–64.
Lemley, Mark A., and David McGowan. 1998. ‘Could Java Change Everything? The Competitive Propriety of
a Proprietary Standard,’ 43 Antitrust Bulletin 715–67.
Lemley, Mark A., and Carl Shapiro. 2007a. ‘Patent Holdup and Royalty Stacking,’ 85 Texas Law Review
1991–2049.
Lemley, Mark A., and Carl Shapiro. 2007b. ‘Reply: Patent Holdup and Royalty Stacking,’ 85 Texas Law Review
2163–5.
Lemley, Mark A., Peter S. Menell, Robert P. Merges, Pamela Samuelson, and Brian W. Carver. 2011. Software
and Internet Law, vol. 1. New York, NY: Aspen Publishers. 4th ed.
Lerner, Josh, and Jean Tirole. 2004. ‘Efficient Patent Pools,’ 94 The American Economic Review 691–711.
Lerner, Josh, and Jean Tirole. 2005. ‘The Economics of Technology Sharing: Open Source and Beyond,’ 19
Lichtman, Douglas. 2000. ‘Property Rights in Emerging Platform Technologies,’ 29 The Journal of Legal Studies
615–48.
Liebowitz, Stan J., and Stephen Margolis. 1994. ‘Network Externality: An Uncommon Tragedy,’ 8 Journal of
Economic Perspectives 133–50.
Malamud, Carl. 2014. Testimony of Carl Malamud, Public. Resource.Org, Hearing on the Scope of Copyright
Protection, House Judiciary Committee, U.S. House of Representatives (Jan. 14, 2014), accessed March 21,
2019 http://www.itworld.com/article/2738675/mobile/the-impact-of-oracle-s-defense-of-api-copyrights.html
at https://public.resource.org/edicts/#p60.
Markoff, John. 1997. ‘Sun Sues Microsoft in Dispute over Java,’ The New York Times (Oct. 8, 1997).
Mattioli, Michael. 2019. ‘Empirical Studies of Patent Pools,’ in Peter S. Menell and David L. Schwartz, eds.,
Research Handbook on the Economics of Intellectual Property, vol. II. Cheltenham, UK and Northampton,
McGowan, David. 2001. ‘Innovation, Uncertainty, and Stability in Antitrust Law,’ 16 Berkeley Technology Law
Journal 729–811.
McKusick, Marshall Kirk. 1999. ‘Twenty Years of Berkeley Unix: From AT&T Owned to Freely Redistributable,’
in Chris DiBona, Sam Ockman, and Mark Stone, eds., Open Sources: Voices from the Open Source Revolution.
Sebastopol, CA: O’Reilly Media, Inc.
Meeker, Heather. 2015. Open (Source) for Business: A Practical Guide to Open Source Software Licensing. North
Charleston, SC: CreateSpace Independent Publishing Platform.
Menell, Peter S. 1987. ‘Tailoring Legal Protection for Computer Software,’ 39 Stanford Law Review
1329–72.
Menell, Peter S. 1989. ‘An Analysis of the Scope of Copyright Protection for Application Programs,’ 41 Stanford

Menell, Peter S. 1998. ‘An Epitaph for Traditional Copyright Protection of Network Features of Computer
Software,’ 43 Antitrust Bulletin 651–714.
Menell, Peter S. 2007. ‘The Property Rights Movement’s Embrace of Intellectual Property: True Love or
Menell, Peter S. 2011. ‘Governance of Intellectual Resources and Disintegration of Intellectual Property in the
Digital Age,’ 26 Berkeley Technology Law Journal 1523–60.
Menell, Peter S. 2013. ‘It’s Time to Make Vague Software Patents More Clear,’ Wired (Feb. 7, 2013),
accessed March 21, 2019 at http://www.wired.com/opinion/2013/02/its-time-to-make-vague-software-pate
nts-more-clear/.
Menell, Peter S. 2016. ‘API Copyrightability Bleak House: Unraveling and Repairing the Oracle v. Google
Jurisdictional Mess,’ 31 Berkeley Technology Law Journal 1515–95.
Menell, Peter S. 2018. ‘Rise of the API Copyright Dead? An Updated Epitaph for Copyright Protection of
Network and Functional Features of Computer Software,’ 31 Harvard Journal of Law & Technology 305–490.
Analysis 1–59.
Shavell, eds., Handbook of Law and Economics, vol. II. Amsterdam: North-Holland.
Menell, Peter S., Mark A. Lemley, and Robert P. Merges. 2018. Intellectual Property in the New Technological
Age, Volume I: Perspectives, Trade Secrets and Patents. Berkeley, CA: Clause 8 Publishing.
Merges, Robert. 1988. ‘Reflections on Current Legislation Affecting Patent Misuse,’ 70 Journal of the Patent
and Trademark Office Society 793–804.
Merges, Robert P. 2000. ‘One Hundred Years of Solicitude: Intellectual Property Law 1900–2000,’ 88 California
Merges, Robert P. 2001. ‘Institutions for Intellectual Property Transactions: The Case of Patent Pools,’ in
Rochelle Cooper Dreyfuss, Diane Leenheer Zimmerman, and Harry First, eds., Innovation Policy for the
Knowledge of Society, vol. 1, New York: Oxford University Press.
Merges, Robert P. 2004. ‘A New Dynamism in the Public Domain,’ 71 University of Chicago Law Review
183–203.
Mock, Dave. 2005. The Qualcomm Equation: How a Fledgling Telecom Company Forged a New Path to Big
Profits and Market Dominance. New York, NY: Amacom Division American Management Association.
Moore, Gordon E. 1965. ‘Cramming More Components onto Integrated Circuits,’ 38 Electronics Magazine
114–17.
Moskowitz, Nelson. 1982. ‘The Metamorphosis of Software-Related Invention Patentability,’ 3 Computer Law
Journal 273–336.
Moy, Russell. 2000. ‘A Case against Software Patents,’ 17 Santa Clara Computer and High Technology Law
Journal 67–99.
Mueller, Janice M. 2002. ‘Patent Misuse through the Capture of Industry Standards,’ 17 Berkeley Technology
National Research Council of the National Academies, Committee on Intellectual Property Management in
Standard-Setting Processes, Keith Maskus and Stephen A. Merrill, eds. 2013. Patent Challenges for Standard-
Setting in the Global Economy: Lessons from Information and Communications Technology. Washington, DC:
The National Academies Press.
Nordhaus, William D. 1969. Invention, Growth, and Welfare: A Theoretical Treatment of Technological Change.
Page, William H. 2014. ‘Judging Monopolistic Pricing: F/RAND and Antitrust Injury,’ 22 Texas Intellectual
Property Journal 181–208.
Parchomovsky, Gideon, and Polk R. Wagner. 2005. ‘Patent Portfolios,’ 54 University of Pennsylvania Law Review
1–77.
Patterson, Mark R. 2012. ‘Leveraging Information about Patents: Settlements, Portfolios, and Holdups,’ 50
Houston Law Review 483–522.
Perzanowski, Aaron, and Jason Schultz. 2016. The End of Ownership: Personal Property in the Digital Economy.
Phelps, Marshall, and David Kline. 2010. Burning the Ships: Transforming Your Company’s Culture through
Intellectual Property Strategy. Hoboken, NJ: John Wiley & Sons.
Phillips, Douglas E. 2009. The Software License Unveiled: How Legislation by License Controls Software Access.
New York, NY: Oxford University Press, Inc.
Pitofsky, Robert. 2001. ‘Challenges of the New Economy: Issues at the Intersection of Antitrust and Intellectual
Property,’ 68 Antitrust Law Journal 913–24.
Profitt, Brian. 2011. ‘The Impact of Oracle’s Defense of API Copyrights,’ ITWorld (Aug. 23, 2011), accessed
March 21, 2019 at http://www.itworld.com/article/2738675/mobile/the-impact-of-oracle-s-defense-of-api-
copyrights.html.

Pruitt, Scarlet, and Paul Roberts. 2004. ‘Microsoft to Pay $700 Million for Antitrust Issues, $900 Million to
Resolve Patent Dispute,’ InfoWorld (Apr. 2, 2004), accessed March 21, 2019 at http://www.infoworld.com/
article/2667124/operating-systems/update--sun--microsoft-settle-suit-in-billion-dollar-pact.html.
Puffert, Douglas J. 2000. ‘The Standardization of Track Gauge on North American Railways, 1830–1890,’ 60
The Journal of Economic History 933–60.
Ratner, James R. 1988. ‘Should There Be an Essential Facility Doctrine?,’ 21 UC Davis Law Review 327–82.
Rice, James, 2015. ‘The Defensive Patent Playbook,’ 30 Berkeley Technology Law Journal 725–75.
Rivette, Kevin G., and David Kline. 1999. Rembrandts in the Attic: Unlocking the Hidden Value of Patents.
Boston, MA: Harvard University Press.
Royall, Sean, Amanda Tessar, and Adam D. Vincenzo. 2009. ‘Deterring “Patent Ambush” in Standard Setting:
Lessons from Rambus and Qualcomm,’ 24 Antitrust American Bar Association 34–7.
Rubinfeld, Daniel L., and Robert Maness. 2005. ‘The Strategic Use of Patents: Implications for Antitrust,’ in
François Lévêque and Howard A. Shelanski, eds., Antitrust, Patents, and Copyright: EU and US Perspectives,
vol. I. Oxford: Hart Publishing.
Russell, Andrew L. 2006. ‘Industrial Legislatures: The American System of Standardization,’ in International
Standardization as a Strategic Tool. Geneva, Switzerland: IEC.
Sammi, P. Anthony, Christopher A. Lisy, and Andrew Gish. 2013. ‘Good Clean Fun: Using Clean Room
Procedures in Intellectual Property Litigation,’ 25 Intellectual Property & Technology Law Journal 3–14.
Samuelson, Pamela, Randall Davis, Mitchell D. Kapor, and J.H. Reichman. 1994. ‘A Manifesto Concerning the
Legal Protection of Computer Programs,’ 94 Columbia Law Review 2308–431.
Samuelson, Pamela, and Suzanne Scotchmer. 2002. ‘The Law and Economics of Reverse Engineering,’ 111 The
Saunders, Kurt M. 2002. ‘Patent Nonuse and the Role of Public Interest as a Deterrent to Technology
Suppression,’ 15 Harvard Journal of Law & Technology 389–452.
Scherer, F.M. 1987. ‘Antitrust, Efficiency, and Progress,’ 62 New York University Law Review 998–1019.
& Technology 1–67.
Seaman, Christopher B. 2016. ‘Permanent Injunctions in Patent Litigation after eBay: An Empirical Study,’ 101
Iowa Law Review 1949–2018.
Shankland, Stephen. 2002a. ‘Sun, Microsoft Settle Java Suit,’ CNET (Mar. 15, 2002), accessed March 21, 2019
at https://www.cnet.com/news/sun-microsoft-settle-java-suit/.
Shankland, Stephen. 2002b. ‘Sun Brings Antitrust Suit against Microsoft: The Company Files a Private
Antitrust Suit against Microsoft Seeking Damages that could Top $1 Billion,’ CNET (July 20, 2002), accessed
March 21, 2019 at https://www.cnet.com/news/sun-brings-antitrust-suit-against-microsoft-1/.
Shannon, Jim. 2012. ‘Statement of Jim Shannon on behalf of the National Fire Protection Association’ (1 June
2012), submitted to the Office of the Federal Register, Request for Comments, Federal Register, Vol. 77, No.
38, Feb. 27, 2012, NARA 12-0002, accessed March 21, 2019 at https://law.resource.org/pub/us/cfr/regulations.
gov.docket.01/0900006481025751.pdf.
Shapiro, Carl. 2001. ‘Navigating the Patent Thicket: Cross Licenses, Patent Pools, and Standard-Setting,’ in
Adam B. Jaffe, Josh Lerner, and Scott Stern, eds., Innovation Policy and the Economy, vol. 1. Cambridge,
MA: MIT Press.
Shapiro, Carl, and Hal R. Varian. 1999. Information Rules: A Strategic Guide to the Network Economy. Boston,
MA: Harvard Business School Press.
Sidak, J. Gregory. 2008. ‘Holdup, Royalty Stacking, and the Presumption of Injunctive Relief for Patent
Infringement: A Reply to Lemley and Shapiro,’ 92 Minnesota Law Review 714–48.
Sobel, Gerald. 1984. ‘The Antitrust Interface with Patents and Innovation: Acquisition of Patents, Improvement
Patents and Grant-Backs, Non-Use, Fraud on the Patent Office, Development of New Products and Joint
Research,’ 53 Antitrust Law Journal 681–711.
Tugend, Alina. 2016. ‘For Online Dating Sites, a Bumpy Road to Love,’ The New York Times (Dec. 25,
2016), accessed March 21, 2019 at https://www.nytimes.com/2016/12/24/business/online-dating-sites-jdate-
christianmingle.html.
Van Houweling, Molly Shaffer. 2019. ‘Intellectual Property as Property,’ in Ben Depoorter and Peter S.
Menell, eds., Research Handbook on the Economics of Intellectual Property, vol. I. Cheltenham, UK and
Weber, Steven. 2004. The Success of Open Source. Cambridge, MA: Harvard University Press.
West, Joel. 2007. ‘The Economic Realities of Open Standards: Black, White, and Many Shades of Gray,’ in Shane
Greenstein and Victor Stango, eds., Standards and Public Policy, Cambridge: Cambridge University Press.
Williamson, Oliver. 1985. The Economic Institutions of Capitalism. New York, NY: The Free Press.
Wilson, Bruce B. 1970. Deputy Assistant Attorney Gen., Remarks before the Fourth New England Antitrust

Conference: Patent and Know-How License Agreements: Field of Use, Territorial, Price and Quantity
Restrictions (Nov. 6, 1970).
Ziedonis, Rosemarie Ham. 2004. ‘Don’t Fence Me In: Fragmented Markets for Technology and the Patent
Acquisition Strategies of Firms,’ 50 Management Science 804–20.
Cases and Investigations
321 Studios v. Metro-Goldwyn-Mayer Studios, Inc., 307 F. Supp. 2d 1085 (N.D. Cal. 2004).
Adams v. Burke, 84 U.S. 453 (1873).
Alice Corp. v. CLS Bank International, 573 U.S. 208 (2014).
American Society for Testing and Materials, et al. v. Public.Resource.Org, Inc., 896 F.3d 437 (D.C. Cir. 2018).
Amini Innovation Corp. v. Anthony Cal., Inc., 439 F.3d 1365 (Fed. Cir. 2006).
Apple Computer, Inc. v. Formula Int’l, Inc., 562 F. Supp. 775 (C.D. Cal. 1983), aff’d, 725 F.2d 521 (9th Cir. 1984).
Apple Computer, Inc. v. Franklin Computer Corp., 545 F. Supp. 812 (E.D. Pa. 1982), rev’d, 714 F.2d 1240 (3rd
Cir. 1983).
Apple Computer, Inc. v. Franklin Computer Corp., 714 F.2d 1253 (3rd Cir. 1983).
Apple Computer, Inc. v. Microsoft Corp., 35 F.3d 1435 (9th Cir. 1994).
Apple Computer, Inc. v. Microsoft Corp., 799 F. Supp. 1006 (N.D. Cal. 1992), aff’d in part, rev’d in part, 35
F.2d 1435 (9th Cir. 1994).
Apple Inc. v. Motorola, Inc., 757 F.3d 1286 (Fed. Cir. 2014).
Apple Inc. v. Samsung Elecs. Co., 786 F.3d 983 (Fed. Cir. 2015).
Application of Bergel, 292 F.2d 955 (C.C.P.A. 1961).
Aspen Skiing Co. v. Aspen Highlands Skiing Corp., 472 U.S. 585 (1985).
Assessment Techs., LLC v. WIREdata, Inc., 350 F.3d 640 (7th Cir. 2003).
Baker v. Selden, 101 U.S. 99 (1879).
Bateman v. Mnemonics, Inc., 79 F.3d 1532 (11th Cir. 1996).
Berry Sterling Corp. v. Pescor Plastics, 122 F.3d 1452 (Fed. Cir. 1997).
Best Lock Corp. v. Ilco Unican Corp., 94 F.3d 1563 (Fed. Cir. 1996).
Bilski v. Kappos, 561 U.S. 593 (2010).
Bowers v. Baystate Techs., 320 F.3d 1317 (Fed. Cir. 2003).
Building Officials & Code Admin. v. Code Technology, Inc., 628 F.2d 730 (1st Cir. 1980).
Carbice Corp. v. American Patents Development Corp., 283 U.S. 27 (1931).
Chamberlain Group, Inc. v. Skylink Techs., Inc., 381 F.3d 1178 (Fed. Cir. 2004).
Cincinnati Car Co. v. New York Rapid Transit Corp., 66 F.2d 592 (2d Cir. 1933).
Code Revision Comm’n for General Assembly of Georgia v. Public.Resource.Org, Inc., 906 F.3d 1229 (11th
Cir. 2018).
Computer Associates International v. Altai, Inc., 982 F.2d 693 (2d Cir. 1992).
Cornell Univ. v. Hewlett-Packard Co., 609 F. Supp. 2d 279 (N.D.N.Y. 2009).
Creative Labs, Inc. v. Cyrix Corp., 42 U.S.P.Q. 2d 1872 (N.D. Cal. 1997).
Crocs, Inc. v. Int’l Trade Comm’n, 598 F.3d 1294 (Fed. Cir. 2010).
CSIRO v. Cisco Sys., Inc., 809 F.3d 1295 (Fed. Cir. 2015).
Diamond v. Diehr, 450 U.S. 175 (1981).
Dolbear v. Am. Bell Tel. Co., 126 U.S. 1 (1888).
DSC Communications Corp. v. DGI Technologies, Inc., 681 F.3d 597 (5th Cir. 1996).
Eastman Kodak Co. v. Image Technical Services, Inc., 504 U.S. 451 (1992).
eBay, Inc. v. MercExchange, LLC, 547 U.S. 388 (2006).
Egyptian Goddess, Inc. v. Swisa, Inc., 543 F.3d 665 (Fed. Cir. 2008).
Elliott v. Google, Inc., 860 F.3d 1151 (9th Cir. 2017).
Eng’g Dynamics, Inc. v. Structural Software, Inc., 26 F.3d 1335 (5th Cir. 1994).
Ericsson, Inc. v. D-Link Sys., Inc., 773 F.3d 1201 (Fed. Cir. 2014).
Gates Rubber Co. v. Bando Chem. Indus., Ltd., 9 F.3d 823 (10th Cir. 1993).
Georgia-Pacific Corp. v. U.S. Plywood Corp., 318 F. Supp. 1116 (S.D.N.Y. 1970).
Gottschalk v. Benson, 409 U.S. 63 (1972).
Halo Electronics, Inc. v. Pulse Electronics, Inc., 136 S. Ct. 1923 (2016).
Harper House, Inc. v. Thomas Nelson, Inc., 889 F.2d 197 (9th Cir. 1989).
Hewlett-Packard Co. v. Nu-Kote Intern., Inc., 2000 WL 33992123 (N.D. Cal. 2000).
Image Tech. Servs., Inc. v. Eastman Kodak Co., 125 F.3d 1195 (9th Cir. 1997).
Impression Prods. v. Lexmark Int’l, Inc., 137 S. Ct. 1523 (2017).
In re Alappat, 33 F.3d 1526 (Fed. Cir. 1994).

In re Bilski, 545 F.3d 943 (Fed. Cir. 2008) (en banc).

In re Comiskey, 554 F.3d 967 (Fed. Cir. 2009).
In re Dembiczak, 175 F.3d 994 (Fed. Cir. 1999).
In re Independent Service Organizations Antitrust Litigation, 203 F.3d 1322 (Fed. Cir. 2000).
In re Innovatio IP Ventures LLC Patent Litig., No. 11 C 9308, 2013 U.S. Dist. LEXIS 144061, 2013 WL 5593609
(N.D. Ill. Oct. 3, 2013).
In re Intel Corp., No. 9288, 1999 F.T.C. LEXIS 145 (Aug. 3, 1999).
In re Morton-Norwich Prods., Inc., 671 F.2d 1332 (C.C.P.A. 1982).
In re Motorola Mobility LLC, F.T.C. File No. C-4410 (2013).
In re Negotiated Data Solutions LLC (N-Data), No. C-4234, 2008 WL 4407246 (F.T.C.) (2008).
In re Nuijten, 500 F.3d 1346 (Fed. Cir. 2007).
Intel v. Advanced Micro Devices, 765 F. Supp. 1292 (N.D. Cal. 1991).
Intergraph Corp. v. Intel Corp., 195 F.3d 1346 (Fed. Cir. 1999).
Inwood Labs., Inc. v. Ives Labs., Inc., 456 U.S. 844 (1982).
Jacobsen v. Katzer, 535 F.3d 1373 (Fed. Cir. 2008).
Kellogg Co. v. National Biscuit Co., 305 U.S. 111 (1938).
King-Seely Thermos Co. v. Aladdin Indus., Inc., 321 F.2d 577 (2d Cir. 1963).
KSR Int’l Co. v. Teleflex Inc., 550 U.S. 398 (2007).
L.A. Gear, Inc. v. Thom McAn Shoe Co., 988 F.2d 1117 (Fed. Cir. 1993).
Lasercomb America, Inc. v. Reynolds, 911 F.2d 970 (4th Cir. 1990).
LaserDynamics Inc. v. Quanta Computer, Inc., 694 F.3d 51 (Fed. Cir. 2012).
Lee v. Dayton-Hudson Corp., 838 F.2d 1186 (Fed. Cir. 1988).
Lexmark Int’l, Inc. v. Static Control Components, Inc., 387 F.3d 522 (6th Cir. 2005).
Lotus Dev. Corp. v. Borland Int’l, Inc., 516 U.S. 233 (1996).
Lotus Dev. Corp. v. Borland Int’l, Inc., 49 F.3d 807 (1st Cir. 1995), aff’d equally divided Court, 516 U.S. 233
(1996).
Lucent Techs., Inc. v. Gateway, Inc., 580 F.3d 1301 (Fed. Cir. 2009).
M. Witmark & Sons v. Jensen, 80 F. Supp. 843 (D. Minn. 1948), appeal dismissed sub nom., M. Witmark & Sons
v. Berger Amusement Co., 177 F.2d 515 (8th Cir. 1949).
Matter of Dell Computer Corp., 121 F.T.C. 616, 1996 WL 33412055 (May 20, 1996).
Mayo Collaborative Services v. Prometheus Laboratories, Inc., 566 U.S. 66 (2012).
MDY Indus., LLC v. Blizzard Entm’t, Inc., 629 F.3d 928 (9th Cir. 2010).
Microsoft Corp. v. Lindows.com, Inc., 2002 WL 32153471 (W.D. Wash. 2002).
Microsoft Corp. v. Motorola, Inc., No. C10-1823JLR, 2013 U.S. Dist. LEXIS 60233, 2013 WL 2111217 (W.D.
Wash. Apr. 25, 2013).
Microsoft Corp. v. Motorola, Inc., 795 F.3d 1024 (9th Cir. 2015).
MiTek Holdings, Inc. v. ARCE Engineering Co., 89 F.3d 1548 (11th Cir. 1996).
Mitel, Inc. v. Iqtel, Inc., 124 F.3d 1366 (10th Cir. 1997).
Morton Salt Co. v. G.S. Suppiger Co., 314 U.S. 488 (1942).
Motion Picture Patents Co. v. Universal Film Mfg. Co., 243 U.S. 502 (1917).
Nautilus, Inc. v. Biosig Instruments, Inc., 134 S. Ct. 2120 (2014).
New York v. Microsoft Corp., No. 98-1233 (D.D.C. May 18, 1998).
Nichols v. Universal Pictures Corp., 45 F.2d 119 (2d Cir. 1930).
OddzOn Prods., Inc. v. Just Toys, Inc., 122 F.3d 1396 (Fed. Cir. 1997).
Oracle America, Inc. v. Google, Inc., 872 F. Supp. 2d 974 (N.D. Cal. 2012).
Oracle America, Inc. v. Google, Inc., 750 F.3d 1339 (Fed. Cir. 2014).
Oracle America, Inc. v. Google LLC, 886 F.3d 1179 (Fed. Cir. 2018).
Oracle America, Inc. v. Google, Inc., 750 F.3d 1339 (Fed. Cir. 2014) (Opening Brief and Addendum of Plaintiff-
Appellant, Oracle America, Inc. v. Google, Inc., 2013, pp. 12–13).
Otter Tail Power Co. v. United States, 410 U.S. 366 (1973).
Parker v. Flook, 437 U.S. 584 (1978).
PHG Techs., LLC v. St. John Cos., 469 F.3d 1361 (Fed. Cir. 2006).
Plains Cotton Coop. Ass’n v. Goodpasture Computer Serv., Inc., 807 F.2d 1256 (5th Cir. 1987).
Practice Management Information Corporation v. American Medical Assoc., 121 F.3d 516, as amended, 133
F.3d 1140 (9th Cir. 1997).
Princeton Graphics Operating, L.P. v. NEC Home Electronics (U.S.A.), Inc., 732 F. Supp. 1258 (S.D.N.Y.
1990).
Qualcomm Inc. v. Broadcom Corp., 548 F.3d 1004 (Fed. Cir. 2008).
Rambus Inc. v. Infineon Technologies AG, 318 F.3d 1081 (Fed. Cir. 2003).
Richardson v. Stanley Works, Inc., 597 F.3d 1288 (Fed. Cir. 2010).
Rosco, Inc. v. Mirror Lite Co., 304 F.3d 1373 (Fed. Cir. 2002).

Samsung Elecs. v. Apple Inc., 137 S. Ct. 429 (2016).

SCM Corp. v. Xerox Corp., 645 F.2d 1195 (2d Cir. 1981).
Sega Enterprises, Ltd. v. Accolade, Inc., 977 F.2d 1510 (9th Cir. 1992).
Seiko Epson Corp. v. Nu-Kote Int’l, Inc., 190 F.3d 1360 (Fed. Cir. 1999).
Sony Computer Entertainment, Inc. v. Connectix Corp., 203 F.3d 596 (9th Cir. 2000).
Stac Elec. v. Microsoft Corp., No. 93-0413 (S.D. Cal. decided Feb. 23, 1994), appeal dismissed per stipulation,
38 F.3d 1222 (Fed. Cir. 1994).
State Street Bank and Trust Company v. Signature Financial Group, Inc., 149 F.3d 1368 (Fed. Cir. 1998).
Storage Technology Corp. v. Custom Hardware Eng’g & Consulting, Inc., 421 F.3d 1307 (Fed. Cir. 2005).
Sun Microsystems, Inc. v. Microsoft Corp., 46 U.S.P.Q. 2d 1531 (N.D. Cal. Mar. 24, 1998).
Sun Microsystems, Inc. v. Microsoft Corp., 87 F.Supp. 2d 992 (N.D. Cal. 2000).
Synercom Technology, Inc. v. University Computing Co., 462 F. Supp. 1003 (N.D. Tex. 1978).
Teleflex, Inc. v. KSR Intern. Co., 119 Fed.Appx’ 282 (Fed.Cir. 2005).
The Murphy Door Bed Co., Inc. v. Interior Sleep Systems, Inc., 874 F.2d 95 (2d Cir. 1989).
Uniloc USA, Inc. v. Microsoft Corp., 632 F.3d 1292 (Fed. Cir. 2011).
United States v. Microsoft Corp., 253 F.3d 34 (D.C. Cir. 2001).
United States v. Microsoft Corp., No. 98-1232 (D.D.C. May 18, 1998).
Universal City Studios, Inc. v. Corley, 273 F.3d 429 (2d Cir. 2001).
Vault Corp. v. Quaid Software, Ltd., 847 F.2d 255 (5th Cir. 1988).
Veeck v. S. Bldg. Code Congress Int’l, Inc., 293 F.3d 791 (5th Cir. 2002) (en banc).
VirnetX, Inc. v. Cisco Sys., Inc., 767 F.3d 1308 (Fed. Cir. 2014).
Whelan Associates, Inc. v. Jaslow Dental Laboratory, Inc., 797 F.2d 1222 (3d Cir. 1986).
Williamson v. Citrix Online LLC, 792 F.3d 1339 (Fed. Cir. 2015) (en banc).
17 U.S.C. § 101, § 102(b), § 105, §109(a), § 117, § 1201

35 U.S.C. § 100(b), § 101, § 102, § 103, § 112, § 171, § 271(d)(1)–(3), § 271(d)(4), § 271(d)(5), § 284, § 289
Act to amend the patent and trademark laws, Pub. L. No. 96-517, 94 Stat. 3015 (1980).
America Invents Act, Pub. L. No. 112-29, 125 Stat. 284 (2011).
Clayton Antitrust Act, Pub. L. No. 63-212, 38 Stat. 730, codified at 15 U.S.C. §§ 12–27 (1914).
Copyright Act of 1976, Pub. L. 94-553, 90 Stat. 254, codified at 17 U.S.C. § 101 et seq.
Establishing the President’s Commission on the Patent System, Executive Order No. 11, 215, 30 Fed. Reg. 4661
(1965)).
Federal Trade Commission Act, Pub. L. No. 63-203, 38 Stat. 717 (1914).
H.R. Rep. No. 94-1476, 94th Cong., 2d Sess. (1976).
Institute of Electrical and Electronics Engineers (IEEE), 2015. IEEE-SA Standards Board Bylaws, accessed
March 21, 2019 at http://standards.ieee.org/develop/policies/bylaws/approved-changes.pdf.
Lanham (Trademark) Act (Pub. L. 79-489, 60 Stat. 427, codified at 15 U.S.C. § 1051 et seq.
Patent Act of 1952, Pub. L. 93-596, 66 Stat. 792, codified at 15 U.S.C. § 1 et seq.
Patent Misuse Reform Act of 1988, Pub. L. No. 100-703, 102 Stat. 4674 (H.R. 4972).
President’s Commission on the Patent System, To Promote the Progress of Useful Arts, Report to the Senate
Judiciary Committee, S. Doc. No. 5, 90th Cong., 1st sess. (1967).
S. Rep. No. 105-190 (1998).
Sherman Antitrust Act, Pub. L. No. 103-325, 26 Stat. 209, codified at 15 U.S.C. §§ 1–7 (1890).
The Digital Millennium Copyright Act of 1998, Pub. L. No. 105-304, 112 Stat. 2860 (1998) (codified in scattered
sections of 5, 17, 28, and 35 U.S.C.).
United States General Accountability Office. 2016. Intellectual Property: Patent Office Should Define Quality,
Reassess Incentives, and Improve Clarity, GAO-16-490.
Unlocking Consumer Choice and Wireless Competition Act, Pub. L. 113-144, 128 Stat. 1751 (2014).
Administrative and Other Materials
Comments of Law and Business Scholars Submitted to the U.S. Department of Justice and Federal Trade
Commission Regarding a Proposed Update to the Antitrust Guidelines for the Licensing of Intellectual
Property (2016).
Federal Trade Commission. 2003. To Promote Innovation: The Proper Balance of Competition and Patent Law
and Policy (Oct. 2003), accessed March 21, 2019 at https://www.ftc.gov/sites/default/files/documents/reports/
promote-innovation-proper-balance-competition-and-patent-law-and-policy/innovationrpt.pdf.

Competition (Mar. 2011), accessed March 21, 2019 at https://www.ftc.gov/reports/evolving-ip-marketplace
-aligning-patent-notice-remedies-competition.
Federal Trade Commission. 2012a. FTC Order Restores Competition in U.S. Market for Equipment Used to
Recharge Vehicle Air Conditioning Systems (Nov. 26, 2012), accessed March 21, 2019 at http://www.ftc.gov/
news-events/press-releases/2012/11/ftc-order-restores-competition-us-market-equipment-used-recharge.
Federal Trade Commission. 2012b. Third Party Statement on the Public Interest, In re Certain Gaming
and Entertainment Related Software, and Components Thereof, ITC Investigation No. 337-TA-752 (June 6,
2012), accessed March 21, 2019 at https://www.ftc.gov/sites/default/files/documents/advocacy_documents/ftc-
comment-united-states-international-trade-commission-concerning-certain-wireless-communication/1206ftcwi
relesscom.pdf.
Library of Congress. 2018. U.S. Copyright Office, Exemption to Prohibition on Circumvention of Copyright
Protection Systems for Access Control Technologies, 83 Federal Register 54010 (Oct. 26, 2018).
Restatement of Torts, § 742.
U.S. Dep’t. Justice. 1988. Antitrust Enforcement Guidelines for International Operations (Nov. 10, 1988).
U.S. Dep’t. Justice. 2006. Business Review Letter relating to VMEbus International Trade Association.
U.S. Dep’t. Justice. 2015. Business Review Letter relating to IEEE from Acting Assistant Attorney General Renata
B. Hesse to Michael A. Lindsay.
U.S. Dep’t. Justice and Federal Trade Commission. 1995. Antitrust Enforcement Guidelines for International
Operations.
U.S. Dep’t. Justice and Federal Trade Commission. 2007. ‘Letter from Assistant Attorney General, Thomas O.
Barnett, to Dorsey & Whitney Partner, Michael A. Lindsay’ (Apr. 30, 2007).
U.S. Dep’t. Justice and Federal Trade Commission. 2017. Intellectual Property Guidelines.
U.S. Dep’t. Justice and PTO. 2013. Policy Statement on Remedies for Standards-Essential Patents Subject to
Voluntary FRAND Commitments.

8. Intellectual property and competition
Herbert Hovenkamp* 5
Contents
I. Introduction: The Political Economy of Intellectual Property Law

II. The Relationship Between the IP Policy and Antitrust Policy
A. Approaches to Market Diversity
B. Changing Attitudes toward Antitrust and IP
C. Assessing Anticompetitive Restraints: ‘Scope-of-the-Patent’ Test
III. IP and Antitrust: Specific Issues and Applications
A. Assessing Market Power in IP-Intensive Markets
B. Horizontal Restraints: Price Fixing and Market Division
C. Vertical Restraints Involving IPRs
D. Patent Pools
E. Exclusionary Practices
1. Walker process and unreasonable infringement claims
2. Acquisitions
3. Refusal to license
IV. IP Law’s Own Internal Rules for Facilitating Competition
A. The First Sale (Exhaustion) Doctrine
B. ‘Misuse’
C. Competition-Based Limitations on ‘Functionality’ Protection
References
I. INTRODUCTION: THE POLITICAL ECONOMY OF

A legal system that relies on private property rights to promote economic development
and progress must consider that profits can come from two different sources. First,
both competition under constant technology and innovation promote economic growth
by granting some returns to the successful developer and some to society. An effective
innovation policy will ensure developer returns adequate to compensate for its investment
and risk. Competition and innovation both increase output. Second, however, profits can
also come from practices that reduce output, in some cases by reducing quantity, or in
others by reducing innovation.
Intellectual property rights (IPRs) and competition policy were once regarded as being
* James G. Dinan University Professor, Penn Law and the Wharton School, University of
Pennsylvania. Thanks to Peter S. Menell and Erik Hovenkamp for comments.
231

in conflict. IPRs create monopoly, which was thought to be inimical to competition. By

contrast, competition policy favors free entry and asset mobility, which IPRs limit in
order to create incentives. Today our view of this relationship is more complex. First,
most IPRs are insufficient to produce durable monopoly, although they do facilitate
product differentiation. Second, we tend to see intellectual property (IP) rules as creating
a property rights system in which competition exists for the property rights themselves.
Firms compete by innovating and appropriating whatever payoffs they can capture,
including IPRs. Third, and most importantly, we define competition in terms of output or
welfare rather than simple rivalry. A market structure or practice that increases output is
more ‘competitive’ than a lower output alternative, even though the amount of immediate
rivalry among firms is less. For example, output in the cellular phone market is much
higher because hardware, software, and telecommunications links are all networked by
cooperative agreements and standard setting.
Under conventional neoclassical assumptions, both innovation and competition
increase output, whether measured by the number of units or their quality. At the same
time, however, excessive IP protection limits competition by reducing asset mobility
further than necessary to facilitate innovation. The policy trick is to find the ‘sweet spot’
where the aggregate effects of IP competition and exclusion are optimized.
It is firmly established that innovation contributes significantly more to economic
growth than does competition under constant technology (Solow, 1957; Bohannan and
Hovenkamp, 2012; Grossman and Helpman, 1994; Aghion and Howitt, 1998). While the
theoretical and empirical literature employ different and sometimes inconsistent models,
all agree on this basic conclusion (Helpman, 2004; Schumpeter, 1943; Solow, 1956; Romer,
1990; Aghion and Howitt, 2007). In addition, the ‘debate’ between Joseph Schumpeter’s
(1943) position that monopoly is more favorable to innovation and Kenneth Arrow’s
(1962) position that competition is more favorable is somewhat settled, mainly in Arrow’s
favor. A broad consensus today is that the market structure/innovation curve is a lopsided,
inverted ‘U’ (Scott and Scott, 2014; Arai, 2013; Aghion et al., 2005). Neither monopoly
nor atomistic competition is especially conducive to innovation. Rather, most innovation
occurs in moderately competitive, product differentiated markets. Some more recent
literature tilts the inverted U more to the competitive side, concluding that on balance
more competition yields more innovation (Hashmi, 2011; Schmitz and Holmes, 2010;
Menell and Scotchmer, 2007, pp. 1526–30)).
Although the relationship between innovation and economic growth is clear, the rela-
tionship between innovation rates and particular IPR systems is not. One problem is that
while IP systems may encourage innovation, they also act as impediments to the diffusion
or cumulation of ideas through the economy (Menell and Scotchmer, 2007; Moser, 2013).
The literature on the relationship between the strength of patent systems and the rate of
economic growth is at best inconclusive, with most of it suggesting little or no correlation
(Gould and Gruben, 1996; Park and Ginarte, 1997; Belleflamme, 2006).1 Relatively little
1
One study finds a correlation between the existence of a patent system and total factor pro-
duction (TFP) growth, but also concludes that there is an inverse correlation between the strength
of patent rights and TFP growth (Chang et al., 2014). The authors conclude that while patent
rights lead to more patents,

Intellectual property and competition 233
literature exists that correlates innovation or growth rates with the existence or strength of
any specific patent doctrine, although there is a robust ‘meta’ empirical literature on the
behavior of courts or judges with respect to certain doctrines (Rantanen, 2013; Mojibi,
2010; Anderson and Menell, 2013; Seaman, 2012; Crouch, 2010).
Further, the innovation effect of IPRs is market specific, just as is true of other market
characteristics such as economies of scale, product differentiation, ease of entry, or
nature of information flow. The competitive impact of IPRs also varies with differences
in industry structure and the market position of the rights owners. For example, for a
dominant firm additional IP protection may serve to entrench or prolong its monopoly
position, while the same right held by a small rival might serve to destabilize the dominant
firm and make the industry more competitive. Therefore, who gets a particular IPR can
be important for competition policy.
That IPR performance varies from one market to another seems beyond dispute.
For example, chemical and pharmaceutical innovations tend to benefit from a robust
patent system with protection of fairly long duration. By contrast, in some markets for
information technologies the patent system is much less valuable and may even produce
greater harms than benefits. The same thing is true of copyright. For example, many
books have long economic lives and can benefit from a lengthy term of protection, while
more journalistic writing and software does not. The optimal term may also vary with the
degree of market competitiveness, with greater competition conducive to shorter terms.
Further, a trade-off exists between duration and breadth: a patent with a shorter life but
broader protection may provide the same incentives as one with longer life but narrower
protection (Gilbert and Shapiro, 1990; Merges and Nelson, 1990; Khoury, 2010).
Unfortunately, our knowledge about market diversity has had little impact on the
creation or application of intellectual property law, which is not particularly sensitive to
issues of market structure, information transmittal, ease of copying, and other barriers to
market entry or mobility. This is in very sharp contrast to antitrust law, which is acutely
sensitive to market differences, perhaps overly so. For example, questions concerning
the legality of a merger or allegedly monopolistic practice can be answered only after a
detailed expert inquiry into the markets at issue and the rationally expected results of
certain practices. By contrast, questions of patent validity, scope, and infringement are
largely indifferent to the markets in which these queries occur. Likewise, the legislated
term of IPR protections is largely invariant to the particular market in which the pro-
tected product is sold.
Because our information about the relationship between innovation and specific IP
rules is so inadequate, opinions often go to extremes. Some believe that the patent or
other IP systems are worthless or even harmful because they hinder rather than promote
our findings also suggest that patent rights slow the diffusion of new innovations throughout
the economy, as we find that the effect of patents on TFP growth is weaker in countries with
stronger patent rights. Our results suggest that finding the optimum level of patent protections
requires the consideration of these two offsetting effects. (Falvey, 2006)
According to Falvey (2006), at least in middle income countries, IPRs cause more harm by
restricting the dissemination of technology than they contribute to economic growth. In Torrance
and Tomlinson (2011), an experimental test showed an inverse relationship between innovation and
patenting.

innovation (Boldrin and Levine, 2008), while others defend IPRs enthusiastically
(Epstein, 2010; Mossoff, 2013). Even within the United States Supreme Court these views
have gyrated from periods when the Court was extremely tolerant of patenting and patent
practices, to periods in which it struck down nearly every patent it encountered and held
exaggerated views about the anticompetitive effects of patent practices (Hovenkamp,
2015d).
Further, IPRs are hardly the only inducement to innovation, and their relative impor-
tance varies from one market to another. When firm managers are questioned, a plurality
believe that the biggest inducement is first mover advantages, while patent protection is
at best secondary (Bohannan and Hovenkamp, 2012, pp. 100–102).2 In some markets,
such as digital content, copying is so cheap and quick that little innovation would occur
but for IP protection. Innovations in processes that are not readily observable or reverse
engineered might be better protected by simple first mover advantages or trade secrets.
Patent protection is secondary and may even be counterproductive to the extent that
patenting requires disclosure. Some markets exhibit high rates of innovation without any
intellectual property protection at all (Raustiala and Sprigman, 2012).
The lack of empirical validation for specific IP rules is troublesome, because some
rules may be far from optimal. A good example is the way that patent law’s requirement
of nonobvious subject matter (35 U.S.C. § 103) is administered. Because patent infringe-
ment does not require copying or even knowledge of another’s patent, it is crucial that
the nonobviousness requirement be interpreted strictly, keeping patent issuance within
proper bounds: we do not want to give patents on things that independent entrepreneurs
would develop on their own. An empirically based inquiry into nonobviousness would
consider the extent to which new technology results from copying rather than independ-
ent invention, a forward-looking inquiry. Whether one can infer nonobviousness from
commercial success is debatable, but doubtful (Merges, 1988). But in any event, that
is not how nonobviousness subject matter is actually determined. Patent examiners or
courts deciding infringement cases assess nonobviousness, or ‘inventive step,’ by looking
backward through prior art. By contrast, entrepreneurs think forward, considering new
things to try from their current position. The likely result is that far too many patents
are granted on things that other businesses develop on their own in the ordinary course
of competition. The recent experience with non-practicing patent holders in informa-
tion technology markets suggests as much. Most of the defendants in those cases are
independent developers rather than copyists.
The statutory systems of competition law and IPRs differ significantly from one
another. Most of the United States antitrust laws are highly general and do not reflect
specific ‘deals’ between legislators and particular special interests. The Sherman and
Clayton Acts simply condemn practices that ‘restrain trade,’ ‘monopolize,’ or have effects
that ‘may be substantially to lessen competition’ (15 U.S.C. §§ 1, 2, 14, 18).3 As a result,
assessment of specific practices is left largely to judges. In addition, after more than 30
2
On the use of alternative funding mechanisms, such as prizes or direct government finance
of research (Menell and Scotchmer, 2007, pp. 1530–34).
3
One exception is the Robinson-Patman Act, which was in fact the product of a deal
between retailers and Congress during the Great Depression (15 U.S.C. § 13; Hovenkamp, 2015d,
pp. 225–32).

years of redefinition and retrenchment, antitrust policy in the United States has become
much more focused on promoting consumer welfare, which it does by facilitating struc-
tures or practices that maximize output, measured by quantity or quality (Bohannan and
Hovenkamp, 2012).
By contrast, most IPR systems are detailed codes that reflect considerable producer
involvement but relatively little input from consumers. The 1976 Copyright Act currently
in force, together with the Copyright Term Extension Act, are examples (Bohannan, 2006;
Patry, 1996), but the patent laws are not far behind (Merges, 2000). For example, over history
Congress has repeatedly granted retroactive term extensions to both patents and copyrights
(Ochoa, 2001; Hovenkamp, 2016c). Retroactive extensions do not facilitate innovation
to the extent that the inventions to which they apply have already been created. They are
pure rent seeking, prolonging exclusive rights, and reducing output. Such extensions have
come to the Supreme Court twice, 150 years apart. In Bloomer v. McQuewan, 55 U.S. 539
(1852), the Supreme Court held that retroactive patent extensions could not be applied to
patented articles that had already been sold, thus creating the foundation for the modern
patent ‘exhaustion’ doctrine (Hovenkamp, 2016b, 2016c). In Eldred v. Ashcroft, 537 U.S.
186 (2003), the Supreme Court upheld a retroactive extension of the copyright term.
This history is unsettling because consumer welfare should be the ultimate goal of inno-
vation policy just as it is of traditional competition policy. Consumers profit from lower
prices and higher innovation rates, giving them the correct set of incentives to determine
optimal IP rules. By contrast, producer incentives are more mixed. While producers profit
from lower costs and increased innovation, they also profit from increased protection for
their own IPRs or reduced protection for the innovations of rivals, whether or not these
protection levels are optimal (Hovenkamp, 2014).
II. THE RELATIONSHIP BETWEEN THE IP POLICY AND

ANTITRUST POLICY
A. Approaches to Market Diversity
While the antitrust laws do not explicitly require different analysis for different markets,
the spare, highly general statutes have been interpreted that way at least since the Supreme
Court’s Chicago Board of Trade decision in 1918. That decision approved an agreement
that literally fixed prices for after-hours trading occurring after the open market had
closed. Such price-fixing was unique to that market and, in the Court’s view, promoted
rather than restricted competition. The antitrust merger provision contained in the
Clayton Act, 15 U.S.C. § 18, condemns acquisitions whose ‘effect may be substantially to
lessen competition or tend to create a monopoly’—a requirement that has always been
held to require highly specific market analysis. For well over a half-century the law of
monopolization under Sherman Act § 2, 15 U.S.C. § 2, has required detailed inquiries
into market structure, producing different outcomes in different industries (United States
v. Aluminum Co. of Am., 148 F.2d 416 (2d Cir. 1945); United States v. E.I. du Pont de
Nemours and Co., 351 U.S. 377 (1956); Brooke Group Ltd. v. Brown and Williamson
Tobacco Corp., 509 U.S. 209 (1993)).
By contrast, IP law largely disregards market differences (Burk and Lemley, 2003;

Menell and Scotchmer, 2007). Terms of protection are largely invariant to the industry,
even though rates of technological turnover vary widely. If protected technology or
expression routinely becomes obsolete in the market before IPRs expire, then the
Constitution’s provision authorizing Congress to create patents or copyrights only for
‘limited times’ (U.S. Const. art. I, § 8, cl. 8) is largely meaningless. The premise of that
provision is that the protection period should be sufficient to induce innovation, but after
expiration the protected good goes into the public domain. Even requirements such as
nonobvious subject matter for patents generally avoid market specific questions about
how information is disseminated in a particular market.
An ideal IP policy truly concerned with innovation would need to develop more empiri-
cally driven, market specific rules, reflecting how innovation works in different situations,
what amount and nature of inducement is required, and the extent of harm caused by the
resulting exclusion. Offsetting this, of course, would be the higher transaction and enforce-
ment costs involved in enforcing a system that contemplates greater market diversity.
B. Changing Attitudes toward Antitrust and IP
Competition policy and IP policy should be regarded as complements. They share eco-
nomic welfare as a goal, and an optimal policy includes elements of both. Public policy
has been erratic, however, and the two legal systems have not always accommodated
each other in socially beneficial ways. Prior to 1917 the Supreme Court approved nearly
every patent practice that had been alleged to restrict competition, including toleration
of product price-fixing in a patent pool (E. Bement & Sons v. Nat’l Harrow Co., 186 U.S.
70 (1902)); granting a dominant firm an injunction against infringement of an externally
acquired but unpracticed patent (Cont’l Paper Bag Co. v. E. Paper Bag Co., 210 U.S. 405
(1908)); and permitting tying of patented and unpatented goods (Henry v. A.B. Dick Co.,
224 U.S. 1 (1912)). One important exception was Standard Sanitary Mfg. Co. v. United
States, 226 U.S. 20 (1912), which condemned a product price fix covering the entire
bathroom fixture industry. The price stipulation was included in a patent license for an
enameling process that represented a minor component of the finished product.
A single mention of patents in the 1914 Clayton Act, 15 U.S.C. § 14, provoked a
dramatic change. Beginning in the 1917 Motion Picture Patents case, which overruled
Henry, the Supreme Court embarked on a war against patent practices thought to be
anticompetitive, in the process developing an expansive, judge-made doctrine of patent
‘misuse,’ of which more later (Motion Picture Patents Co. v. Universal Film Co., 243 U.S.
502 (1917); Henry v. A.B. Dick Co., 224 U.S. 1 (1912)).
Beginning in the late 1930s the Supreme Court applied increasingly harsh standards
for patent issuance, eliciting Justice Robert H. Jackson’s famous complaint that ‘the only
patent that is valid is one which this Court has not been able to get its hands on’ (Jungersen
v. Ostby & Barton Co., 335 U.S. 560 (1949)).
Three dispersed events gradually turned the tide again. First was the 1952 Patent
Act, a significant revision, which restated the patentability requirement as ‘nonobvious’
subject matter and also limited the reach of patent misuse law (Duffy, 2007; Hovenkamp,
2015d). The second was the establishment of the Federal Circuit Court of Appeals in
1982, with a mandate to unify and strengthen patent law (Dreyfuss, 1989, 2008). The
third development, which occurred more gradually and within antitrust and patent

misuse law, was a doctrinal reformulation that required much more explicit proof of
anticompetitive effects (Bohannan and Hovenkamp, 2012). The high point of antitrust
hostility toward perceived patent abuses was 1970, when the US Antitrust Division issued
its ‘nine no nos’ of patenting that were almost certain to provoke an antitrust challenge
(Wilson, 1970; Hovenkamp, 2015d). Today nearly all of the ‘nine no nos,’ including such
things as mandatory packaging licensing, grantback clauses, reach through royalties,
and resale price maintenance, are widely regarded as competitively benign in most situ-
ations (Hovenkamp, 2015a). Antitrust courts and scholars increasingly came to believe
that many post-issuance patent practices that had been condemned as ‘misuse’ were in
fact competitively harmless. This was particularly true of tying arrangements, the most
frequent generator of misuse findings, as well as vertical price and nonprice restraints,
package licensing, provisions that tied royalty payments to unpatented goods, and most
unilateral refusals to license (Bowman, 1973; Hovenkamp, 2018a).
One important result of significant antitrust revision is that overreaching is less likely to
occur today than it was 30 years ago. By contrast, patent law has continued on an expan-
sion course in both issuance and doctrine that until recently seemed unstoppable. Today
antitrust law is in a much better position to accommodate concerns about innovation
than patent law is to accommodate concerns for competition. Antitrust law’s sensitivity to
innovation manifests itself in several ways. One is a very broad rule that innovation itself
can almost never be an antitrust violation, no matter how exclusionary, as several courts
have held (Allied Orthopedic Appliances, Inc. v. Tyco Health Care Grp. LP, 592 F.3d 991
(9th Cir. 2010); In re Apple iPod iTunes Antitrust Litig., 2014 WL 6783763 (N.D. Cal.
2014); Areeda and Hovenkamp, 2009–15). One limited exception is situations where the
cost of product changes is very small in relation to competitive harm and the changes are
readily reversible. This is true mainly of software, where a minor change in code can serve
to make rivals’ products incompatible (Newman, 2012). Another area is the deferential
treatment that the courts have afforded to settlements of IP lawsuits. For example, in
Clorox Co. v. Sterling Winthrop, Inc., 117 F.3d 50 (2d Cir. 1997), the court approved a
market division agreement settling a trademark dispute (Hovenkamp, 2015a). A third
area is increasingly strict limitations on the use of antitrust to challenge anticompetitive
IP infringement actions, as created by the Supreme Court in Walker Process Equip., Inc.
v. Food Mach. and Chem. Corp., 382 U.S. 172 (1965), but limited by Dippin’ Dots, Inc. v.
Mosey, 476 F.3d 1337 (Fed. Cir. 2007) (Bohannan and Hovenkamp, 2012, pp. 290–324).
Yet another is deferential treatment of technology sharing agreements under antitrust law,
which rarely condemns them unless they involve explicit restraints in the product market
(Hovenkamp, 2015a).
By contrast, patent case law sometimes operates as if competition were the affirmative
evil to be resisted. One example is the Federal Circuit’s 2014 decision in Trebro Mfr., Inc.
v. Firefly Equip., LLC, 748 F.3d 1159 (Fed. Cir. 2014) (Hovenkamp and Cotter, 2015). The
court permitted a dominant firm in a concentrated market to enjoin patent infringement
on unpracticed patents. The dominant firm had purchased two patents from a third party
that covered an alternative technology to that in its own product. After the acquisition,
it continued to use its older technology and brought an infringement suit against the
defendant, a recent entrant whose technology very likely infringed the acquired patents.
The Federal Circuit distinguished a line of lower court decisions which had refused
injunctions to non-practicing entities, following the Supreme Court’s decision in eBay Inc.

v. MercExchange, L.L.C., 547 U.S. 388 (2006) (Sichelman, 2014). In this case the patentee
was actually competing in the market, even though it was not practicing the patent whose
infringement was claimed. No one apparently raised an antitrust issue. Nevertheless, the
Court’s lack of foresight did considerable harm to competition by giving dominant firms
an excuse to buy up competing technologies in order to keep them out of production, thus
limiting the avenues through which new entry can occur.
C. Assessing Anticompetitive Restraints: ‘Scope-of-the-Patent’ Test
Historically, competition policy presumed that an IP practice that increased the profitabil-
ity of an IP right would also increase the incentive to innovate. Competition law enforcers
should stand aside if the practice fell ‘within the scope of the patent’ (Hovenkamp, 2015e).
This formulation originated in the nineteenth century as a rationale for the exhaustion,
or ‘first sale,’ doctrine, which held that ‘when the machine passes to the hands of the
purchaser, it is no longer within the limits of the monopoly’ (Bloomer v. McQuewan, 55
U.S. 539 (1852)). For example, even if a patent license limited the geographic range over
which a good could be used, once the good was sold that right could no longer be enforced
against the purchaser by means of a patent infringement suit (Adams v. Burke, 84 U.S. 453
(1873)). Later on, the Supreme Court used the ‘beyond the scope’ formulation to describe
overly broad patent claim constructions, as in Coupe v. Royer, 155 U.S. 565 (1895); or
overly broad interpretations of the patent doctrine of equivalents, which extended patent
coverage to things that did not literally fall within the patent’s claims. Johnson & Johnston
Assocs., Inc. v. R.E. Serv. Co., 285 F.3d 1046 (Fed. Cir. 2002) concluded that a broad
infringement claim under the doctrine of equivalents was an attempt to extend patent
beyond its rightful scope (Sarnoff, 2005). Beginning in the 1930s, the formulation was also
employed in patent ‘misuse’ cases, particularly those involving the tying of unpatented
goods. The tie was said to extend the patent’s power beyond its proper scope by bringing
the unpatented tied product within the patent monopoly. For example, in Carbice Corp.
of Am. v. Am. Patents Dev. Corp., 283 U.S. 27 (1931), the Supreme Court held that a
patentee’s tie of unpatentable dry ice to its patented ice box was an attempt to control
‘unpatented material’ and thus ‘beyond the scope of the patentee’s monopoly.’
The scope-of-the-patent formulation was also used defensively, however, to exonerate
practices challenged as anticompetitive but that were found to be within the patent’s
scope. For example, in 1926 the Supreme Court upheld product price-fixing contained in
patent licenses on the theory that setting the product price was the patentee’s right, and
the license agreement did no more than retain that right, while transferring the right to
produce to the licensee (United States v. General Electric Co., 272 U.S. 476 (1926)). In the
1970s Ward Bowman’s important book on patent and antitrust law envisioned the patent
as a walled garden protecting everything within its scope, but not necessarily activities that
spilled outside (Bowman, 1973). In its decision in United States v. Line Material Co., 333
U.S. 287 (1948), however, the majority condemned a product price fix in a cross-license,
over the dissent of three Justices who objected that the price fix was within the scope
of the patent. The dissenters in the Supreme Court’s 2013 Actavis decision would have
exonerated a settlement agreement in which a patentee paid an accused infringer a large
sum to delay its entry into production, provided that the permitted entry date was prior
to the expiry of the patent (F.T.C. v. Actavis, Inc., 570 U.S. 136 (2013)). In that case, the

settlement agreement would be no more exclusionary than a judicial determination of

validity and infringement; thus, the agreement fell within the scope of the patent (F.T.C.
v. Actavis, Inc., 570 U.S. 136 (2013); Edlin et al., 2015).
Pay-for-delay settlements came into existence with the passage of the Hatch-Waxman
Act, which rewards a generic firm for being the first to challenge a pioneer’s patent or
entering upon that patent’s expiry. Under the Act, no subsequent generic can enter the
market until 180 days after the first generic to file an Abbreviated New Drug Application
(ANDA) actually starts producing. Prior to generic production the patent is virtually
immune from challenge by other potential competitors, because they have no right to
produce in any event. The situation gives the patentee and the generic infringement
defendant a strong incentive to share the patent monopoly, thus largely eliminating adver-
sity between them. Under the ‘scope-of-the-patent’ test the equilibrium duration of such
an agreement is the remaining term of the patent, assuming that the antitrust laws permit
such an agreement (Edlin et al., 2014; Hovenkamp, 2015e). That is, the joint-maximizing
agreement for the settling parties would share the returns permitted by the patent for its
full period.
But the Actavis majority rejected a scope-of-the-patent approach, perhaps heralding an
important change in antitrust analysis of patent practices. If patent rights are presumed
to be valid, valuable, and clearly defined, then the scope-of-the-patent formulation func-
tions much like similar scope formulations might do for, say, real property. But if patents
are of questionable validity, dubious value, or ambiguous scope, then the scope-of-the-
patent formulation can permit significant anticompetitive overreaching. This issue was
highlighted in Actavis because the legislative framework largely immunized suspiciously
weak patents from challenge while the pay-for-delay agreement was pending. Further,
because the owner of a robust patent would not pay much more than avoided litigation
costs in order to enforce its rights, the high pay-for-delay payment (often several hundred
million dollars) is a strong signal that the patent is invalid or, in a few cases, not infringed
(Edlin et al., 2013; Edlin, et al., 2014). For example, a landowner attempting to exclude a
trespasser would not pay the trespasser a large sum of money to stay off her land unless
she had serious doubts about the validity of her legal claim. If her title were good she
could exclude the trespasser by paying nothing more than litigation costs.
In other cases, the scope-of-the-patent formulation fails, not because the patents in
question are invalid, but because their value is very low in relation to the restraints in ques-
tion. Licenses that include product price-fixing are a good illustration. Even for relatively
sound patents, license fees range from 0.5 to 6 percent of sales, with rates below 3 percent
being the norm. The rates on individual patents can be much lower in patent intensive
technologies such as computers and telecommunications. Further, these rates are for
licensed patents, and only a small percentage of patents are ever licensed. By contrast, the
markups of successful cartels often run in the range of 10 to 50 percent (Connor, 2014).
If the firms in an industry cross-license their patents and also fix the product price, the
agreement as measured by a scope-of-the-patent test attributes the value of the entire
cartel markup to the patents.
Justice Breyer’s majority opinion in Actavis held that courts evaluating such settle-
ments need not address questions of patent validity or infringement. That proposition is
consistent with long-standing reluctance by federal judges to review the IP merits when
considering competition-based challenges to settlements, except for obvious cases of

patents that are almost certainly invalid or not infringed. Most of those cases go on to
uphold the settlement, however, while the Actavis decision did not.
More importantly, as Actavis recognized, antitrust’s economic approach is designed to
create appropriate incentives at the point of decision. The relevant question is not the ex
post one whether the patent was valid and infringed, but rather the ex ante question of
what the parties’ expectations were at the time the settlement was entered. By settling, the
parties have already implicitly agreed that getting a judicial determination of patent valid-
ity and infringement is not worth the cost and attendant risk of a judicial determination.
As a result, it makes little sense to insist on that same query before passing judgment on
the settlement (Edlin, et al., 2015).
Of course, most settlements raise no competition issues because the settlements them-
selves tend to increase rather than decrease output. The most common settlement of an IP
infringement dispute is a production license under which the defendant pays the plaintiff
for the right to produce. Such a license is likely to increase rather than decrease output,
but in any event production licenses are explicitly authorized by § 261 of the Patent Act.
They are legal whether or not they are in settlement of litigation. The more problematic
settlements are those that fix product prices, divide product markets (as in Actavis), or
in some cases that involve an agreement among the settlors not to license to or otherwise
deal with third parties, such as the Supreme Court condemned in United States v. Singer
Mfg. Co., 374 U.S. 174 (1963).
Finally, one thing that makes an Actavis style pay-for-delay settlement unusual is that it
does not involve a license at all, but at most an agreement to license at some future date.
That is why Justice Breyer’s opinion for the Court observed that, while the Patent Act
explicitly permits licensing, the agreement providing for delayed entry was not authorized
by the Patent Act. Indeed, an equilibrium agreement under the scope-of-the-patent test
advocated by the dissenters would never be a license: for the entire remaining duration of
the patent the generic would not produce. Once the patent expires the generic is free to
produce without a license. Until actual production under a license occurs, the settlement
is nothing more than a naked market division agreement. Even so, Actavis held that
the agreement in question should be addressed under antitrust’s rule of reason, which
requires proof of market power and anticompetitive effects. It also held, however, that
both power and harmful effects could be inferred from the large payment itself.4
III. IP AND ANTITRUST: SPECIFIC ISSUES AND

APPLICATIONS
Prior to patent issuance the patent process operates under intensive government supervi-
sion and control. To be sure, improper conduct in patent prosecution is not rare, but
the patent system itself has tools for policing it. Further, riding herd on the procedures
and rules of other federal agencies is not antitrust’s purpose. Even if we believe that the
existing system issues too many patents, that too many of these are worthless, or that
4
Reverse payments in the context of adjudication before the Patent Trial and Appeal Board
(PTAB) can raise analogous issues (Hovenkamp and Lemus, 2016).

the process has other flaws, these are virtually never antitrust problems. This position
is mandated by the ordinary antitrust rules of implied immunity, which limit or remove
antitrust involvement from activities that are actively regulated by other federal agencies
(Areeda and Hovenkamp, 2009–15).
Once a patent is issued, however, the situation is much different. Patents are largely
treated as property rights requiring little government supervision, other than the United
States Patent and Trademark Office’s (USPTO) power to re-examine, collect renewal fees,
and a few other housekeeping matters (Hovenkamp, 2015c). Because issued patents are
largely subject to private control, antitrust policy becomes relevant. One important factor
is whether the practice in question is expressly authorized by the Patent Act. Under the
rules of express immunity, a practice that is compelled or authorized by a federal statute
cannot be an antitrust violation, provided that the practice stays within the expressly
authorized boundaries.
After considering how market power should be assessed in IP-intensive markets,
this section briefly addresses specific intellectual property practices that might also be
challenged as antitrust violations. All are post-issuance practices and most of them are
either not authorized by the Patent Act itself, or else they fall outside the scope of the
authorization. As a result, antitrust analysis is appropriate. Of course, this does not mean
that they are unlawful. Nor does it entail that the presence of an IP right or license is
irrelevant (Cotter, 2015).
A. Assessing Market Power in IP-Intensive Markets
No anticompetitive practice can succeed unless its participants have significant market
power, which is the power profitably to raise prices above cost by reducing output. This
requirement applies both to anticompetitive exclusion and anticompetitive collusion. To
be sure, certain practices such as price-fixing are said to be unlawful ‘per se,’ which means
that proof of illegality does not require a showing of market power. This is not because
market power is irrelevant, however. To the contrary, naked practices such as price-fixing,
which produce no efficiency gains to the participants, are profitable only on the premise
that power exists. As a result, proper identification of the practice eliminates the need to
assess market power separately (Areeda and Hovenkamp, 2009–15).
In 2006 the Supreme Court overruled a half-century old presumption that a patent
conferred sufficient market power on its owner to make certain anticompetitive practices
such as tying unlawful (Illinois Tool Works Inc. v. Indep. Ink, Inc., 547 U.S. 28 (2006),
overruling International Salt Co. v. United States, 332 U.S. 392 (1947)). In United States v.
Loew’s, Inc., 371 U.S. 38 (1962), the Supreme Court had also extended the presumption to
copyrights, and a few lower courts had applied it to trademarks (Siegel v. Chicken Delight,
Inc., 448 F.2d 43 (9th Cir. 1971)). Most courts limited the presumption to tying cases, but
where it applied the challenger needed to show only that the challenged restraint involved
an IPR-protected product, and the requisite market power would then be presumed. All
these decisions are now overruled.
The end of the power presumption hardly means that IPRs are irrelevant to inquiries
about market power. Today, they are properly regarded as an important factor in estab-
lishing power (Areeda and Hovenkamp, 2009–15). A few very powerful patents and some
software copyrights may have so much exclusionary power that they give their owners

dominant market positions. One likely historical example is Microsoft’s Windows operat-
ing system, which is protected from duplication by copyright and some patents (United
States v. Microsoft Corp., 253 F.3d 34 (D.C. Cir. 2001)). Other good historical examples
are the patents that protected Polaroid’s self-developing camera and film system, which
Kodak tried in vain to invent around (Fierstein, 2015), and the array of patents that
Xerox acquired from outside inventors that led to its long-held dominance of plain paper
copying technology (SCM Corp. v. Xerox Corp., 645 F.2d 1195 (2d Cir. 1981)).
Some aggregations of patents can become so essential to operation that they give
significant market power to their owners, at least when the aggregation is owned by a
single firm. Of course, aggregations of essential patents are often owned by pools in
which a large number of firms have nonexclusive rights. Good examples are MPEG-LA,
a patent pool and standards association whose members control standards for digital
video technology; and 3GPP, whose members control the technology for 3G and 4G
wireless telecommunications. Once a particular patent in such a pool is declared ‘stand-
ards essential,’ it may be necessary for any firm wishing to compete in that technology
to purchase a license. That obligation can confer significant market power, limited by
the fact that standards-essential patents, or SEPS, are also typically subject to FRAND
(‘fair, reasonable, and non-discriminatory’) licensing obligations, which are generally
interpreted to require licensing to willing participants at fair and nondiscriminatory rates
(Contreras, 2015).
So far there have been few antitrust cases challenging the creation and enforcement
of SEPs or FRAND obligations, and these have been largely unsuccessful. For example,
Golden Bridge Tech., Inc. v. Motorola, Inc., 547 F.3d 266 (5th Cir. 2008), rejected the
antitrust claim of an inventor whose technology was rejected by a standard setting organi-
zation (SSO) in favor of alternative technologies. Legal control of SEPs lies largely with
patent law, contract law, or the court’s general equity powers. In Apple, Inc. v. Motorola,
Inc., 757 F.3d 1286 (Fed. Cir. 2014), the court held that the owner of a SEP could not
obtain an injunction against a user; and in Qualcomm, Inc. v. Broadcom Corp., 548 F.3d
1004 (Fed. Cir. 2008), the court applied the judge-made doctrine of estoppel against one
who reneged on its promise to subject its patents to a FRAND commitment. At this writ-
ing Qualcomm is facing separate lawsuits brought by the Federal Trade Commission and
Apple, claiming that Qualcomm is tying SEPs to devices that it sells, or licensing only on
the condition that its patents be used exclusively with Qualcomm devices (E. Hovenkamp,
2018).
IPRs of all forms can limit asset mobility and facilitate product differentiation. In such
markets prices will be higher than short-run marginal cost, even though the market has
several competing firms. The impact of IPRs in these situations depends heavily on the
number of firms in a market and the strength of the IPRs in question. Suffice it to say that
many products from automobiles to computers to kitchen appliances contain numerous
patents but are yet sold in moderately competitive, product differentiated markets.
One technical difficulty for assessing power is that IP development often requires high
fixed costs invested at the front end, and fairly low marginal costs. Whether acquisition
costs are fixed or variable depends heavily on whether the IPRs in question are developed
internally or licensed from outside inventors. For example, internal research often is very
costly at the front end and these costs, once invested, do not vary with output. By contrast,
licensing in the same technology by per unit or per dollar royalties becomes a variable

cost to the licensee. Most of the technical tools used for market power measurement
examine the relationship between price and marginal cost. The result can be false posi-
tives, depending on the prevalence of IPRs and the extent of fixed costs. For example, an
unpatented living room chair with a patented recliner button may sell at a small markup
over cost, reflecting licensing of the button patent. At the other extreme, purely digital
products such as streamed e-books, songs, or software may have distribution costs very
close to zero, meaning that the licensor’s entire price is markup. In that case, any measure
of market power based on the relationship between price and short-run marginal cost will
exaggerate the seller’s power.
Because digital content is so easily duplicated, the ability to sell at a substantial markup
over short-run cost is largely a result of IP protection. For example, one can obtain an
e-book version of Moby Dick at a price of zero, even though it is very famous and widely
read. Moby Dick is in the public domain, which means that no one is earning a royalty
on its sales and copying is free. By contrast, the e-book version of a mediocre but recent
novel will be much higher because royalties must be paid and it cannot be copied without
a license from the publisher or author. For antitrust purposes, the main takeaway from
these situations is that assessment of price-cost margins is rarely a useful way of assessing
market power in markets for purely digital goods. Theoretically, one could address the
problem by querying whether the returns to a product are significantly positive over
its entire life. For example, the fact that a digital computer program sells at a high ratio
of price to short-run cost tells us nothing if the product becomes obsolete or loses its
commercial viability before recouping development costs. As a practical matter, these
measurements can be very difficult to make, particularly when the IP right in question is
a copyright with an effective duration of a century (Hovenkamp, 2016a).
With some exceptions, non-patent IPRs make even smaller contributions to power than
do patents. Copyrights and trademarks are easier to obtain than patents are. A few highly
popular publications or computer programs are counterexamples, but generally one
cannot infer significant power merely from the existence of an IPR of any kind (Areeda
and Hovenkamp, 2009–15).
B. Horizontal Restraints: Price Fixing and Market Division
A restraint is ‘horizontal’ if the participants are competitors or would be competitors

but for the restraint. Identification of firms as ‘competitors’ is usually a reference to the
product or service markets in which the firms operate, although it may also refer to the
technologies that they develop or license. In any event, it is always important to distin-
guish restraints in the patent and licensing market from restraints in the product market.
An example is price-fixing. Setting a price is inherent in licensing and rarely anticom-
petitive. If firms cross-license, they must necessarily agree on the price that each will
charge to the others, even if the price is zero. Price-fixing in the product market is another
matter and is highly suspicious. For example, firms with worthless patents or other IPRs
might use licenses or cross-licenses as a cover for price-fixing, as Judge Posner observed
in his opinion in Asahi Glass Co. v. Pentech Pharm., Inc., 289 F. Supp. 2d 986 (N.D. Ill.
2003).
As noted previously, the competition problem with product price-fixing actually
reaches far beyond invalid patents. Even if a patent is valid and essential, it may contribute

only a small amount to a product’s value. As a result, the market price of the license can
be far less than the cartel markup on the product. For this reason, product price fixes in IP
licenses should be regarded as competitively harmful whether or not the IPRs in question
are valid. The corollary is that product price fixes in patent licenses can be condemned
without inquiry into patent validity or infringement.
Market division agreements operate economically much like price-fixing. By dividing
up the market (by territory, customer, or product) a group of firms can create individual
monopolies for themselves. As a cartel device, market division can be superior to price-
fixing if the firms have differing costs or, for other reasons, disagree about the price that a
cartel should charge. One important difference between price-fixing and market division
is that the Patent Act expressly authorizes patentees to grant exclusive licenses to ‘any
part’ of the United States (35 U.S.C. § 261), thus making most domestic territorial division
agreements lawful. While the Patent Act says nothing about licenses restricted to specific
customers or products, these ‘field of use’ restrictions are treated leniently, mainly because
they are viewed as organizers of production enabling the patentee to take advantage of
the unique characteristics of different producers. For example, in Gen. Talking Pictures
Corp. v. W. Elec. Co., 304 U.S. 175 (1938), the Supreme Court upheld an arrangement in
which the patentee reserved to itself the market for commercial use of its patented sound
amplifier, while other licensees were authorized to make the amplifiers only for residential
customers. The Federal Circuit Court of Appeals has held that field of use restrictions
must be evaluated under antitrust’s rule of reason (B. Braun Med., Inc. v. Abbott Labs.,
124 F.3d 1419 (Fed. Cir. 1997)). Field of use restrictions become more suspect, however,
if they take the form of product market division among competing manufacturers. It
is also worth noting that while § 261 of the Patent Act authorizes an exclusive territory
agreement between a patent owner and a licensee, it does not authorize agreements among
the licensees themselves.
As is true of price-fixing, the tolerance for market division agreements applies to the
IP right, not to products that might include it. For example, suppose that Ford patents
a desirable windshield wiper blade and licenses Chrysler to sell cars with the patented
blade in any state except California. That would be a territorially restricted license
expressly authorized by the Patent Act. Ford could very likely also authorize Chrysler
to put the blade only on its pickup trucks, but not its cars. That would be a field of use
restriction and would ordinarily be lawful under antitrust law’s rule of reason. What
Ford could not do, however, is agree that Chrysler would not sell any pickup trucks in
California, whether or not they contain the patented blade. That would be a restraint
on the product market rather than on use of the patent. Unless other factors suggesting
joint development were present, that agreement would be unlawful per se under the
antitrust laws.
C. Vertical Restraints Involving IPRs
A restraint is purely vertical when the parties stand in a buyer-seller relationship but are
not actual or potential competitors in either the product market or the licensing market.
Because every license agreement has a buyer and seller, they are all vertical as to the IPR
license itself. The more important question is the relationship of the parties in the underly-
ing product (or service) market. Today the antitrust attitude toward vertical restraints is

benign, although it was not always so (Hovenkamp, 2015d). Resale price maintenance
(RPM), vertical nonprice restraints, and tying were all once unlawful per se.
Vertical restraints come in two classes, generally called ‘intrabrand’ and ‘interbrand,’
even though at least some of the products and services that they control are not branded
at all. A restraint is said to be intrabrand if it controls distribution only of the supplier’s
own product. It is interbrand if it places limits on the products of rivals. The principal
intrabrand price restraint is RPM, or seller dictation of the price at which its own product
can be resold. As a general rule, the fact that the price restraints are included in an IP
license is irrelevant, and when the per se rule against RPM was in place it applied to
patented and copyrighted goods, as well as those protected by trade secrets (Areeda and
Hovenkamp, 2009–15). Since 2007, RPM has been assessed under the rule of reason
(Leegin Creative Leather Prod., Inc. v. PSKS, Inc., 551 U.S. 877 (2007)). Today, few
instances are found to be unlawful.
Vertical nonprice restraints restrict a dealer’s or retailer’s sale of the supplier’s own
product in some way other than by setting price. The most common ones are territorial
restrictions, customer restrictions, and product restrictions. There are also numerous
others, such as restrictions regulating the hours that a firm is open for business, fast food
franchise restrictions dictating menu items, employee uniforms, hours of operation, and
the like. Section 261 of the Patent Act expressly permits territorial restrictions in patent
licenses, but in any event the Supreme Court has been applying the rule of reason to purely
vertical nonprice restraints since its decision in Continental TV, Inc. v. GTE Sylvania,
Inc., 433 U.S. 36 (1977). They are rarely found to be unlawful.
The Copyright Act expressly permits many nonprice restrictions, both horizontal and
vertical, by making separate statutory authorizations for the right to reproduce, to prepare
derivative works, to distribute, to perform, and to display, depending on the nature of the
copyrighted good (17 U.S.C. § 106). However, even a purely vertical licensing restriction
that is not expressly authorized by the Copyright Act would probably be legal under the
antitrust laws. Restrictions that attached to a copyrighted article after it is sold might not
be enforceable under copyright law’s statutory first sale doctrine (Kirtsaeng v. John Wiley
& Sons, Inc., 133 S. Ct. 1351 (2013)). These are not antitrust challenges, however.
Interbrand vertical restraints, which include tying and exclusive dealing, have histori-
cally been treated with greater suspicion than intrabrand restraints, mainly because they
can reduce the opportunities of rivals. Under exclusive dealing a firm, typically a retailer
or other intermediary, promises to deal exclusively in the supplier’s product or service.
An ‘output contract’ does the same thing except that it places the exclusivity obligation
on the seller rather than the buyer. In the IP context, the exclusive license is a form of
output contract, under which the IP holder promises to license only one firm for all or
a particular subset of production under the license. The Patent Act expressly authorizes
domestic exclusive patent licenses in 35 U.S.C. § 261. The Copyright Act also authorizes
exclusive licenses, although without an express territorial limitation (17 U.S.C. § 106).
By contrast, § 3 of the Clayton Act makes it unlawful to sell a good or article, ‘whether
patented or unpatented,’ on the condition that the buyer not deal in the goods of a com-
petitor, provided that the agreement’s effect ‘may be to substantially lessen competition or
tend to create a monopoly . . .’ (15 U.S.C. § 14). In addition, anticompetitive exclusionary
contracts are prohibited by § 1 of the Sherman Act, as well as § 2 if the firm imposing
them is a monopolist (United States v. Dentsply Int’l, Inc., 399 F.3d 181 (3d Cir. 2005)).

The Patent Act’s explicit authorization of exclusive licenses exists in some tension
with the Sherman Act’s prohibition of anticompetitive contracts. The tension is partly
reduced by the first sale doctrine (patent exhaustion), which exhausts the patents in any
good once it is sold, thus preventing many instances of post-sale exclusivity enforced by
infringement actions. However, that solution is incomplete because it is still possible to
license or lease patented goods, and the first sale doctrine does not apply unless there is
a sale. The best view is that the antitrust laws qualify the Patent Act in those cases where
competitive harm can be shown. This is consistent with the general rule of statutory
interpretation in this area, which is that the simple statutory authority to do something
is not authority to do so in violation of the antitrust laws. For example, numerous state
and some federal corporation acts authorize corporations to acquire the stock or assets of
other corporations, but that does not mean that they can make an anticompetitive stock or
asset acquisition in violation of § 7 of the Clayton Act, which condemns anticompetitive
mergers. In F.T.C. v. Phoebe Putney Health Sys., Inc., 568 U.S. 216 (2013), the Supreme
Court held that a hospital corporation’s statutory power to acquire other hospitals did not
immunize an anticompetitive acquisition. Already in the 1916 antitrust decision in United
States v. Am. Can Co., 230 F. 859 (D. Md. 1916), the court held that it was unlawful for
the defendant to purchase exclusive rights to patented can making machinery when the
effect was to deny mechanization to rivals. Because American Can made no can making
machinery itself, these agreements were purely vertical. In any event, after a lengthy period
of hostility the courts now accept that exclusivity provisions in sale or license contracts
are only infrequently anticompetitive, depending on such factors as the market share
foreclosed by the agreement, ease of entry, the duration of the contracts and frequency
of rebidding, and offsetting efficiencies that might justify an exclusive deal (Areeda and
Hovenkamp, 2009–15).
A tying arrangement occurs when a firm conditions the sale of one product or service
(the ‘tying’ product) on the purchaser’s taking of a second product or service (the ‘tied’
product). The courts first confronted tying arrangements in patent law cases, long before
the antitrust laws were enacted (Hovenkamp, 2018a). Most ties are contractual, which
means that the tie is imposed by agreement. A few are ‘technological,’ or ‘tech ties,’ which
means that tying is accomplished by a technological design or feature that makes the
seller’s tying product incompatible with the tied products of rivals. Well-known examples
are Microsoft’s ‘commingling’ of the code for the Internet Explorer browser into the
Windows operating system in United States v. Microsoft Corp., 253 F.3d 34 (D.C. Cir.
2001); or Apple’s technological tying of its iTunes content and devices so that they worked
best only when used together, in Apple iPod iTunes Antitrust Litig., 2014 WL 4809288
(N.D. Cal. 2014). Another was Kodak’s introduction of the Instamatic pocket camera,
which was compatible only with its own film cartridges, approved in Berkey Photo, Inc.
v. Eastman Kodak Co., 603 F.2d 263 (2d Cir. 1979). Because product design is a unilateral
act, technological ties are usually treated under antitrust’s monopolization provision, § 2
of the Sherman Act. By contrast, ordinary contractual ties are addressed under § 1 of the
Sherman Act or § 3 of the Clayton Act, both of which require an agreement.
Today legal and scholarly opinion has shifted dramatically from the belief expressed in
the mid-1900s that tying arrangements serve ‘hardly any purpose beyond the suppression
of competition,’ to the view that most ties are competitively benign or beneficial (Standard
Oil Co. of Calif. v. United States, 337 U.S. 293 (1949). The Patent Act expressly addresses

tying arrangements in the 1988 Patent Misuse Reform Act (35 U.S.C. § 271(d)(5)), which
provides that patent ties are not unlawful unless the tying firm has market power in the
tying product. This provision was passed in response to serious excesses in the application
of misuse doctrine, which refused to enforce patent ties in markets where there was no
serious risk of competitive harm. For example, Justice Brandeis’s opinion in Carbice
Corp. of Am. v. Am. Patents Dev. Corp., 283 U.S. 27 (1931) found patent misuse when
the manufacturer of a patented ice box required purchasers of the box to use its own
unpatentable dry ice. The market for the unmechanized boxes was competitive and dry
ice was a common commodity produced in ‘carbonic’ plants in many places.
Proving unlawful tying under the antitrust laws requires proof of tying product market
power. In Illinois Tool Works Inc. v. Indep. Ink, Inc., 547 U.S. 28 (2006), the Supreme
Court held that this power may not be inferred simply from the existence of a patent
covering the tying product. Beyond that, the analysis depends in part on whether the
tying and tied products are used in fixed or variable proportions. The tie of Microsoft’s
Windows computer operating system and its Internet Explorer browser is a fixed propor-
tion tie. A buyer purchases one copy of each, although she might use them in different
proportions. The most common anticompetitive rationale for such a tie is exclusion of
rivals in the tied product market, which requires that the firm employing the tie have
dominance in the tying market. For example, by tying Windows and Internet Explorer
(initially by contract and later by technological integration of the code), Microsoft was
able to dry up the market for independent web browser Netscape. However, fixed propor-
tion ties can also be beneficial when joint production or distribution reduces costs or
makes a product work better. Many fixed proportion ties are a consequence of nothing
more than technological improvement, which often proceeds by making products more
integrated (Hovenkamp, 2018a). For example, IBM’s desktop computers were attacked as
tying arrangements to the extent that they merged processors, motherboards, controllers,
storage devices, and memory into a single box. Previously these products had been sold
separately and connected by cables, making it possible for one person to sell the processing
box and another the disc drive, and so on. But the new systems were more reliable, faster,
and cheaper (Fisher, 1983). In sum, overly aggressive use of the law governing technologi-
cal tying can serve to limit innovation. Thus, the decision in California Computer Prod.,
Inc. v. IBM Corp., 613 F.2d 727 (9th Cir. 1979) rejected the claim that innovative product
integration resulting in a technological tie was simple ‘design manipulation’ intended to
ruin competing disc drive manufacturers.
Variable proportion ties can serve all of the same functions as fixed proportion ties. In
addition, however, variable proportioning can operate as metering or price discrimination
devices. These are usually either competitively harmless or beneficial to the extent that
they increase overall output. For example, the manufacturer of a printer that uses ink
cartridges might tie the printer and the cartridges, either by contract or technological
design. It then drops the price of the printer, often substantially and sometimes even
to zero, but charges a premium for the ink cartridges. The result of this practice is that
the firm obtains a higher overall return from high intensity users who consume more
ink. This is a form of second degree price discrimination which results in larger printer
sales to the extent the printer price is lower, but also captures more revenue from higher
intensity users. In most cases consumers as a group are better off, although high intensity
consumers may be injured (Hovenkamp and Hovenkamp, 2015). While such ties may

reduce welfare if the firm has an absolute monopoly in the tying product and the tie
reduces output (Elhauge and Nalebuff, 2017), they clearly increase welfare in cases where
the seller is not a monopolist or in collusion with the other sellers in the market. In such
cases any purchaser required to pay a monopoly price will purchase from a different seller
(Hovenkamp, 2018a). In virtually all litigated variable proportion tie cases the seller is
not a monopolist, and often it is not even among the largest firms in the market. For
example, in Impression Prod., Inc. v. Lexmark Int’l, Inc., 137 S. Ct. 1523 (2017), a vari-
able proportion tying case that the Supreme Court decided under the patent exhaustion
doctrine, Lexmark’s share of the highly competitive computer printer market was less
than 5 percent (Hovenkamp, 2018b).
Variable proportion ties are also commonly used by franchisors, such as those in the
fast food industry, where the tying product is commonly trademarked or occasionally
copyrighted rather than patented. For example, the franchisor might charge a very low
price or even zero for the right to a franchise, but then place an overcharge on various
food products or supplies that the franchisee uses in variable proportions. As a result, the
franchisor makes more money from franchisees that sell more (Queen City Pizza Inc. v.
Domino’s Pizza Inc., 124 F.3d 430 (3d Cir. 1997)). Importantly, the profitability of these
variable proportion ties does not depend on the seller’s market power, but only on suffi-
cient product differentiation or branding to make its tying product attractive. Many of the
defendants in franchise tying claims have been relatively minor firms. For example, Siegel
v. Chicken Delight, Inc., 448 F.2d 43 (9th Cir. 1971) found antitrust liability for a variable
proportion franchise by a struggling, minor fast food franchisor. The market power
requirement was thought to be met because the defendant’s name was trademarked. As
noted earlier, that legal conclusion has now been overruled.
D. Patent Pools
In a patent pool, several patent holders license their patents into a common ‘pool,’ or
organization that relicenses them out to members. The pool may also license to outside
manufacturers who do not own relevant patents of their own. While patent pools have
existed for a long time, in recent decades their growth has been exponential, thanks largely
to the rapid development of networked products sold by multiple firms. The products
include telecommunications and other digital devices as well as computer and video
technology.
A patent pool is a ‘horizontal’ restraint to the extent that its participants are competi-
tors in the product market or license competing technologies. The relationship among the
parties is often more complex, however. First of all, one important value of patent pooling
is to facilitate the coordinated use of complementary technologies, but firms that produce
pure complements are not competitors. Another use, equally or more important in some
markets, is to minimize the transaction costs of patent management. For such savings to
occur the pool members can have any relationship, from competitor, to vertically related,
or complementary.
Historical rationales for pooling focused on the differences between complements and
substitutes (Gilbert, 2004; Santore, 2010). Pooling or cross-licensing among different
producers in the same industry was thought to be procompetitive if the shared patents
were complements with one another, but more likely to be anticompetitive if they were

substitutes (Shapiro, 2001). If a patent is a discrete unit of innovation whose validity and
boundaries are easily assessed, this distinction makes sense. Complements are ordinarily
used together, meaning that joining them can create social benefits. For example, if one
firm owns a patent on a desirable lawn mower motor and another owns a patent on a
desirable throttle, pooling will enable the two firms to produce mowers that have both
improvements. By contrast, substitutes are used in the alternative rather than together.
This makes price-fixing a more likely explanation.
The distinction between complements and substitutes is less important, however, in a
world where patents are numerous and complex, with many claims, costly to interpret,
and of uncertain quality (Lerner and Tirole, 2004). In such a setting, the economics of
transaction costs and boundary enforcement dominates other explanations. That is to
say, the modern patent pool in information technologies is a type of commons for which
sharing with management rules is cheaper and more effective than individual appropria-
tion and enforcement.
Patent pools for complex information technologies can contain thousands of patents.
For example, the MPEG-LA pool for digital video technology controls about 5,000.5
Subsequent to issuance, most of these patents have never been assessed for validity and
scope. Even getting a legal opinion on these issues can be very costly. Many of the patents
have numerous claims, often making the relationship among different patents far more
complex than simple substitutes or complements. Further, the substitute/complement
question often cannot be answered until one examines the device that intends to use them.
For example, the licensees of the MPEG-LA pool include multiple competing makers
of digital cameras and phones, but also producers of complementary products such as
flash memories, video displays or projectors, film or photo editing software, and the like.
Whether two patents are substitutes, complements, or simply not practiced at all can
depend on the type of device that the manufacturer produces.
One illustration of these complex relationships is the Princo litigation in the Federal
Circuit, which involved shared technology for developing rewritable digital discs (Princo
Corp. v. Int’l Trade Comm’n, 616 F.3d 1338 (Fed. Cir. 2010)). A feature that located the
stylus on a rewritable disc came in both patented analog and digital versions. The two
versions were substitutes in the product market, because a manufacturer would use one of
them but not both. However, at least one claim of the analog version wrote on the digital
patent, making the two patents legal complements as well. Regardless of their function,
two patents are complements if someone cannot practice one without infringing the
other (Princo Corp. v. Int’l Trade Comm’n, 616 F.3d 1338 (Fed. Cir. 2010); Bohannan,
2011, pp. 510–11). Indeed, not until after costly claim construction would we know how
the patents relate to one another. Even then, claim construction reversal rates are unac-
ceptably high (Teva Pharmaceuticals USA, Inc. v. Sandoz, Inc., 2014 WL 2885379 (U.S.)
(U.S., 2014)) (amicus brief).
Finally, the relationship between substitutes and complements is relevant to one
additional factor: in markets where multiple technological alternatives exist, the pooling
of substitutes may serve to deny access to outsiders. For example, if there are three
known technological alternatives for solving a particular problem a pool that acquires
5
See http://www.mpegla.com/main/Pages/AboutHistory.aspx.

the patents to all three may be in a position to exclude non-pool members from the
product market.
A more robust explanation for pooling in high tech information markets comes from
the theory of commons development (Bohannan and Hovenkamp, 2012). As Ronald
Coase argued in The Nature of the Firm (Coase, 1937), the boundaries of a firm are
determined by its costs of doing business. A firm makes for itself or buys from others,
depending on which alternative is more profitable. When it makes something for itself its
boundaries expand, and when it buys from others they contract (Hovenkamp, 2011a).
Property rights, including IPRs, provide the legal power to exclude, but they do not
always make individual exclusion the best choice. The traditional property world is full
of instances, ranging from shared driveways and party walls to community owned tennis
courts and swimming pools, where individuals share because sharing is cheaper or more
productive than exclusion. The same thing has been true for many centuries of so-called
common pool resources such as fisheries, irrigation rights, or grazing rights. One hundred
fishermen who own a large lake could if they wished build fences cutting the lake into
100 pieces. But subdividing would be costly, would very likely produce many disputes,
and would be devastating to the yield of a mobile species of fish (Ostrom, 1990). Patent
rights in information technologies are similar. While firms are certainly entitled to own
patents individually, defining and defending boundaries might be much less productive
than sharing them reciprocally with others.
While patent pools created to limit boundary problems have some similarities to
common pool resources, they are not identical and some differences are critical to com-
petitive impact. Most traditional common pool resources are ‘rivalrous,’ which means that
each unit taken depletes what is left. Uncontrolled sharing leads to overuse, as shown by
the fate of the American Bison as well as chronic overfishing in many fishing commons.
For this reason, catch or harvest limitations are essential if the commons is to operate
efficiently. Indeed, a principal difficulty in commons management is making and enforc-
ing access or harvest rules (Ostrom, 1990). By contrast, IPRs are generally nonrivalrous.
An IP right such as a patent can be used many times without depleting what is left over.
As a result, output restraints in the product market are more suspicious in IP commons
than in traditional commons (Bohannan and Hovenkamp, 2012).
Another difference between modern information technology patent pools and tradi-
tional commons is the diversity of the participants. All those sharing rights on a common
fishery or pasture are likely to be fishermen and ranchers with relatively undifferentiated
businesses. By contrast, given the nature of multi-claim patents, those who practice them
might be highly diverse, producing complements or unrelated technologies as much as
they produce substitutes. That is clearly true of the previously described MPEG-LA pool.
A complex device such as a digital video camera might practice many patents in the pool,
while a simple device such as a memory chip or photo editing software employs only a few.
These differences have led to claims akin to anticompetitive tying. A firm that is
required to license all of a pool’s package of 5,000 patents may complain that it really uses
only 200 of the patents. In this case, the patents it uses are the ‘tying product,’ while the
unwanted patents are the ‘tied product.’ These antitrust claims nearly always fail, for the
simple reason that competition is not injured. Forcing someone to take a bigger product
than he might want—such as 100 acres of land instead of a single residential lot—might
present a bargaining problem. It is not anticompetitive, however, because no one is being

excluded. Further, the cost of determining which of the pool’s patents the complaining
licensee actually practiced could be much greater than the cost of the license fee itself. As
one court observed in rejecting such an antitrust claim, ‘[I]t is not anticompetitive for a
patent pool to include numerous potentially blocking patents, patents which may or may
not be essential but which are more efficient to license as part of the pool than to risk
the expense of future litigation’ (Nero AG v. MPEG LA, LLC, 2010 WL 4366448 (C.D.
Cal. 2010)).
E. Exclusionary Practices
1. Walker process and unreasonable infringement claims

By their nature IPRs create a power to exclude. Asserting that right is protected by §
271(d)(3) of the Patent Act as well as constitutional protections of the right of access
to the courts (Areeda and Hovenkamp, 2009–15). No protection exists, however, for
‘baseless’ claims having no foundation in fact or law. The Supreme Court’s decision
in Walker Process Equip., Inc. v. Food Mach. and Chem. Corp., 382 U.S. 172 (1965)
recognized that improper patent infringement actions could violate § 2 of the Sherman
Act if they excluded another firm unreasonably. As noted previously, antitrust has little
or no role in policing pre-issuance patent applicant conduct. Many Walker Process cases,
including Walker Process itself, involve some pre-issuance conduct. In Walker Process,
the patent applicant failed to inform the patent examiner about sales made prior to
patent application that, if disclosed, would have barred the patent. See 35 U.S.C. § 102(a)
(1). Other cases have involved other forms of inequitable conduct before the USPTO,
including failures to inform the examiner of known prior art that would have defeated
patentability (Nobelpharma AB v. Implant Innovations, Inc., 129 F.3d 1463 (Fed. Cir.
1997)). Importantly, however, merely obtaining a patent improperly is not an antitrust
violation. Further the patent system has numerous remedies for improper conduct before
the USPTO. Walker Process itself involves the post-issuance practice of filing an infringe-
ment suit, or engaging in other enforcement activity such as a threat to sue, on a patent
that the owner should reasonably have known to be invalid or unenforceable under the
circumstances.
Antitrust liability for Walker Process violations requires not only the improper infringe-
ment action, but also a structural basis for thinking that the lawsuit either perpetuates or
threatens to create a market monopoly. This makes the antitrust remedy much less readily
available than the ‘exceptional case’ remedy in § 285 of the Patent Act, which authorizes
a judge to award attorney’s fees against a patentee who makes an improper claim or
commits other litigation misconduct. The Supreme Court’s decision in Octane Fitness,
LLC v. ICON Health & Fitness, Inc., 134 S. Ct. 1749 (2014) made such awards easier to
obtain. At this writing, it remains to be seen if this decision will have any impact on the
availability of antitrust remedies. One 2014 decision did find some basis for comparison
(Tyco Healthcare Group LP v. Mutual Pharma. Co., Inc., 762 F.3d 1338 (Fed. Cir. 2014)).
As formulated, Walker Process reaches no more than a small portion of socially harm-
ful infringement actions. It applies in cases where questions of validity or infringement
are relatively clear and, given that, the patent holder has sued unreasonably. A far larger
number of harmful infringement actions arise out of problems of ‘notice failure’ (Menell
and Scotchmer, 2007; Hovenkamp, 2011b). Because patent infringement does not require

advance notice of someone else’s patents, many innocent innovators are later caught by
surprise simply because the cost of searching patents or interpreting their claims is so high
and the results so unreliable. These are features of the patent system itself, however, and
generally cannot be remedied by the antitrust laws.
2. Acquisitions
Section 261 of the Patent Act expressly authorizes patentees to assign their patents to
others, although it does not authorize anticompetitive assignments. Even monopolists
should be entitled to acquire technology that they need to improve their own products
or processes. Nevertheless, a few qualifications are important. First, to improve its own
technology a dominant firm requires no more than a nonexclusive license. Second,
acquiring a patent in order to improve one’s own technology necessitates practicing the
patent; acquisition and nonuse does nothing to improve the acquirer’s own product,
but may serve to exclude rivals from access to alternative technologies. Third, under
some circumstances the acquisition of exclusive rights in competing patents can be
a significant threat to competition, particularly when the patents represent all of the
best alternatives for operating in some market. Suppose, for example, that patent
portfolios Alpha, Beta, and Gamma represent the only three commercially feasible
ways of providing a particular service. If a dominant firm acquired exclusive rights in
all three it would be able to exclude competitors from the market even though it would
likely be practicing only one of the three alternatives. Such acquisitions of competing
portfolios should be challengeable as unlawful mergers under § 7 of the Clayton Act,
which expressly applies to asset as well as stock acquisitions. The courts have repeatedly
held that a patent is a qualifying ‘asset,’ provided that anticompetitive effects are shown
(Areeda and Hovenkamp. 2009–15). If market dominance results, the acquisitions
could also be unlawful monopolization under § 2 of the Sherman Act, at least if followed
by an infringement action. For example, in Intellectual Ventures I, LLC v. Capital One
Financial Corp., 99 F. Supp. 3d 610 (D. Md. 2015), the court held that a non-practicing
entity’s acquisition of some 3,500 patents covering an entire segment of the banking
industry and subsequent enforcement could amount to unlawful monopolization
(Hovenkamp and Hovenkamp, 2017; Areeda and Hovenkamp, 2009–15). If the firm
has acquired only one or two patents, however, the courts have been willing to sustain
infringement actions even if one or more of the patents are unused, as the Supreme
Court held in Cont’l Paper Bag Co. v. E. Paper Bag Co., 210 U.S. 405 (1908), and the
Federal Circuit much more recently confirmed in Trebro Mfr., Inc. v. Firefly Equip.,
LLC, 748 F.3d 1159 (Fed. Cir. 2014). Neither case raised an antitrust issue (Hovenkamp
and Cotter, 2015).
3. Refusal to license

Neither United States antitrust law nor the Patent Act recognizes a general duty to
license a patent or to share patented technology. Indeed, § 271(d)(4) of the Patent Act
provides that a refusal to license is not ‘misuse or illegal extension of the patent right. . . .’
This language appears to refer to unilateral refusals to deal, but some language in a
Federal Circuit dissent in Princo Corp. v. Int’l Trade Comm’n, 616 F.3d 1338 (Fed. Cir.
2010), suggested that it could reach concerted refusals to deal as well (Bohannan and
Hovenkamp, 2011). That rule, if accepted, would be unfortunate, for concerted refusals

have significantly more anticompetitive potential than unilateral refusals (Areeda and
Hovenkamp, 2009–15).
In its decision in Image Tech. Servs., Inc. v. Eastman Kodak Co., 125 F.3d 1195 (9th Cir.
1997), the Ninth Circuit held that the antitrust laws did give Kodak a duty to sell patented
parts to independent repair persons who wished to fix its high-speed photocopiers. The
decision has been widely criticized and was expressly rejected by the Federal Circuit in the
ISO Antitrust Litigation, 203 F.3d 1322 (Fed. Cir. 2000), a similar case involving Xerox.
As of this writing the Supreme Court has not resolved the split (Areeda and Hovenkamp,
2009–15).
Currently the courts have not settled on the scope of a patentee’s duty to license a patent
encumbered by a FRAND commitment, or a promise made as part of a standard setting
process to license on ‘fair reasonable, and nondiscriminatory terms.’ Patentees generally
make these commitments as a condition of having their patents declared ‘standards essen-
tial,’ or part of the technology to be adopted by a network. In Apple, Inc. v. Motorola,
Inc., 757 F.3d 1286 (Fed. Cir. 2014), a panel of the Federal Circuit split three ways on
the question. Nonetheless, it seems clear that if the owner of a FRAND-encumbered
patent has executed a contractual promise to license to all on FRAND terms, then an
infringement suit in conflict with that contractual obligation would be improper, as the
court found in Microsoft Corp. v. Motorola, Inc., 795 F.3d 1024 (9th Cir. 2015). It could
occasion antitrust liability, although antitrust would not necessarily be the best vehicle
for assessing it because the challenger must also establish market power and the creation
of monopoly. The Patent Act’s own remedial measures, such as the previously discussed
‘exceptional case’ provision, would provide a more direct route to relief, although not
treble damages.
One must also distinguish between simple and conditional refusals to grant an IP
license or sell a good covered by an IP right. Section 3 of the Clayton Act, discussed
above, expressly limits certain conditional licensing contracts that threaten competition,
and expressly covers goods that are patented as well as unpatented. For example, a tying
arrangement can readily be construed as a refusal to sell or license a tying product unless
the buyer also takes the tied product. The legality of a conditional refusal depends on the
legality of the underlying condition.
Another important distinction is between compulsory licensing as a general antitrust
duty and compulsory licensing as a remedy for some other antitrust violation. In many
antitrust cases the courts have determined that compulsory licensing is the best way to
undo the effects of some other violation. Merger remedies often require IP licensing as a
condition for approval and many structural breakups, such as the one that dismantled the
nationwide AT&T telephone network, could not have succeeded unless each of the spun
off companies received nonexclusive licenses to the parent’s intellectual property portfolio
(United States v. AT&T, 552 F. Supp. 131 (D.D.C. 1982)).
Finally, it is worth noting that the question of compulsory licensing duties is an impor-
tant and controversial one, raising serious questions in market dominating networks,
lifesaving pharmaceuticals, national defense, and other areas involving questions about
patent social value. Antitrust is a poor vehicle for resolving most of these issues. It offers
no useful tools for determining when such a dealing obligation is socially necessary. Nor
does it have any mechanisms for setting a price, other than through ordinary damages
measurement tools.

IV. IP LAW’S OWN INTERNAL RULES FOR FACILITATING

COMPETITION
Intellectual property law also has internal rules for facilitating competition in the scope or
use of IPRs (Cotter, 2006). Most of these rules are quite different from antitrust rules. IP
law’s internal competition rules do not require a market definition, proof of anticompeti-
tive market exclusion, or even an attempt to measure the impact of a particular practice
on price or output. Rather they are best seen as rules that limit IP ‘overreaching’ of
various kinds.
A. The First Sale (Exhaustion) Doctrine
Patent and copyright law’s ‘first sale,’ or exhaustion rule, limits the power of a right
holder to restrict the use or resale of a protected good once it has been sold. Both the
patent and copyright exhaustion rules were originally judge-made, but subsequently
Congress enacted the copyright first sale doctrine into the statute, at 17 U.S.C. §109(a).
In its simplest form the exhaustion rule states that the sale, not the license or lease, of a
patented or copyrighted good exhausts the right holder’s legal interest in that particular
copy. The exhaustion doctrine has been used to strike down resale price restrictions,
under both copyright (Bobbs-Merrill Co. v. Straus, 210 U.S. 339 (1908)); and patent
(United States v. Univis Lens Co., 316 U.S. 241 (1942)). It has been applied to quasi-
exclusive dealing restrictions in patents (Quanta Computer, Inc. v. LG Elecs., Inc., 553
U.S. 617 (2008)), and has also been used to strike down limitations on where purchased
goods can be sold, in both copyright (Kirtsaeng v. John Wiley & Sons, Inc., 133 S. Ct.
1351 (2013)) and patents (Adams v. Burke, 84 U.S. 453 (1873)); Hovenkamp, 2011c). In
2017 the Supreme Court held that a patentee could not avoid exhaustion by making its
sale ‘conditional’ on certain post-sale restrictions, no matter how clearly the condition
was stated (Impression Products, Inc. v. Lexmark, Inc., 137 S. Ct. 1523 (2017)). The
Court located the policy rationale for this concern in the common law rules limiting
restraints on alienation.
Even though the first sale rule reaches many of the same practices that antitrust law
reaches, such as tying, exclusive dealing, or RPM, it cannot be counted as an antitrust
provision. Its application is completely indifferent to competitive effects, requiring only
that there be a qualifying ‘sale.’ Once such a sale has been found6 the post-sale restraint
becomes unenforceable per se, without regard to competitive effects. One view of the
doctrine is that it is intended to limit the reach of patent law in order to leave space for
other bodies of commercial law (Duffy and Hynes, 2016; Hovenkamp, 2011c). Another
view is that it is a creature of federalism, limiting the reach of federal IP supremacy in
order to ensure that states retain the power to control post-sale restraints on protected
goods (Hovenkamp, 2016b).
6
Bowman v. Monsanto Co., 133 S. Ct. 1761 (2013) held that exhaustion did not apply to self-
replicating seed where the replanted seed was a subsequent generation and not the selfsame seed
that the infringement defendant had purchased.

B. ‘Misuse’
The misuse doctrine originated in patent law and was later expanded to copyright. It
is entirely judge-made, and the only references to it in the Patent Act are limitations
contained in the Patent Misuse Reform Act, 35 U.S.C. § 271(d), rather than recognition.
Almost from the beginning misuse doctrine got off to a troublesome start. The Supreme
Court’s decision in the Motion Picture Patents Co. v. Universal Film Co., 243 U.S. 502
(1917) case, where a version of the doctrine was first recognized, almost surely involved
an anticompetitive restraint. The owner of a market dominant projector sold it subject to
a license restriction permitting projection only of its own films. But the Clayton Act had
already been passed, and in 15 U.S.C. § 14 it expressly prohibited patent ties that injured
competition. To be sure, the competition issue was raised as a defense to an infringement
action rather than an antitrust claim, but by this time the Supreme Court had already held
in Cont’l Wall Paper Co. v. Louis Voight and Sons Co., 212 U.S. 227 (1909) that illegality
under the antitrust laws was a complete defense to an enforcement action, such as a breach
of contract suit. Therefore, a patent-law-generated misuse doctrine was unnecessary.
In subsequent decades, misuse doctrine was applied aggressively to patent tying,
exclusive dealing, royalty extensions, and other licensing arrangements where injury to
competition was nowhere in view. It is no wonder that during a later era of antitrust con-
traction both courts and commentators severely criticized ‘misuse’ unless its application
was limited to conduct that violated the antitrust laws (USM Corp. v. SPS Tech., Inc., 694
F.2d 505 (7th Cir. 1982); Bohannan, 2011; Lim, 2013).
The courts have also recognized copyright ‘misuse’ where the copyright infringement
plaintiff had attempted to ‘sequester’ uncopyrightable public data in a copyrighted
database program. Judge Posner made clear that the violation in question was not of
the antitrust laws. It was a unilateral act and there was no finding that the infringement
plaintiff had sufficient market power (Assessment Tech. of Wisconsin., LLC v. Wiredata,
Inc., 350 F.3d 640, 646 (7th Cir. 2003)).7 Rather, the concern was strictly one of copyright
law, which both protects copyrighted content but also seeks to ensure that material in the
public domain can be accessed.
In 2015, the Supreme Court breathed new life into one particular variation of the
misuse doctrine. In Kimble v. Marvel Entm’t Co., LLC, 135 S. Ct. 2401 (2015), it adhered
to the much-criticized rule adopted by the Supreme Court in Brulotte v. Thys Co., 379
U.S. 29 (1964), that a license agreement basing royalties on post-expiration patent use is
unenforceable per se. Kimble did not attempt to justify the original Brulotte rule, but held
only that reliance plus stare decisis required adherence (Hovenkamp, 2015b).
The overuse and subsequent undermining of ‘misuse’ doctrine was an opportunity lost.
Courts have legitimate concerns about IP overreaching as a matter of intellectual property
law. The 70-year-long escapade, stretching from the 1917 Motion Picture Patents case
7
The court also relied on the Fourth Circuit’s decision in Lasercomb Am., Inc. v. Reynolds,
911 F.2d 970 (4th Cir. 1990), which found misuse in a restriction on a copyright licensee’s develop-
ment of new technology, even though that restraint did not violate the antitrust laws. The concur-
ring opinion in Omega S.A. v. Costco Wholesale Corp., 776 F.3d 692 (9th Cir. 2015) argued that a
copyright holder committed misuse by trying to leverage copyrighted design into prohibition on
product sale in the United States.

to the 1988 Patent Misuse Reform Act, represented gross overreaching in tying doctrine
under both misuse and antitrust, including the doctrine developed in Morton Salt Co.
v. G.S. Suppiger Co., 314 U.S. 488 (1942), that someone who was misusing its patent as
against one firm could not sue anyone else, even an admitted infringer. In U.S. Gypsum
Co. v. National Gypsum Co., 352 U.S. 457 (1957), the Supreme Court globalized that
remedy, holding that once misuse was found the patent would be unenforceable against
the entire world until the misuse was ‘purged.’
A revitalized doctrine of patent misuse should do two things. First, it should identify
conduct that is socially harmful because it threatens the balance between exclusion and
access created by the intellectual property laws themselves. Second, the remedy should
ordinarily be limited to an injunction against the practice, or recognition that violation
of the practice is not an act of infringement (Bohannan, 2011).
C. Competition-Based Limitations on ‘Functionality’ Protection
One important limiting principle of IP law is that protection of a ‘function’ must come
by means of a utility patent, which is more difficult to obtain than other IPRs and has
a shorter duration than copyrights or trademarks. For this purpose, design patents are
treated more like copyrights and trademarks rather than patents.
The fundamental principle undergirding functionality limitations is that people should
not be able to use IPRs to create product monopolies any more than is justified by the
limitations inherent in utility patents. An important corollary is that people should not be
able to turn other IPRs into quasi-patent rights in order to broaden their scope or dura-
tion. The grandparent of the doctrine is Baker v. Selden, 101 U.S. 99 (1879), which held
that a copyrighted book teaching the author’s method of accounting served to protect
against making unauthorized copies of the book, but not against using or teaching the
method itself. That decision was followed in Bikram’s Yoga College of India, L.P. v.
Evolation Yoga, LLC, 803 F.3d 1032 (9th Cir. 2015), holding that a book illustrating yoga
poses and movements did not prohibit someone from teaching those motions in a class. In
the area of trademark or trade dress, the Supreme Court’s important decision in TrafFix
Devices, Inc. v. Market Displays, Inc., 532 U.S. 23 (2001) held that a traffic sign support
that was an essential element of an expired utility patent could not be grandparented
into a trademark, effectively extending the right over this functional design indefinitely
(Rosetta Stone, Ltd. v. Google, Inc., 676 F.3d 144 (4th Cir. 2012); 1-800-CONTACTS,
Inc. v. WhenU.Com, Inc., 414 F.3d 400 (2d Cir. 2005)).8 Similarly, copyright protection
does not extend to purely functional names or commands in computer programs if good
alternatives are lacking and the result serves to limit the ability of rival programmers to
replicate that function (Lotus Dev. Corp. v. Borland Int’l, Inc., 49 F.3d 807 (1st Cir. 1995);
Oracle Am., Inc. v. Google Inc., 750 F.3d 1339 (Fed. Cir. 2014)).
Several lower courts have also held that design patents cannot be expanded so as to
perform a utility function, thus excluding the products of rivals. For example, in Chrysler
8
The so-called ‘trademark use’ doctrine in its most extreme form would make it unlawful
under the Lanham Act to use the trademarked name of a competitor for purposes such as com-
parative advertising.

Motors Corp. v Auto Body Panels of Ohio, Inc., 908 F.2d 951 (Fed. Cir. 1990), and Best
Lock Corp. v. Ilco Unican Corp., 94 F.3d 1563 (Fed. Cir. 1996), the Federal Circuit held
that the owner of design patents could not use a design feature to exclude complementary
products. In the first, it rejected Chrysler’s attempt to enforce a design patent on an auto-
mobile bumper mount so as to exclude competing bumper makers who could not make
a Chrysler-compatible bumper without infringing the design patent (Chrysler Motors
Corp. v Auto Body Panels of Ohio, Inc., 908 F.2d 951 (Fed. Cir. 1990)). In the second it
prevented the owner of a design patent on a door key from enforcing a patent so as to
make competitors’ keys incompatible with its locks (Best Lock Corp. v. Ilco Unican Corp.,
94 F.3d 1563 (Fed. Cir. 1996)). In both cases the design patentees were attempting to create
‘technological ties.’ No determination of the patentee’s market power was made, and the
ties may or may not have been antitrust violations. But that is beside the point. The issue
was sequestration, not monopoly. Utility patents may permit firms to keep rivals out of
the product market, but design patents should not.
REFERENCES
Aghion, Philippe, and Peter Howitt. 1998. Endogenous Growth Theory. Cambridge, MA: MIT Press.
Aghion, Philippe, and Peter Howitt. 2007. ‘Capital, Innovation, and Growth Accounting,’ 23 Oxford Review of
Economic Policy 79–93.
Aghion, Philippe, Nick Bloom, Richard Blundell, Rachel Griffith, and Peter Howitt. 2005. ‘Competition and
Innovation: An Inverted-U Relationship,’ 120 Quarterly Journal of Economics 701–28.
Anderson, J. Jonas, and Peter S. Menell. 2013. ‘Informal Deference: A Historical, Empirical, and Normative
Analysis of Patent Claim Construction,’ 108 Northwestern Univ. Law Review 1–79.
Arai, Koki. 2013. ‘Patents, Competition Policy and Growth,’ 18 Journal of Technology Law and Policy 83–97.
Areeda, Phillip E., and Herbert Hovenkamp. 2009–15. Antitrust Law. 20 vols. New York, NY: Aspen Pub.
Arrow, Kenneth. 1962. ‘Economic Welfare and the Allocation of Resources for Innovations,’ in Richard Nelson,
ed., The Rate and Direction of Inventive Activity. Princeton, N.J.: Princeton University Press.
Belleflamme, Paul. 2006. ‘Patents and Incentives to Innovate: Some Theoretical and Empirical Economic
Evidence,’ 13 Ethical Perspective: Journal of the European Ethics Network 267–88.
Boldrin, Michele, and David Levine. 2008. Against Intellectual Monopoly. New York, NY: Cambridge
University Press.
Bohannan, Christina. 2006. ‘Reclaiming Copyright,’ 23 Cardozo Arts and Entertainment Law Journal 567–634.
Bohannan, Christina. 2011. ‘IP Misuse as Foreclosure,’ 96 Iowa Law Review 510–11.
Bohannan, Christina, and Herbert Hovenkamp. 2011. ‘Concerted Refusals to License Intellectual Property
Rights,’ 1 Harvard Business L. Rev. 21–5.
Bohannan, Christina, and Herbert Hovenkamp. 2012. Creation Without Restraint. New York, NY: Oxford
University Press.
Bowman, Ward S. 1973. Patents and Antitrust Law. Chicago, IL: University of Chicago Press.
Brief of the ABA as Amicus Curiae. 2014. Teva Pharmaceuticals USA, Inc. v. Sandoz, Inc., 2014 WL 2885379.
Chang, Xin., R. David McLean, Bohui Zhang, and Wenrui Zhang. 2014. ‘Patents and Productivity Growth:
Evidence from Global Patent Awards,’ 26 SSRN Working Paper, accessed January, 2013 at http://papers.ssrn.
com/sol3/papers.cfm?abstract_id=2371600.
Coase, Ronald H. 1937. ‘The Nature of the Firm,’ 4 Economica (New Series), 386–405.
Connor, John M. 2014. ‘Price-Fixing Overcharges,’ Working Paper, accessed March 6, 2019 at http://papers.ssrn.
com/sol3/papers.cfm?abstract_id=2400780.
Contreras, Jorge L. 2015. ‘A Brief History of FRAND,’ 80 Antitrust L.J. 39–114.
Cotter, Thomas F. 2006. ‘The Procompetitive Interest in Intellectual Property Law,’ 48 William and Mary Law
Review 483–557.
Cotter, Thomas F. 2015. ‘Innovation and Antitrust Policy,’ in Roger D. Blair and D. Daniel Sokol, eds., 2 The
Oxford Handbook of International Antitrust Economics. NY: Oxford University Press.
Crouch, Dennis. 2010. ‘An Empirical Study of the Role of the Written Description Requirement in Patent
Examination,’ 104 Northwestern University Law Review 1665–79.

Dreyfuss, Rochelle. 1989. ‘The Federal Circuit: A Case Study in Specialized Courts,’ 64 NYU Law Review 1–76.
Dreyfuss, Rochelle. 2008. ‘In Search of Institutional Identity: The Federal Circuit Comes of Age,’ 23 Berkeley
Tech. L.J. 787–828.
Duffy, John F. 2007. ‘Inventing Invention: A Case Study of Legal Innovation,’ 86 Texas Law Review 1–72.
Duffy, John F., and Richard M. Hynes. 2016. ‘Statutory Domain and the Commercial Law of Intellectual
Property,’ 102 Va. L. Rev. 1–77.
Edlin, Aaron S., Scott C. Hemphill, Herbert Hovenkamp, and Carl Shapiro. 2013. ‘Activating Actavis,’ 28
Antitrust 16–23.
Edlin, Aaron S., Scott C. Hemphill, Herbert Hovenkamp, and Carl Shapiro. 2014. ‘Actavis and Error Costs,’
2014 Antitrust Source 1–8.
Edlin, Aaron S., Scott C. Hemphill, Herbert Hovenkamp, and Carl Shapiro. 2015. ‘The Actavis Inference:
Theory and Practice,’ 67 Rutgers Univ. L. Rev. 585–635.
Elhauge, Einer, and Barry J. Nalebuff. 2017. ‘The Welfare Effects of Metering Ties,’ (2017), Harvard Public
Law Working Paper, No. 16-20, accessed March 6, 2019 at https://papers.ssrn.com/sol3/papers.cfm?abstr
act_id=2591577.
Epstein, Richard A. 2010. ‘The Disintegration of Intellectual Property? A Classical Liberal Response to a
Falvey, Rod, Neil Foster, and David Greenaway. 2006. ‘Intellectual Property Rights and Economic Growth,’ 10
Review of Development Economies 700–719.
Fierstein, Ronald K. 2015. A Triumph of Genius: Edwin Land, Polaroid, and the Kodak Patent War. Washington
D.C.: ABA Press.
Fisher, Franklin M. 1983. Folded, Spindled and Mutilated: Economic Analysis and United States Versus I.B.M.
Law Review 3–33.
Gilbert, Richard J., and Carl Shapiro. 1990. ‘Optimal Patent Length and Breadth,’ 21 Rand J. Econ. 106–12.
Gould, David M., and William C. Gruben. 1996. ‘The Role of Intellectual Property Rights in Economic
Growth,’ 48 Journal of Development Economics 328–38.
Grossman, Gene M., and Elhanan Helpman. 1994. ‘Endogenous Innovation in the Theory of Growth,’ 8
Hashmi, Aamir Rafique. 2011. ‘Competition and Innovation: The Inverted-U Relationship Revisited,’ accessed
March 31, 2019 at https://papers.ssrn.com/sol3/papers.cfm?abstract_id=1762388.
Helpman, Elhanan. 2004. The Mystery of Economic Growth. Cambridge, MA: Harvard Univ. Press.
Hovenkamp, Erik N. 2018. ‘Tying, Exclusivity, and Standard Essential Patents,’ 19 Columbia Science &
Hovenkamp, Erik N., and Thomas F. Cotter. 2015. ‘Anticompetitive Patent Injunctions,’ 100 Minn. L. Rev. 871–920.
Hovenkamp, Erik N., and Herbert Hovenkamp. 2015. ‘Tying Arrangements,’ in Roger D. Blair and D. Daniel Sokol,
eds., The Oxford Handbook of International Antitrust Economics. New York, NY: Oxford University Press.
Hovenkamp, Erik, and Herbert Hovenkamp. 2017. ‘Buying Monopoly: Antitrust Limits on Damages for
Externally Acquired Patents,’ 25 Tex. Intel. Prop. L.J. 39–72.
Hovenkamp, Erik N., and Jorge Lemus. 2016. ‘Reverse Settlement and Holdup at the Patent Office,’ forthcom-
ing, International Review of Law and Economics, accessed March 31, 2019 at https://papers.ssrn.com/sol3/
Hovenkamp, Herbert. 2011a. ‘Coase, Institutionalism, and the Origins of Law and Economics,’ 86 Indiana Law
Journal 499–542.
Hovenkamp, Herbert. 2011b. ‘Notice and Patent Remedies,’ 88 Texas L.Rev. Online 221–33.
Hovenkamp, Herbert. 2011c. ‘Post-Sale Restraints and Competitive Harm: The First Sale Doctrine in
Perspective,’ 66 NYU Annual Survey of American Law 487–547.
Hovenkamp, Herbert. 2014. ‘Consumer Welfare in Competition and Intellectual Property Law,’ 9 Competition
Policy International 53–68.
Hovenkamp, Herbert. 2015a. ‘Antitrust and the Patent System: A Reexamination,’ 76 OSU L.J. 467–564.
Hovenkamp, Herbert. 2015b. ‘Brulotte’s Web,’ 11 J. Comp. L. and Econ. 527–48.
Hovenkamp, Herbert. 2015c. ‘Inventing the Classical Constitution,’ 101 Iowa Law Rev. 1–53.
Hovenkamp, Herbert. 2015d. The Opening of American Law: Neoclassical Legal Thought, 1870-1970. New York
and London: Oxford University Press.
Hovenkamp, Herbert. 2015e. ‘The Rule of Reason and the Scope of the Patent,’ 52 San Diego Law Rev. 515–54.
Hovenkamp, Herbert. 2016a. ‘Antitrust and Information Technologies,’ 68 Fla. L. Rev. 419–65.
Hovenkamp, Herbert. 2016b. ‘Patent Exhaustion and Federalism: A Historical Note,’ 102 Va. L. Rev. Online
25–33.
Hovenkamp, Herbert. 2016c. ‘The Emergence of Classical Patent Law,’ 58 Ariz. L. Rev. 263–307.
Hovenkamp, Herbert. 2018a. ‘Antitrust and the Design of Production,’ 102 Cornell L. Rev. 1155.

Hovenkamp, Herbert. 2018b. ‘Reasonable Patent Exhaustion,’ 35 Yale J. on Regulation 513.

Khoury, Amir H. 2010. ‘Differential Patent Terms and the Commercial Capacity of Innovation,’ 18 Tex. Intell.
Prop. L.J. 373–417.
Lerner, Josh, and Jean Tirole. 2004. ‘Efficient Patent Pools,’ 94 Am. Econ. Rev. 691–711.
Lim, Daryl. 2013. Patent Misuse and Antitrust Law: Empirical, Doctrinal and Policy Perspectives. Cheltenham,
UK and Northampton, MA, USA: Edward Elgar Publishing.
Shavell, eds., Handbook of Law and Economics, vol. II. Amsterdam, Netherlands: Elsevier.
Merges, Robert P. 1988. ‘Commercial Success and Patent Standards: Economic Perspectives on Innovation,’ 76
Cal. L. Rev. 803–76.
Merges, Robert P. 2000. ‘One Hundred Years of Solicitude: Intellectual Property Law, 1900–2000,’ 88 California
Merges, Robert P., and Richard R. Nelson. 1990. ‘On the Complex Economics of Patent Scope,’ 90 Colum. L.
Rev. 839–916.
Mojibi, Ali. 2010. ‘An Empirical Study of the Effect of KSR v. Teleflex on the Federal Circuit’s Patent Validity
Jurisprudence,’ 20 Albany Law Journal of Science and Technology 559–96.
Moser, Petra. 2013. ‘Patents and Innovation: Evidence from Economic History,’ 27 J. Econ. Persp. 23–44.
Mossoff, Adam. 2013. ‘The Trespass Fallacy in Patent Law,’ 65 Florida Law Review 1687–712.
Newman, John M. 2012. ‘Anticompetitive Product Design in the New Economy,’ 39 Florida State University
Ochoa, Tyler T. 2001. ‘Patent and Copyright Term Extension and the Constitution: A Historical Perspective,’
49 Journal of the Copyright Society of the U.S.A. 19–125.
Ostrom, Elinor. 1990. Governing the Commons: The Evolution of Institutions for Collective Action. New York,
NY: Cambridge Univ. Press.
Park, Walter G., and Carlos Ginarte. 1997. ‘Intellectual Property Rights and Economic Growth,’ 15
Contemporary Economic Policy 54–6.
Patry, William. 1996. ‘Copyright and the Legislative Process: A Personal Perspective,’ 14 Cardozo Arts and
Rantanen, Jason. 2013. ‘The Federal Circuit’s New Obviousness Jurisprudence: An Empirical Study,’ 16
Stanford Technology Law Review 709–68.
Raustiala, Kal, and Christopher Sprigman. 2012. The Knockoff Economy: How Imitation Sparks Innovation.
New York, NY: Oxford University Press.
Romer, Paul M. 1990. ‘Endogenous Technical Change,’ 98 Journal of Political Economy 71–102.
Santore, Rudy, Michael McKee, and David Bjornstad. 2010. ‘Patent Pools as a Solution to Efficient Licensing
of Complementary Patents? Some Experimental Evidence,’ 53 J.L. and Econ. 167–84.
Sarnoff, Joshua D. 2005. The Historic and Modern Doctrines of Equivalents and Claiming the Future: Part II
(1870–1952),’ 87 Journal of Patent and Trademark Office Society 441–92.
Schmitz, James A., and Thomas J. Holmes. 2010. ‘Competition and Productivity: A Review of the Evidence,’ 2
Annual Review of Economics 618–42.
Schumpeter, Joseph A. 1943. Capitalism, Socialism, Democracy. New York, NY: Allen and Unwin.
Scott, John T., and Troy J. Scott. 2014. ‘Innovation Rivalry: Theory and Empirics,’ 41 Economia e Politica
Industriale-Journal of Industrial and Business Economics 25–53.
Seaman, Christopher B. 2012. ‘Willful Patent Infringement and Enhanced Damages after in re Seagate: An
Empirical Study,’ 97 Iowa Law Review 417–71.
Shapiro, Carl. 2001. ‘Navigating the Patent Thicket: Cross Licenses, Patent Pools, and Standard-Setting,’ in
A. Jaffe, J. Lerner, and S. Stern, eds., Innovation Policy and the Economy, vol. 1. Cambridge, MA: MIT Press.
Sichelman, Ted. 2014. ‘Purging Patent Law of “Private Law Remedies,”’ 92 Texas Law Review 517–71.
Solow, Robert M. 1956. ‘A Contribution to the Theory of Economic Growth,’ 70 Quarterly Journal of Economics
65–94.
Solow, Robert M. 1957. ‘Technical Change and the Aggregate Production Function,’ 39 Review of Economics
Torrance, Andrew W., and Bill Tomlinson. 2011. ‘Property Rules, Liability Rules, and Patents: One Experimental
View of the Cathedral,’ 14 Yale Journal of Law and Technology 138–62.
Wilson, Bruce B., Deputy Ass’t Atty. Gen., Antitrust Div., Dep’t of Justice. 1970. ‘Patent and Know-How
License Agreements: Field of Use, Territorial, Price and Quantity Restrictions,’ Remarks before the Fourth
New England Antitrust Conference, (Nov. 6, 1970).
Cases
1-800-CONTACTS, Inc. v. WhenU.Com, Inc., 414 F.3d 400 (2d Cir. 2005).

Adams v. Burke, 84 U.S. 453 (1873).

Allied Orthopedic Appliances, Inc. v. Tyco Health Care Grp. LP, 592 F.3d 991 (9th Cir. 2010).
Apple, Inc. v. Motorola, Inc., 757 F.3d 1286 (Fed. Cir. 2014).
Apple iPod iTunes Antitrust Litig., 2014 WL 4809288 (N.D. Cal. 2014).
Asahi Glass Co. v. Pentech Pharm., Inc., 289 F. Supp. 2d 986 (N.D. Ill. 2003).
Assessment Tech. of Wisconsin., LLC v. Wiredata, Inc., 350 F.3d 640 (7th Cir. 2003).
B. Braun Med., Inc. v. Abbott Labs., 124 F.3d 1419 (Fed. Cir. 1997).
Baker v. Selden, 101 U.S. 99 (1879).
Berkey Photo, Inc. v. Eastman Kodak Co., 603 F.2d 263 (2d Cir. 1979).
Best Lock Corp. v. Ilco Unican Corp., 94 F.3d 1563 (Fed. Cir. 1996).
Bikram’s Yoga College of India, L.P. v. Evolation Yoga, LLC, 803 F.3d 1032 (9th Cir. 2015).
Bloomer v. McQuewan, 55 U.S. 539 (1852).
Bobbs-Merrill Co. v. Straus, 210 U.S. 339 (1908).
Bowman v. Monsanto Co., 133 S. Ct. 1761 (2013).
Brooke Group Ltd. v. Brown and Williamson Tobacco Corp., 509 U.S. 209 (1993).
Brulotte v. Thys Co., 379 U.S. 29 (1964).
California Computer Prod., Inc. v. IBM Corp., 613 F.2d 727 (9th Cir. 1979).
Carbice Corp. of Am. v. Am. Patents Dev. Corp., 283 U.S. 27 (1931).
Chicago Board of Trade v. United States, 246 U.S. 231 (1918).
Chrysler Motors Corp. v Auto Body Panels of Ohio, Inc., 908 F.2d 951 (Fed. Cir. 1990).
Clorox Co. v. Sterling Winthrop, Inc., 117 F.3d 50 (2d Cir. 1997).
Continental TV, Inc. v. GTE Sylvania, Inc., 433 U.S. 36 (1977).
Cont’l Paper Bag Co. v. E. Paper Bag Co., 210 U.S. 405 (1908).
Cont’l Wall Paper Co. v. Louis Voight and Sons Co., 212 U.S. 227 (1909).
Coupe v. Royer, 155 U.S. 565 (1895).
Dippin’ Dots, Inc. v. Mosey, 476 F.3d 1337 (Fed. Cir. 2007).
E. Bement & Sons v. Nat’l Harrow Co., 186 U.S. 70 (1902).
eBay Inc. v. MercExchange, L.L.C., 547 U.S. 388 (2006).
Eldred v. Ashcroft, 537 U.S. 186 (2003).
F.T.C. v. Actavis, Inc., 570 U.S. 136 (2013).
F.T.C. v. Phoebe Putney Health Sys., Inc., 568 U.S. 216 (2013).
Gen. Talking Pictures Corp. v. W. Elec. Co., 304 U.S. 175 (1938).
Golden Bridge Tech., Inc. v. Motorola, Inc., 547 F.3d 266 (5th Cir. 2008).
Henry v. A.B. Dick Co., 224 U.S. 1 (1912).
Illinois Tool Works Inc. v. Indep. Ink, Inc., 547 U.S. 28 (2006).
Image Tech. Servs., Inc. v. Eastman Kodak Co., 125 F.3d 1195 (9th Cir. 1997).
Impression Prod., Inc. v. Lexmark Int’l, Inc., 137 S. Ct. 1523 (2017).
In re Apple iPod iTunes Antitrust Litig., 2014 WL 6783763 (N.D. Cal. 2014).
Intellectual Ventures I, LLC v. Capital One Financial Corp., 99 F. Supp. 3d 610 (D. Md. 2015).
International Salt Co. v. United States, 332 U.S. 392 (1947).
ISO Antitrust Litigation, 203 F.3d 1322 (Fed. Cir. 2000).
Johnson & Johnston Assocs., Inc. v. R.E. Serv. Co., 285 F.3d 1046 (Fed. Cir. 2002).
Jungersen v. Ostby & Barton Co., 335 U.S. 560 (1949).
Kimble v. Marvel Entm’t Co., LLC, 135 S. Ct. 2401 (2015).
Kirtsaeng v. John Wiley & Sons, Inc., 133 S. Ct. 1351 (2013).
Lasercomb Am., Inc. v. Reynolds, 911 F.2d 970 (4th Cir. 1990).
Leegin Creative Leather Prod., Inc. v. PSKS, Inc., 551 U.S. 877 (2007).
Lotus Dev. Corp. v. Borland Int’l, Inc., 49 F.3d 807 (1st Cir. 1995).
Motion Picture Patents Co. v. Universal Film Co., 243 U.S. 502 (1917).
Nero AG v. MPEG LA, LLC., 2010 WL 4366448 (C.D. Cal. 2010).
Nobelpharma AB v. Implant Innovations, Inc., 129 F.3d 1463 (Fed. Cir. 1997).
Octane Fitness, LLC v. ICON Health & Fitness, Inc., 134 S. Ct. 1749 (2014).
Omega S.A. v. Costco Wholesale Corp., 776 F.3d 692 (9th Cir. 2015).
Oracle Am., Inc. v. Google Inc., 750 F.3d 1339 (Fed. Cir. 2014).
Princo Corp. v. Int’l Trade Comm’n, 616 F.3d 1338 (Fed. Cir. 2010).
Qualcomm, Inc. v. Broadcom Corp., 548 F.3d 1004 (Fed. Cir. 2008).
Quanta Computer, Inc. v. LG Elecs., Inc., 553 U.S. 617 (2008).
Queen City Pizza Inc. v. Domino’s Pizza Inc., 124 F.3d 430 (3d Cir. 1997).
Rosetta Stone, Ltd. v. Google, Inc., 676 F.3d 144 (4th Cir. 2012).

SCM Corp. v. Xerox Corp., 645 F.2d 1195 (2d Cir. 1981).
Siegel v. Chicken Delight, Inc., 448 F.2d 43 (9th Cir. 1971).
Standard Oil Co. of Calif. v. United States, 337 U.S. 293 (1949).
Standard Sanitary Mfg. Co. v. United States, 226 U.S. 20 (1912).
Teva Pharmaceuticals USA, Inc. v. Sandoz Inc., 2014 WL 2885379 (U.S.) (U.S., 2014).
TrafFix Devices, Inc. v. Market Displays, Inc., 532 U.S. 23 (2001).
Trebro Mfr., Inc. v. Firefly Equip., LLC, 748 F.3d 1159 (Fed. Cir. 2014).
Tyco Healthcare Group LP v. Mutual Pharma. Co., Inc., 762 F.3d 1338 (Fed. Cir. 2014).
U.S. Gypsum Co. v. National Gypsum Co., 352 U.S. 457 (1957).
United States v. Aluminum Co. of Am., 148 F.2d 416 (2d Cir. 1945).
United States v. Am. Can Co., 230 F. 859 (D. Md. 1916).
United States v. AT&T, 552 F. Supp. 131 (D.D.C. 1982).
United States v. Dentsply Int’l, Inc., 399 F.3d 181 (3d Cir. 2005).
United States v. E.I. du Pont de Nemours and Co., 351 U.S. 377 (1956).
United States v. General Electric Co., 272 U.S. 476 (1926).
United States v. Line Material Co., 333 U.S. 287 (1948).
United States v. Loew’s, Inc., 371 U.S. 38 (1962).
United States v. Microsoft Corp., 253 F.3d 34 (D.C. Cir. 2001).
United States v. Singer Mfg. Co., 374 U.S. 174 (1963).
United States v. Univis Lens Co., 316 U.S. 241 (1942).
USM Corp. v. SPS Tech., Inc., 694 F.2d 505 (7th Cir. 1982).
Walker Process Equip., Inc. v. Food Mach. and Chem. Corp., 382 U.S. 172 (1965).

15 U.S.C. §§ 1, 2, 13, 14, 18, 68b–68c, 70b.
17 U.S.C. §§ 106, 109(a).
35 U.S.C. §§ 102(a)(1), 103, 261, 271, 271(d)(3), 271(d)(4), 271(d)(5), 285.
Copyright Act of 1976, Pub. L. No. 94-533, 90 Stat. 2541 (1976), 17 U.S.C. § 101 et seq.
Copyright Term Extension Act of 1998, Pub. L. No. 105-298, 112 Stat. 2827 (1998), codified in relevant part
at 17 U.S.C. § 302.

9. Intellectual property and the economics of
product differentiation
Christopher S. Yoo* 9
Contents
I. Introduction
II. The Economics of Product Differentiation
A. Monopolistic Competition
B. Spatial Competition
C. Implications
III. Patent
IV. Copyright
A. Differentiated Products Competition between Different Works
B. Spatial Competition and Impure Public Goods
V. Trademark
VI. Conclusion
References
I. INTRODUCTION
To date, economic analyses of intellectual property (IP) have been dominated by models
that frame IP as monopolies and/or public goods. These approaches have given rise to
general policy inferences, such as systematic underproduction, deadweight losses resulting
from pricing above marginal cost, and the supposedly inevitable tradeoff between access
to IP and the incentives to create it. These inferences in turn have led many commentators
to call for calibrating different aspects of IP protection to mitigate these problems.
A new literature is emerging that relaxes the assumption implicit in these models that
the relevant products are homogeneous and instead proceeds from the premise that all
IP faces competition from imperfect substitutes. Allowing for the possibility of product
differentiation provides new insights into the economics of IP. It helps explain persistent
features of IP markets that the traditional approaches cannot. It challenges the extent to
which IP allows rightsholders to earn monopoly profits. It makes the market dynamics
more complex by opening up new dimensions along which companies can compete. This
in turn allows for possible sources of welfare outside of price and quantity and yields equi-
libria with different welfare characteristics. The inevitable tendency towards systematic
underproduction is replaced by a more contingent world in which either underproduction
* John H. Chestnut Professor of Law, Communication, and Computer and Information

Science, University of Pennsylvania.
262

Intellectual property and the economics of product differentiation 263
or overproduction is possible. It also suggests a broader range of policy options for

promoting optimality in IP markets. These insights have led some commentators to call
this literature ‘the most important development in the economic analysis of copyright in
recent years’ (Bracha and Syed, 2014, p. 1842).
This chapter will review the economics of product differentiation and the literature
applying that literature to IP. Section II introduces the economics of product differentia-
tion. Sections III, IV, and V survey the literature applying these approaches to patent,
copyright, and trademark respectively.
II. THE ECONOMICS OF PRODUCT DIFFERENTIATION

The seminal works in the economies of product differentiation are the monopolistic
competition theory advanced by Edward Chamberlin (1933) and the spatial competition
models pioneered by Harold Hotelling (1929). Each approach emphasizes different
aspects of the underlying economics. Monopolistic competition models the market
dynamics in the traditional price-quantity space of Marshallian economics and takes
product differentiation indirectly. Conversely, spatial competition portrays differentiation
directly and deals with price and quantity indirectly.
A. Monopolistic Competition
Monopolistic competition retains most of the assumptions underlying perfect competi-

tion, including free entry and the presence of a substantial number of sellers.1 The key
difference is that monopolistic competition relaxes the assumption that products are
homogeneous and serve as perfect substitutes for one another. When products are dif-
ferentiated, the structure of demand for any particular product allows producers to raise
their prices without losing all of their demand, because they will be able to retain those
customers who place the highest value on the particular variant that they offer. Product
differentiation thus provides each producer with a degree of power over price sufficient to
justify modeling each product as facing a downward-sloping demand curve.
Monopolistic competition theory simplifies the analysis by assuming that consumer
preferences are symmetric with respect to each product in the group. The differences
across different products were abstracted away by product differentiation as consumer
demand as reflecting a generic preference for variety rather than for any particular form
of differentiated product. Most importantly for the purposes of this chapter, Chamberlin
(1933, pp. 57–64) regarded patents, copyrighted works, and trademarks as examples of
differentiated products to which his theory applied.
The primary effect of this assumption is to place each product in equal competition
with all other products in the group rather than in localized competition with a smaller
set of near neighbors. Chamberlin’s original formulation also assumed that each producer
1
Robinson (1933) proposed a related theory of imperfect competition at roughly the same time
as Chamberlin. Chamberlin’s approach focused explicitly on product differentiation and provides
the more appropriate focus for this chapter.

faced identical cost curves. This allowed him to employ a single graph portraying the
price-quantity response of a representative firm to model the entire market, although
relaxing the symmetry assumptions were later shown not to affect the core analysis
(Kaldor, 1935; Archibald, 1961).
Because monopolistic competition portrays market interactions in a classic price-
quantity space, it is quite easily integrated into a conventional welfare analysis based on
economic surplus. In the short run, profit-maximizing producers will produce at the quan-
tity (QSR) where their marginal revenue (MRSR) and marginal cost (MC) curves intersect.
Because monopolistically competitive producers face downward-sloping demand curves
(DSR), in the short run they will set prices in the same manner as a monopolist (PSR), as
depicted in Figure 9.1. This results in a significant short-run deadweight loss (DWLSR).
Should price exceed average cost (AC), producers may also earn short-run supracompeti-
tive profits (ProfitSR).
Were entry impossible, this short-run equilibrium would be stable, and the long-run
outcome would be the same as under the monopoly analysis. Monopolistic competi-
tion, however, assumes that entry by close substitutes is always possible. As a result, the
presence of supracompetitive profits attracts other producers selling similar products.
Because all of the products in the market are in equal competition with one another, new
entrants take business equally from each of the incumbents, with the demand and supply
curves representing the decision confronting a representative firm, with the entire industry
constituting the sum of all such graphs. Entry causes the demand curve confronting each
incumbent to shift inwards (DLR), as customers substitute purchases of the new product
for those of the incumbents. Although some have suggested that the demand curve could
either increase or decrease in slope, the increase in the number of imperfect substitutes
should generally cause demand to become more elastic (Varian, 2010). The resulting
long-run equilibrium is depicted in Figure 9.2.
Entry continues until no profits remain, at which point the long-run equilibrium price
PSR
ProfitSR
DWLSR
AC
MC
DSR
QSR Q
MRSR
Figure 9.1 Short-run equilibrium under monopolistic competition

PSR
PLR
DWLLR AC
MC
DLR DSR
QLR QSR Q
Figure 9.2 Long-run equilibrium under monopolistic competition
(PLR) will be lower than the short-run equilibrium price (PSR). Under Chamberlin’s
original formulation, this occurs when the surplus appropriated by each producer is just
enough to cover the fixed costs of entry, a condition which exists when the long-run
demand curve (DLR) is tangent to the average cost curve (AC). There is, however, a well-
known exception to Chamberlin’s zero-profit result. The indivisibility of fixed costs may
create a situation in which n products would earn small profits while n + 1 products would
run losses. This so-called ‘integer problem’ allows for an equilibrium in which n products
each earn sustainable profits (Kaldor, 1935). So long as the economy is sufficiently ‘large’
(i.e., so long as n is relatively sizeable), such profits will be negligible (Eaton and Lipsey,
1989).
As entry causes the demand curve to flatten, long-run deadweight loss shrinks
(DWLLR). Note that whether the market will reach long-run equilibrium on a flatter
portion of the demand curve depends on the assumption that the relevant demand curve
is linear. If the demand function is curved, it could be tangent to the average cost curve at
any one of a number of points. In that case, it is no longer inevitable that the market will
reach equilibrium at a place where the spread between price and marginal cost is narrower.
In addition, the tangency solution also presupposes that the producer is charging the
same price to all consumers. Allowing for price discrimination raises the possibility that
a firm might still earn supracompetitive profits even at the point of tangency, meaning
that further entry would occur.
The equilibrium number of products can be determined by dividing the total
surplus associated with the entire market by the fixed costs needed for entry. Indeed,
as the size of the market expands or the size of the fixed costs declines, the number
of products asymptotically approaches infinity and the deadweight loss approaches
zero. In this way, as Yoo (2004) pointed out, that access to existing products may be
promoted indirectly by stimulating entry instead of being promoted directly by reduc-
ing the scope of the entitlement. Whether an economy is ‘large’ in this manner does
not depend upon the size of the fixed costs relative to the size of the marginal costs,

as suggested by the traditional approach. Instead, it depends on the magnitude of the

fixed costs relative to the overall market. Markets for differentiated products are also
biased against products whose demands are structured in a way that makes it difficult
for the producer to appropriate surplus (Spence, 1976a, 1976b; Koenker and Perry,
1981; Beath and Katsoulacos, 1991).
Chamberlin regarded the welfare characteristics of the resulting equilibrium as
demonstrating the systematic inefficiency of markets. For example, Chamberlin observed
that monopolistically competitive markets necessarily set prices above marginal cost. In
addition, the fact that markets reach equilibrium on the declining portion of the average
cost curve implied that the monopolistically competitive markets exhibit a sustainable ten-
dency towards excess capacity and that overall costs could be reduced if the total number
of sellers were reduced and the existing sellers were permitted to increase their production.
He also implied that many forms of product differentiation, particularly those associated
with trademarks, are spurious and represent attempts to increase profits that do not yield
any consumer benefits. The models demonstrated the existence of ‘wastes of competition’
that can be as severe as the ‘wastes of monopoly’ and required regulatory correction
(Chamberlin, 1933, pp. 104–109).
Chamberlin’s work prompted a vigorous attack from the Chicago School, who
appeared to regard its finding of endemic market failure as an attack on perfect com-
petition as the benchmark for optimal economic performance (see, e.g., Stigler, 1949;
Friedman, 1953; see Chamberlin, 1957, for his response). Subsequent development of the
literature on monopolistic competition appears to have narrowed the scope of some of
these disagreements. For example, Chamberlin’s excess capacity theorem, which is often
regarded as the theory’s most significant result, is based entirely on measuring welfare on
the price-quantity space. It ignores the fact that there may be compensating welfare gains
from the product characteristics space (Bishop, 1967; Spence, 1976a; Dixit and Stiglitz,
1977), as Chamberlin (1950, p. 89) recognized in his later work.
A more trenchant criticism offered centers on monopolistic competition’s failure to
generate testable hypotheses (Archibald, 1961). As Paul Samuelson (1967) has noted, a
theory should be measured by the insights that it provides, not for its elegance or for the
simplicity of the policy inferences it is able to generate. Commentators have noted the
difficulties in measuring monopolistic competition empirically (Spence, 1976a).
B. Spatial Competition
The fact that monopolistic competition depicts market interactions in the classic price-
quantity space of microeconomics allows for a natural representation of economic
welfare associated with total surplus. This approach does not portray product differentia-
tion directly. Spatial competition models, in contrast, adopt the opposite tack, making
product differentiation the primary variable instead of price and quantity. The original
formulations assumed that, rather than competing on price, firms instead vie for business
by choosing a location along a linear geographic space. Because of transportation costs,
customers derive greater utility from purchasing from sellers that are closer to their
locations. Utility declines as the distance from the store increases until the entire surplus
is completely consumed by transportation costs, at which point the customer decides not
to purchase from that vendor. Spatial differentiation thus gives sellers a degree of power

utility
A B C A B C D E F G
Figure 9.3 Impact of fixed costs on the equilibrium under spatial competition
over price, as they can increase price without losing those customers who are situated
closest (Hotelling, 1929).
Economists quickly recognized that the same framework could be used to model
competition among products distributed along a characteristics space rather than a geo-
graphic space (Hotelling, 1929; Chamberlin, 1933; Kaldor, 1935; see generally Lancaster,
1979). Under this approach, customers decide whether to purchase a particular product
based on how closely that product matches their ideal preferences. The decline in utility
represented by transportation costs in geographic location models is replaced by diver-
gence from a consumer’s preferred characteristics.
Examples of two such characteristics spaces are depicted in Figure 9.3. The horizontal
dimension depicts where along the continuum of characteristics a particular product is
located. The vertical dimension in the graph represents the net surplus available from
consumers occupying any particular location. Consumers’ ideal preferences are assumed
to be distributed uniformly across the characteristics space. Each product is produced by
a different firm, and the surplus captured by each product is depicted by a triangle. The
decline in utility resulting from the good’s divergence from the consumer’s ideal prefer-
ences is represented by the slope of the triangle’s sides. The slope of this line is determined
by the structure of demand, as reflected in the relevant cross-price elasticities. If a product
serves as a relatively good substitute for similar products, the slope will be relatively flat.
If not, the slope will be relatively steep.
Consumers are assumed to purchase whichever product lies closest to their ideal
preferences. As noted earlier, the symmetric preference assumption posits that all
products within a group are in equal competition with one another and that entry by a
new product takes business from all incumbents evenly. This assumption is represented
in spatial models either by positing that all products enter simultaneously or by assuming
that incumbent products respond to product entry by costlessly shifting their position.
The result is an equilibrium in which products are evenly distributed across the relevant
product space. To ensure the existence of equilibrium, the characteristics space is typically

modeled as a unit line or a unit circle. Moreover, firms face competition in the areas
where two triangles overlap, but will act as monopolists in the areas closer to the peaks of
the triangles, which effectively means that they will face a kinked demand curve and the
accompanying difficulties with defining equilibrium in terms of tangency (Salop, 1979).2
As was the case under monopolistic competition, entry by additional products divides
the available surplus into increasingly smaller fragments until the size of those fragments
equals the fixed costs of entry, at which point no product earns supracompetitive profits,
subject again to a mild exception for fixed-cost indivisibilities. Again, the equilibrium
number of products can be determined by dividing the total surplus by the fixed costs of
entry. If the total surplus is significantly greater than the fixed costs of entry, the economy
will be relatively ‘large’ and the competition among sellers will be relatively substantial. As
the available surplus increases or the size of the fixed costs approaches zero, the number of
products will approach infinity and all customers will purchase products exactly matched
to their tastes. If the total surplus is small relative to the fixed costs, the economy will be
relatively ‘small’ and products may well enjoy a degree of local monopoly or oligopoly
power. Although entry by competitors remains possible under these conditions, the
volume may be too low to support additional sellers (Eaton and Wooders, 1985; Eaton
and Lipsey, 1989).
Spatial competition provides for a direct representation of product differentiation,
although the fact that these models do not employ the price-quantity space of neoclas-
sical economics means that price competition must be modeled separately and makes it
more difficult to integrate spatial models into conventional analyses of economic welfare.
In addition, spatial competition models are subject to a number of qualifications and
refinements. Not all products can be organized into a simple spectrum of characteristics.
Furthermore, while the assumption that utility falls linearly with distance is natural when
measuring the impact of distance in a geographic space, it appears to be less plausible
in a characteristics space, where the decay in utility could take just about any shape. As
it turns out, the assumption that utility falls linearly is a strong one. For example, the
assumption that utility decays quadratically instead of linearly leads to radically different
results (d’Aspremont et al., 1979).
The two different depictions in Figure 9.3 illustrate how the relative size of the total
available surplus affects the degree of competition. The available surplus in the right-hand
graph is 50 percent larger than that in the left-hand graph. In both graphs, each individual
product captures an identical surplus. As the graphs illustrate, increasing the size of the
overall market yields a fairly substantial increase in the degree of competition. More
sophisticated models that allow for price reactions among competitors further underscore
the importance of the relative size of the economy. These models demonstrate that so
2
Interestingly, the assumption that prices are constant and endogenously determined can
give rise to a different equilibrium in which products, rather than being spread evenly across
the characteristics space, divide the market by entering at the same point (Hotelling, 1929). The
oddities resulting from refusing to allow for price competition are well recognized. For example,
the assumption that firms do not compete on price necessarily implies that effective competition
exists only with respect to consumers located equidistantly from two works. This effect disappears,
however, if the model is broadened to allow for endogenous pricing in which price competition is
possible.

long as marginal costs are nonzero, increases in the size of the economy cause prices
to approach marginal cost and reduce profits by bringing the revenue captured by each
producer more into line with fixed costs.
Furthermore, the analysis becomes significantly more complicated when one relaxes the
rather restrictive assumptions that typify the basic model of spatial competition described
above. More refined models allow for the possibility of sunk costs in location and sequen-
tial entry (Baumol, 1967; Hay, 1976; Prescott and Visscher, 1977; Eaton and Lipsey, 1980;
Lane, 1980; Bonanno, 1987; Neven, 1987; Bhaskar and To, 2004). Other models relax the
assumption that prices are fixed and allow prices to be determined endogenously (Salop,
1979; Eaton and Wooders, 1985). Still other models allow for the possibility that a single
firm might produce multiple products occupying multiple locations (Schmalensee, 1978;
Eaton and Lipsey, 1979; Brander and Eaton, 1984; Judd, 1985; Bonanno, 1987). Still
other models allow for the possibility that consumers base their purchases on multiple
dimensions of characteristics, which means that firms compete with more than just two
adjacent competitors. For example, if spatial competition takes place on three dimensions,
each work may compete with as many as six adjacent neighbors. If competition expands
to four dimensions, works may theoretically compete with as many as half the works oper-
ating in the product group (Archibald and Rosenbluth, 1975). An empirical assessment
of the automobile industry concluded that differentiated products compete in as many as
six dimensions (Feenstra and Levinsohn, 1995). Finally, the results change significantly
when one allows for the possibility that consumer preferences are not distributed equally
across the characteristics space (Kaldor, 1935; Eaton and Lipsey, 1976). The basic model
is nonetheless sufficient to capture the key intuitions about how differentiated products
compete and to provide useful insights into markets for differentiated products. As was
the case with monopolistic competition, leading commentators on spatial competition
have noted the difficulty in determining whether any particular equilibrium is optimal
(Lancaster, 1979; Eaton and Lipsey, 1989).
C. Implications
Compared with the traditional approach, the predictions of the differentiated products
approach to copyright fit better with features of real-world markets for IP. One of the
most interesting aspects of the differentiated products approach is that it reconceptualizes
the tension between access and incentives that motivates much of the economic analysis
of IP. Chamberlin’s primary point is that the fact that equilibrium prices in markets for
differentiated products exceed marginal cost raises questions whether marginal cost pric-
ing represents the appropriate benchmark for efficiency.
In addition, two complementary insights reveal how product differentiation both
creates and limits the degree of market power that producers enjoy. The downward-
sloping nature of demand curves associated with product differentiation also provides
an explanation of how producers can maintain the power over price to engage in price
discrimination in markets that are open to free entry and subject to competition (for the
seminal work, see Spulber, 1979, 1981; Katz, 1984; and Borenstein, 1985; for a recent
survey, see Stole, 2007). At the same time, the fact that products serve as imperfect
substitutes for one another allows multiple producers to coexist and prevent markets from
collapsing into natural monopolies even when average costs are constantly declining by

permitting producers to compete on dimensions other than price (Maurer and Scotchmer,
2002). In so doing, product differentiation appears to provide a more realistic description
of real-world markets.
At the same time, in contrast to perfect competition, in which entry generally promotes
economic welfare, product differentiation gives rise to the possibility that entry may be
excessive or insufficient. Whether a market will tend towards too much or too little entry
depends on the extent to which the surplus that a new entrant appropriates is the result
of demand creation, that is incremental sales to new customers who previously were not
in the market or incremental welfare by permitting purchasers to consume products that
provide a better fit with their preferences, or demand diversion, that is surplus redistributed
from producers already in the market.3 The entrant simply compares the total surplus it
appropriates to its costs and enters whenever the former equals or exceeds the latter. While
demand creation represents an incremental contribution to welfare, demand diversion
does not. This causes firms to enter even when the fixed costs of entry exceed incremental
benefits of doing so, in which case, entry is excessive. The greater the proportion of sur-
plus comes from demand diversion, the greater the tendency towards excess entry. Indeed,
Salop (1979) estimates that if surplus falls off linearly with distance, the equilibrium
number of producers will exceed the optimum by a factor of two. As Eaton and Lipsey
(1989) note, there is thus no invisible hand inexorably pushing markets for differentiated
products towards optimality. While demand diversion in homogeneous product markets
unambiguously produces excess entry, differentiated product markets may create either
excess or insufficient entry (Mankiw and Whinston, 1986).
Yoo (2004) points out that the tendency towards excess entry created by demand diver-
sion may be offset by the tendency towards insufficient entry created by producers’ inabil-
ity to appropriate the surplus they create. Whether entry levels will exceed or fall short of
the optimum depends upon which effect dominates.4 The possibility of insufficient entry
created by incomplete appropriability also opens up the policy space by suggesting that
access and incentives need not always be in conflict. If, on the one hand, entry is excessive,
any strengthening of IP rights will push incentives to create in the wrong direction. If, on
the other hand, entry is insufficient, strengthening IP rights can push incentives in the
right direction while simultaneously indirectly promoting access by stimulating greater
price competition among the differentiated products.
3
The terminology used in this discussion is taken from Borenstein (1985). For similar analyses
using other terminology, see Mankiw and Whinston (1986) (‘business stealing effect’) and Beath
and Katsoulacos (1991) (‘cannibalisation’). For other works recognizing the concept without
employing a distinctive moniker, see Spence (1976a, 1976b); Koenker and Perry (1981).
4
Although demand diversion allows a new product to capture the same number of buyers as
would result under complete appropriability, it does not result in the capture of the same buyers.
Instead, demand diversion substitutes buyers who already were purchasing other products for new
buyers whose purchases represent incremental sales. Thus, although the total number of sales may
reach optimal levels, the total surplus generated by those sales is likely to fall somewhat short of
welfare-maximizing levels because the buyers who actually purchase the product are not necessarily
those who place the highest value on the good. Some consumers may purchase goods that provide
a better fit with their ideal preferences, while others may purchase goods in which the fit is worse.
As a result, the equilibrium amounts to a close approximation of a first-best outcome that falls
somewhat short of maximizing welfare.

III. PATENT
During the early 1990s, economists produced a vibrant body of scholarship applying
the economics of product differentiation to IP. This literature focused on the tradeoff
between patent length, measured in the number of years of exclusivity included in the
patent grant, and breadth, determined by the extent to which patents foreclose others from
using similar technologies. For example, Gilbert and Shapiro (1990) offer a model that
measured product differentiation by the ability to raise price above marginal cost to find
that welfare would be maximized if patents were infinitely lived and patent breadth were
adjusted to provide the appropriate reward for investment. This is because adjusting patent
duration gives rise to the familiar access-incentives tradeoff by both increasing the surplus
contained within the patent grant and permitting increases in price, whereas Gilbert and
Shapiro’s measure of breadth implicates only the latter.5 Gallini (1992) defines breadth
in terms of the size of the cost to invent around a patent to present a model finding
short-lived, broad patents would be optimal because longer terms hurt patent holders by
increasing competitors’ incentives to invent around the patent.6 Klemperer (1990) models
production differentiation directly in a Hotelling-style model, characterizing patent scope
as the size of the characteristics space over which the patent gives the patentee exclusivity
rights. He concludes that if consumer preferences are homogeneous across all products,
infinitely lived, narrow patents would maximize welfare. If consumers have strong prefer-
ences for particular product characteristics, short, broad patents would improve welfare.
This literature recognizes the existence of additional policy instruments that would
calibrate the strength of the patent grant. Gilbert and Shapiro define breadth in a way that
effectively focuses on appropriability of surplus. Klemperer demonstrates the ambiguous
impact of demand diversion by showing how a high level of substitution among goods
favors granting a narrow patent.7 The result is a policy space that is richer and more
nuanced than the previous literature which focused exclusively on calibrating patent
duration (Nordhaus, 1969; Scherer 1972).
Later scholars extended this literature in important ways. Matutes et al. (1996) look
at the proper patent regime for basic innovations that benefit multiple markets. In order
to minimize the period during which the inventor postpones disclosing its invention in
order to have more time to develop the expertise to commercialize these multiple markets,
Matutes et al. prefer a patent that is short in duration, but broad in scope. O’Donoghue et
al. (1998) apply a conventional spatial competition model with substitutes from alternative
technologies distributed along a quality scale to distinguish between lagging breadth,
which is breadth that forecloses low-cost imitations, and leading breadth, which forecloses
higher-quality innovations. They found patents that protected lagging breadth provided
sufficient incentives for R&D. With respect to leading breadth, a long, narrow patent was
superior in reducing R&D costs, while a short, broad patent would reduce market distor-
tions. The authors conclude that a long, narrow patent is preferable when the hit rate of
5
Tandon (1982) proposed a similar model favoring infinitely lived patents limited by a
compulsory license.
6
Denicolò (1996) presents a model yielding a similar result.
7
See also Waterson (1990), using a similar spatial competition model to show how demand
diversion can lead to excess entry.

new innovations is high. Maurer and Scotchmer (2002) focus on independent invention,
arguing for narrow breadth protection because the threat of independent invention would
encourage licensing of the patent and discourage duplicative investments in R&D.8
Beschorner (2008) frames his analysis as a tradeoff among patent length, breadth, and
what he terms height, which is the degree of novelty required to justify receiving a patent,
to conclude that a finite patent length would maximize welfare and that a monopolist
would require a lower level of novelty than would be socially optimal. Further explora-
tions of the other aspects of product differentiation theory are likely to follow.
IV. COPYRIGHT
An equally dynamic body of scholarship has emerged applying product differentiation
theory to copyright, this time conducted primarily by legal academics. The initial analyses
used product differentiation to analyse competition between a work and inferior quality
copies of the same work (Johnson, 1985; Liebowitz, 1986; Besen and Kirby, 1989), analyse
competition between creative and noncreative works (Lunney, 1996a), or to briefly men-
tion the impact of demand diversion (Meurer, 2001).
A. Differentiated Products Competition between Different Works
More complete analyses of competition between different copyrighted works began with
Yoo (2004), who offered the most complete exploration of the dynamics and welfare char-
acteristics of both the monopolistic competition and spatial competition approaches to
product differentiation that has yet appeared in the literature. Importantly, it applies the
traditional length versus breadth tradeoff to copyright, while adding a third dimension
termed intensity, which refers to the proportion of the available surplus that copyright
holders can appropriate through price discrimination and other similar mechanisms. He
also notes that the size of the market relative to fixed cost determines the number of works
in equilibrium. As the size of the market increases and the size of fixed costs decreases,
pricing converges to the perfectly competitive outcome, although the problems of excess
entry become worse, which opens the possibility of promoting access by increasing the
number of surplus-generating activities contained within the copyright and facilitating
copyright holders’ ability to engage in price discrimination and relying on the ensuing
price competition to reduce prices, although doing so could exacerbate problems of excess
entry. He then analyses the implications for copyright doctrines such as fair use, the first-
sale doctrine, digital rights management, and derivative works.
Abramowicz (2004) employs a spatial competition model to analyse the impact of
demand diversion on optimal entry. Although he recognizes that a market can reach
equilibrium with too much or too little entry, he relies on the findings of the circle model
put forth by Salop (1979) and a simulation to suggest that excess entry is the more likely
outcome. He acknowledges that Salop’s result depends on the assumption that the welfare
that individual consumers derive from consuming a particular work falls off linearly with
8
See Lemley and McKenna (2012) for a related argument appearing in the legal literature.

distance in the characteristics space. He also addresses a wide range of other economic
issues, including wealth distribution, winner-take-all markets, positional goods, and exter-
nalities, and noneconomic issues, such as democracy. Faulhaber (2006) similarly applies
the Salop model along with sunk costs in location to provide a model finding excess
entry in music. Lemley and McKenna (2012) are similarly concerned about excess entry,
arguing for narrowing breadth by increasing the degree of similarity required to find
infringement and narrowing leading breadth by construing fair use and derivative works
doctrines to provide greater latitude to works that provide incremental contributions.
The divergent emphases of these two lines of research underscore the ambiguity of the
welfare characteristics of differentiated product models, which both authors recognize.
Bracha and Syed (2014) acknowledge the contribution of the Yoo and Abramowicz
papers, but criticize both for drawing overly simple policy inferences. While both papers
emphasize different sides of the divergent possible results, both acknowledge the inherent
ambiguity of the differentiated products equilibria.9
The welfare ambiguity of product differentiation underscores the importance of the
small empirical literature exploring the economic performance of markets for creative
works. Like the theoretical literature, the empirical literature is divided. On the one hand,
Goettler and Shachar (2001) empirically assess the spatial competition among the three
major US television networks, concluding that the networks’ program offerings nearly
fully achieved the optimum suggested by the underlying Nash equilibrium and that the
shortfall was largely (but not completely) explained by the networks’ adherence to the
rules of thumb against airing sitcoms after 10:00 p.m. and against airing news magazines
before 10:00 p.m. On the other hand, Berry and Waldfogel (1999) empirically study entry
patterns in the radio industry, finding that the deadweight losses attributable to excess
entry may be substantial. They acknowledge that their study focuses exclusively on
advertisers and therefore ignores potential benefits to users and that their assumption that
the radio market is composed of homogeneous products causes them to ignore potential
welfare benefits resulting from product differentiation.
B. Spatial Competition and Impure Public Goods
A small literature also exists connecting the economics of product differentiation with the
theory of impure public goods. The conventional wisdom is that IP is a pure public good
and that as such exhibits a systematic tendency towards underproduction. Although most
attribute that to the pricing problems inherent in zero marginal cost goods, a review of
seminal works on public goods indicates that the problem is one of incentive compatibility
inherent in the Samuelson condition (1954), stemming from the inability to get consumers
to reveal their marginal preference for public goods.
The literature on impure public goods associated with club goods (Buchanan, 1965)
or local public goods (Tiebout, 1956) adds a congestion function to the standard public
9
Bracha and Syed (2014) also place great weight on the distinction between the impact of
product differentiation on supramarginal versus inframarginal works. All of this is taken into
account by the symmetry assumption, and the literature has long recognized that relaxing this
assumption leads to inferences that are more ambiguous that Bracha and Syed suggest.

goods set up. The resulting equilibria trade off the incurrence of additional fixed costs
against the reduction in congestion costs associated with creating an additional impure
public good. Unlike the Samuelson condition, congestion is potentially incentive
compatible. Numerous Nobel laureates have identified the connection between spatial
competition and impure public goods, with the congestion function being replaced with
the transportation costs function in the product characteristics space (Samuelson, 1958;
Buchanan, 1965; Stiglitz, 1977).
Yoo (2007) applies the insights of impure public goods theory to copyright. As is the
case with product differentiation, simple policy inferences disappear for impure public
goods, including the systematic tendency towards underproduction. It similarly opens
up the possibility of alternative policy approaches, such as promoting access by increas-
ing the size of the market and stimulating entry that induces greater price competition.
Because the shape of the congestion function can take any form, there is no necessary
reason to believe that the equilibrium induced by the congestion function will prove
optimal. Like markets for differentiated products, markets for impure public goods can
produce too many works or too few.10
V. TRADEMARK
Of all of the forms of IP, Chamberlin (1933) devoted the most attention to trademarks,
presenting an extended discussion of them in Appendix E of his book. He regarded the
differentiation associated with trademarks as motivated primarily to promote monopoly
and argued that lowering the standard for infringement would bring markets closer to the
competitive ideal. He particularly questioned the need for extending trademark protection
to descriptive words, color, shape, design, packaging, and labels, when an ‘inconspicuous
identification mark or the name and address of the producer’ would suffice. Anything
more simply conveys monopoly power to the mark holder. Chamberlin acknowledged
that free imitation of trademarks would harm consumers by eliminating incentives to
maintain product quality, but thought that an alternative regime centered on ‘defining
quality standards by law’ ‘has large possibilities’ and ‘would be equally effective.’ He also
acknowledged that trademarks can stimulate variety and give consumers wider choice,
but such variety came at the cost of higher prices. Although Chamberlin conceded that
theory provides no basis for how best to strike this tradeoff, on balance he nonetheless
came down squarely against trademarks, arguing that weakening trademark protec-
tion would discourage ‘[u]seless innovation,’ would reduce the ‘wastes of advertising,’ and
would focus producers less on creating monopoly and more on production, which in turn
would create ‘[f]ewer “business” men and more laborers.’ Innovations would still be pro-
tected by patent law and by first-mover advantages. If this were insufficient, trademark
protection should be only of limited duration, say five years (Chamberlin, 1933, app. E).
Chamberlin launched a tradition of scholars largely critical of trademarks as a source
of spurious differentiation (see, e.g., Brown, 1948; Scherer, 1970, 1976; Comanor and
10
For a related analysis, see Barnes (2011) arguing that impure public goods theory does not
support strengthening copyright protection.

Wilson, 1979; Schmalensee, 1979). Contemporary antitrust scholars raised similar

concerns (Bain, 1956). These concerns reached their zenith in the late 1970s, exemplified
by the Federal Trade Commission’s (FTC’s) case against Borden’s ReaLemon lemon juice
(for an overview, see Mensch and Freeman, 1990; McClure, 1996). The initial 1974 com-
plaint and 1976 decision by the administrative law judge alleged that Borden’s trademark
promotion and advertising had created an entry barrier and required Borden to license its
ReaLemon trademark to anyone wishing to compete with it. When the full Commission
reviewed this decision, it similarly concluded that ReaLemon represented a classic case
study of spurious and artificial product differentiation created by advertising and not by
the superiority of its product or its greater efficiency. The Commission declined to impose
compulsory licensing of the trademark, concluding that the remedy of preventing Borden
from engaging in selective price reductions was sufficient. The US Court of Appeals for
the Sixth Circuit upheld this decision over a strong dissent from Judge Cornelia Kennedy.
The advent of the Reagan Administration led the FTC to reverse itself and support
Borden’s request to vacate and remand the Sixth Circuit’s decision. The Commission’s
revised 1983 order repudiated the reasoning of its initial 1978 order, concluding instead
that trademark-driven product differentiation can promote competition, particularly by
reducing search costs.
Combined with the FTC’s abandonment of its Cereals case in 1981, which included
similar claims of artificial product differentiation through advertising, and the Maxwell
House case in 1984, the ReaLemon case is widely regarded as signaling the downfall of
the Chamberlinian critique of trademarks as a form of spurious product differentiation
(Mensch and Freeman, 1990; McClure, 1996; Weinberg, 2005). Since that time, the
economics of trademark has generally come to regard trademarks as a way to reduce
search costs and to promote investment (for the leading statement, see Landes and Posner,
1987). Some scholars have continued to advocate a Chamberlinian approach (Lunney,
1996b). Others have taken a more balanced approach, acknowledging that trademarks
can represent a positive source of product differentiation and information for consumers,
but looking for ways to limit the potential abuses of power (Barnes, 2009; Lemley and
McKenna, 2012).
VI. CONCLUSION
Product differentiation has represented a generative force in the economic analysis of IP.
Not only does it offer a better theoretical explanation for a number of market features, it
also provides a basis for formalizing both the access and incentives sides of the tradeoff
in a way that yields insights into their structural interrelationship. Product differentiation
also creates the possibility of excess entry. It also demonstrates the existence of circum-
stances under which strengthening IP protection can promote both access and incentives
simultaneously. This stands in stark contrast to the position that dominates existing
scholarship, which views these two considerations as being in inexorable tension.
It may seem counterintuitive that protection should be the greatest when high fixed
costs and low substitutability cause the market to become the most concentrated, but
this apparent paradox is resolved once one understands the complex manner in which
access and incentives interact with one another. In this sense, the differentiated products

approach captures some of the insights of classic property theory, which emphasizes the
importance of well-defined property rights in ensuring optimal investment and deploy-
ment. In so doing, it corrects for the blind spot that results when markets for IP are treated
as monopolies and allows for serious consideration of the role of short-run profits in
stimulating entry and promoting economic efficiency. At the same time, it moves beyond
classic property theory by identifying ways in which a property right can be too strong.
It bears noting that the differentiated products approach cannot completely resolve
the tension between access and incentives. The presence of a downward-sloping demand
curve renders some degree of deadweight loss endemic. In addition, the fact that perfect
price discrimination is impossible prevents rightsholders from appropriating the entire
surplus created by their IP. As a result, markets may exhibit a systematic tendency towards
having too few products. However, demand diversion makes it possible that the market
will produce the optimal number of products. Any such solution to the incentives side of
the tradeoff necessarily requires accepting a degree of inefficiency in terms of access. As
the theory of the comparative second-best aptly points out, the differentiated products
approach’s inability to generate first-best outcomes is not by itself sufficient to justify
rejecting it.
In addition, the differentiated products approach allows for a more nuanced analysis
by making it possible for policymakers to distinguish among different aspects of IP
protection. This represents a substantial improvement over the traditional approach,
which tends to represent all aspects of the strength of IP with a single variable and fails to
distinguish among different aspects of protection. In so doing, it identifies circumstances
under which efficiency might best be served by making the right large (in terms of surplus-
generating activities within its scope) and intense (in terms of the proportion of that
surplus that producers are able to appropriate), but narrow (in terms of how close another
product can come to an existing product without infringing the IP). Thus, the differenti-
ated products approach does not amount to a blanket endorsement for strengthening IP
protection. On the contrary, the resulting theory allows for a degree of subtlety that is
impossible under other approaches.
Although the application of product differentiation to IP has yielded some interesting
insights, considerable additional work remains to be done before it can be fully operation-
alized. As noted earlier, further work should incorporate elements of cumulative innova-
tion that take into account the extent to which existing IP serves as inputs to subsequent
products. Furthermore, the differentiated products approach should be broadened to
account for endogenous pricing as well as the preemptive strategies available when entry
is sequential and when firms can occupy more than one location.
The models also should consider the implications of relaxing the symmetrical prefer-
ences assumption, either by allowing for variations in the distribution of consumers across
the characteristics space or by allowing the extent to which particular products serve as
substitutes for other products to vary. Relaxing the symmetry assumption allows for the
possibility that the impact of entry by a new product will no longer be spread evenly
across all of the incumbents. Instead, it suggests that the entry will affect only some of the
products. This localization of competition has the effect of dividing the relevant market
into subsegments, with the overall competitiveness of each subsegment determined by the
size of the total surplus of the subsegment relative to the fixed cost. The lack of robust
competition within a subsegment may limit the extent to which entry can push price

towards marginal cost. It can also allow the ‘integer problem’ to arise simultaneously with
respect to multiple portions of the overall market, as the single ‘large economy’ is chopped
into a series of ‘small economies’ that each are capable of supporting sustainable profits.11
If these effects arise with respect to multiple subsegments, the combined adverse impact
may be quite substantial, although the resulting policy prescription may be the same as
when consumer preferences are assumed to be symmetric.
Countervailing considerations exist as well. The discussion of spatial competition
assumes that product characteristics vary along a single dimension, in which case
products compete exclusively with their two adjacent neighbors. The localized nature of
differentiated products competition can be substantially mitigated if spatial competition
occurs along more than one dimension.
The inherent ambiguity of the outcomes under product differentiation suggests that
the early fight between Chamberlin and the Chicago School may have been somewhat
overstated. It also underscores that the study of product differentiation would benefit
from more empirical work. Moreover, the policy instruments that follow from the dif-
ferentiated products approach are by their nature extremely contextual and do not lend
themselves to simple inferences. In addition, the interrelationships among the available
policy instruments make calibrating them simultaneously an extremely difficult empirical
exercise. The fact that the differentiated products approach is contextual and nuanced
should not obscure its basic analytical power. Indeed, the intuitions that the theory reveals
about the relationship between access and efficiency, the manner in which the various
aspects of IP protection interrelate, and the true relationship between IP and public goods
theory are sufficient to justify further inquiry.
REFERENCES
Abramowicz, Michael. 2004. ‘An Industrial Organization Approach to Copyright,’ 46 William and Mary Law
Review 33–125.
Archibald, G.C. 1961. ‘Chamberlin versus Chicago,’ 29 Review of Economic Studies 2–28.
Archibald, G.C., and G. Rosenbluth. 1975. ‘The “New” Theory of Consumer Demand and Monopolistic
Competition,’ 89 Quarterly Journal of Economics 569–90.
Bain, Joe S. 1956. Barriers to New Competition: Their Character and Consequences in Manufacturing Industries.
Barnes, David W. 2009. ‘One Trademark per Source,’ 18 Texas Intellectual Property Law Journal 1–54.
Barnes, David W. 2011. ‘Congestible Intellectual Property and Impure Public Goods,’ 9 Northwestern Journal of
Technology and Intellectual Property 533–63.
Baumol, William J. 1967. ‘Calculation of Optimal Product and Retailer Characteristics: The Abstract Product
Beath, John, and Yannis Katsoulacos. 1991. The Economic Theory of Product Differentiation. Cambridge:
Berry, Steven T., and Joel Waldfogel. 1999. ‘Free Entry and Social Inefficiency in Radio Broadcasting,’ 30 RAND
11
Interestingly, the market need not be divided into discrete subsegments in order for this effect
to occur. Variations in the density of firms across the product space can balkanize the industry
into a chain of ‘overlapping oligopolies,’ each comprised of a small number of firms engaged in
localized competition regardless of how many firms are operating in the overall market. This can
give rise to the same problems even in the absence of actual gaps in the product continuum (Kaldor,
1934; Eaton and Lipsey, 1989).

Beschorner, Patrick F.E. 2008. ‘Optimal Patent Length and Height,’ 35 Empirica 233–40.
Besen, Stanley M., and Sheila Nataraj Kirby. 1989. ‘Private Copying, Appropriability, and Optimal Copyright
Royalties,’ 32 Journal of Law and Economics 255–80.
Bhaskar, V., and Ted To. 2004. ‘Is Perfect Price Discrimination Really Efficient? An Analysis of Free Entry,’ 35
Bishop, Robert L. 1967. ‘Monopolistic Competition and Welfare Economics,’ in Robert E. Kuenne, ed.,
Monopolistic Competition Theory: Studies in Impact. New York: John Wiley & Sons.
Bonanno, Giacomo. 1987. ‘Location Choice, Product Proliferation and Entry Deterrence,’ 54 Review of
Economic Studies 37–45.
Borenstein, Severin. 1985. ‘Price Discrimination in Free-Entry Markets,’ 16 RAND Journal of Economics 380–97.
Bracha, Oren, and Talha Syed. 2014. ‘Beyond the Incentive-Access Paradigm? Product Differentiation and
Copyright Revisited,’ 92 Texas Law Review 1841–920.
Brander, James A., and Jonathan Eaton. 1984. ‘Product Line Rivalry,’ 74 American Economic Review 323–34.
Brown, Ralph S., Jr. 1948. ‘Advertising and the Public Interest: Legal Protection of Trade Symbols,’ 57 Yale
Buchanan, James M. 1965. ‘An Economic Theory of Clubs,’ 32 Economica (new series) 1–14.
Chamberlin, Edward Hastings. 1933. The Theory of Monopolistic Competition: A Re-Orientation of the Theory
of Value, 1st ed. Cambridge, MA: Harvard University Press.
Chamberlin, Edward Hastings. 1950. ‘Product Heterogeneity and Public Policy,’ 40 American Economic Review
85–92.
Chamberlin, Edward Hastings. 1957. Towards a More General Theory of Value. New York: Oxford University Press.
Comanor, William S., and Thomas A. Wilson. 1979. ‘The Effect of Advertising on Competition: A Survey,’ 17
Journal of Economic Literature 453–76.
d’Aspremont, C., J. Jaskold Gabszewicz, and Jean-François Thisse. 1979. ‘On Hotelling’s “Stability in
Competition,”’ 47 Econometrica 1145–50.
Denicolò, Vincent. 1996. ‘Patent Races and Optimal Patent Breadth and Length,’ 44 Journal of Industrial
Economics 249–65.
Dixit, Avinash K., and Joseph E. Stiglitz. 1977. ‘Monopolistic Competition and Optimum Product Diversity,’
67 American Economic Review 297–308.
Eaton, B. Curtis, and Richard G. Lipsey. 1976. ‘The Non-Uniqueness of Equilibrium in the Löschian Location
Model,’ 66 American Economic Review 71–93.
Eaton, B. Curtis, and Richard G. Lipsey. 1979. ‘The Theory of Market Pre-Emption: The Persistence of Excess
Capacity and Monopoly in Growing Spatial Markets,’ 46 Economica 149–58.
Eaton, B. Curtis, and Richard G. Lipsey. 1980. ‘Exit Barriers Are Entry Barriers: The Durability of Capital as
a Barrier to Entry,’ 11 Bell Journal of Economics 721–9.
Eaton, B. Curtis, and Richard G. Lipsey. 1989. ‘Product Differentiation,’ in Richard Schmalensee and Robert
D. Willig, eds., Handbook of Industrial Organization, vol. 1. New York: North-Holland.
Eaton, B. Curtis, and Myrna Holtz Wooders. 1985. ‘Sophisticated Entry in a Model of Spatial Competition,’ 16
Faulhaber, Gerald R. 2006. ‘File Sharing, Copyright, and the Optimal Production of Music,’ 13 Michigan
Telecommunications and Technology Law Review 77–113.
Feenstra, Robert C., and James A. Levinsohn. 1995. ‘Estimating Markups and Market Conduct with
Multidimensional Product Attributes,’ 62 Review of Economic Studies 19–52.
Friedman, Milton. 1953. Essays in Positive Economics. Chicago: University of Chicago Press.
Gallini, Nancy T. 1992. ‘Patent Policy and Costly Imitation,’ 44 RAND Journal of Economics 52–63.
Gilbert, Richard, and Carl Shapiro. 1990. ‘Optimal Patent Length and Breadth,’ 21 RAND Journal of Economics
106–12.
Goettler, Ronald L., and Ron Shachar. 2001. ‘Spatial Competition in the Network Television Industry’, 32
Hay, D.A. 1976. ‘Sequential Entry and Entry-Deterring Strategies in Spatial Competition,’ 28 Oxford Economic
Papers 240–57.
Hotelling, Harold. 1929. ‘Stability in Competition,’ 39 Economic Journal 41–57.
Johnson, William R. 1985. ‘The Economics of Copying,’ 93 Journal of Political Economy 158–74.
Judd, Kenneth L. 1985. ‘Credible Spatial Preemption,’ 16 RAND Journal of Economics 153–66.
Kaldor, Nicholas. 1934. ‘Mrs. Robinson’s “Economics of Imperfect Competition,”’ 1 Economica 335–41.
Kaldor, Nicholas. 1935. ‘Market Imperfection and Excess Capacity,’ 2 Economica 33–50.
Katz, Michael L. 1984. ‘Price Discrimination and Monopolistic Competition,’ 52 Econometrica 1453–71.
Klemperer, Paul. 1990. ‘How Broad Should the Scope of Patent Protection Be?,’ 21 RAND Journal of Economics
113–30.
Koenker, Roger W., and Martin K. Perry. 1981. ‘Product Differentiation, Monopolistic Competition, and Public
Policy’, 12 Bell Journal of Economics 217–31.

Lancaster, Kelvin. 1979. Variety, Equity, and Efficiency. New York: Columbia University Press.
Landes, William M., and Richard A. Posner. 1987. ‘Trademark Law: An Economic Perspective’, 30 Journal of
Lane, W.J. 1980. ‘Product Differentiation in a Market with Endogenous Sequential Entry,’ 11 Bell Journal of
Economics 237–60.
Lemley, Mark A., and Mark P. McKenna. 2012. ‘Is Pepsi Really a Substitute for Coke? Market Definition in
Antitrust and IP,’ 100 Georgetown Law Journal 2055–117.
Liebowitz, S.J. 1986. ‘Copyright Law, Photocopying, and Price Discrimination,’ 8 Research in Law and
Economics 181–200.
Lunney, Glynn S., Jr., 1996a. ‘Reexamining Copyright’s Incentives-Access Paradigm,’ 49 Vanderbilt Law Review
483–656.
Lunney, Glynn S., Jr. 1996b. ‘Trademark Monopolies,’ 48 Emory Law Journal 367–486.
Mankiw, N. Gregory, and Michael D. Whinston. 1986. ‘Free Entry and Social Inefficiency,’ 17 RAND Journal
Matutes, Carmen, Pierre Regibeau, and Katharine Rockett. 1996. ‘Optimal Patent Design and the Diffusion of
Innovations,’ 27 RAND Journal of Economics 60–83.
Maurer, Stephen M., and Suzanne Scotchmer. 2002. ‘The Independent Invention Defence in Intellectual
Property,’ 69 Economica 535–47.
McClure, Daniel M. 1996. ‘Trademarks and Competition: The Recent History,’ 59 Law and Contemporary
Problems 13–43.
Mensch, Elizabeth, and Alan Freeman. 1990. ‘Efficiency and Image: Advertising as an Antitrust Issue,’ Duke
Meurer, Michael J. 2001. ‘Copyright Law and Price Discrimination,’ 23 Cardozo Law Review 55–148.
Neven, Damien J. 1987. ‘Endogenous Sequential Entry in a Spatial Model,’ 5 International Journal of Industrial
O’Donoghue, Ted, Suzanne Scotchmer, and Jean-François Thisse. 1998. ‘Patent Breadth, Patent Life, and the
Pace of Technological Progress,’ 7 Journal of Economics & Management Strategy 1–32.
Prescott, Edward C., and Michael Visscher. 1977. ‘Sequential Location among Firms with Foresight,’ 8 Bell
Robinson, Joan. 1933. The Economics of Imperfect Competition. London: Macmillan & Co.
Salop, Steven C. 1979. ‘Monopolistic Competition with Outside Goods,’ 10 Bell Journal of Economics 141–56.
Samuelson, Paul A. 1954. ‘The Pure Theory of Public Expenditure,’ 36 Review of Economics and Statistics
387–89.
Samuelson, Paul A. 1958. ‘Aspects of Public Expenditure Theories,’ 40 Review of Economic Studies 332–8.
Samuelson, Paul A. 1967. ‘The Monopolistic Competition Revolution,’ in Robert E. Kuenne, ed., Monopolistic
Competition Theory: Studies in Impact. New York: John Wiley & Sons.
Scherer, F.M. 1970. Industrial Market Structure and Economic Performance. Chicago: Rand McNally & Co.
Scherer, F.M. 1972. ‘Nordhaus’ Theory of Optimal Patent Life: A Geometric Reinterpretation,’ 62 American
Scherer, F.M. 1976. ‘Predatory Pricing and the Sherman Act: A Comment,’ 89 Harvard Law Review 869–90.
Schmalensee, Richard. 1978. ‘Entry Deterrence in the Ready-to-Eat Breakfast Cereal Industry,’ 9 Bell Journal
Schmalensee, Richard. 1979. ‘On the Use of Economic Models in Antitrust: The ReaLemon Case,’ 127
University of Pennsylvania Law Review 994–1050.
Spence, Michael. 1976a. ‘Product Differentiation and Welfare,’ 66 American Economic Review 407–14.
Spence, Michael. 1976b. ‘Product Selection, Fixed Costs, and Monopolistic Competition,’ 43 Review of
Spulber, Daniel F. 1979. ‘Non-Cooperative Equilibrium with Price Discriminating Firms,’ 4 Economics Letters
221–7.
Spulber, Daniel F. 1981. ‘Spatial Nonlinear Pricing,’ 71 American Economic Review 923–33.
Stigler, George J. 1949. ‘Monopolistic Competition in Retrospect,’ in Five Lectures on Economic Problems.
London: Longmans, Green & Co. Ltd.
Stiglitz, Joseph E. 1977. ‘The Theory of Local Public Goods,’ in Martin S. Feldstein and Robert P. Inman, eds.,
The Economics of Public Services. London: Macmillan & Co.
Stole, Lars. 2007. ‘Price Discrimination in Competitive Environments,’ in Mark Armstrong and Robert Porter,
eds., Handbook of Industrial Organization, vol. 3. New York: North-Holland.
Tandon, Pankaj. 1982. ‘Optimal Patents with Compulsory Licensing,’ 90 Journal of Political Economy 470–86.
Tiebout, Charles M. 1956. ‘A Pure Theory of Local Expenditures,’ 64 Journal of Political Economy 416–24.
Varian, Hal R. 2010. Intermediate Microeconomics: A Modern Approach, 8th ed. New York: W.W. Norton & Co.

Waterson, Michael. 1990. ‘The Economics of Product Patents,’ 80 American Economic Review 860–69.
Weinberg, Harold R. 2005. ‘Is the Monopoly Theory of Trademarks Robust or a Bust?,’ 13 Journal of
Yoo, Christopher S. 2004. ‘Copyright and Product Differentiation,’ 79 New York University Law Review 212–80.

10. Price discrimination and intellectual property
Michael J. Meurer*and Ben Depoorter**
12
Contents
I. Introduction
II. Price Discrimination: Overview
A. Pigou’s Classification
B. Product Differentiation and Second Degree Price Discrimination
C. Differences between Second and Third Degree Price Discrimination
D. Tying, Merchandising, and Bundling
III. When is Price Discrimination Profitable?
IV. Social Welfare Issues
V. A Social Welfare Analysis of Price Discrimination in the IP Context
A. Copyright
B. Patents
VI. The Many Forms of Price Discrimination by Way of IP Rights
A. Geographic Price Discrimination
B. Restrictions on Type of Use
C. Intensity of Use: Tying, Merchandising, and Bundling
D. Sharing
VII. Conclusion
Bibliography
I. INTRODUCTION
This chapter provides an overview of the literature on price discrimination with special
attention to the relationship between intellectual property (IP) rights and price discrimi-
nation. There are several reasons why a serious student of intellectual property law should
understand the theory of price discrimination. First, although many IP owners practice
price discrimination, these practices come in many guises that may not be apparent to a
casual observer. For instance, patent licensors sometimes price discriminate across uses
of the patented technology and in different geographic regions. Similarly, trademarks
support product differentiation and brand loyalty in ways that often facilitate price
** Abraham and Lillian Benton Scholar and Professor of Law, Boston University School
of Law. I thank workshop participants at the University of California-Berkeley for their helpful
comments.
** Max Radin Chair and Distinguished Professor of Law, University of California, Hastings
College of the Law; Affiliate Scholar, Stanford Law School, Center for Internet and Society;
CASLE, Ghent University, Belgium. Contact: depoorter@uchastings.edu.
281

iscrimination. Trademark lawsuits have been used (not always successfully) to support
d
price discrimination across geographic regions, or within aftermarkets. Price discrimina-
tion is also widespread in the market for copyrighted works. Movie tickets, magazine
and book prices, computer software prices, and music performance licenses are subject
to price discrimination strategies. Magazine subscribers pay much less than newsstand
buyers, movie theaters offer cheaper tickets for daylight or mid-week showings, software
is sold at a discount to students and educators, and establishments offering public
music performances pay widely varying fees for identical blanket licenses. Second, price
discrimination is becoming increasingly common in markets for digital copyrighted works
(Shapiro and Varian, 1999). Third, intellectual property laws influence the profitability
and practice of price discrimination in many ways. As we will demonstrate below, intel-
lectual property law offers unusually fertile ground for analysis of policies that facilitate
or discourage price discrimination.
Roughly speaking, price discrimination occurs when a seller sets more than one price
for a good or service. More precisely, Stigler (1987) specifies that price discrimination
arises when the ratio of prices differs from the ratio of marginal costs for the products
offered by a seller.1 Price discrimination is evident, for instance, to air travelers who
may find the passenger next to them paid a different price to travel the same route.
Price discrimination is less evident when the price for two products is the same but the
marginal cost of providing them differs, or when multiple products are sold together, or
when differentiated products are offered at prices that diverge from the ratio identified by
Stigler. In what follows we consider various strategies used by IP owners to achieve price
discrimination.
We keep formal economics to a minimum and recommend Stole (2007) to readers
seeking a more comprehensive and rigorous review of the price discrimination literature.
Readers may also find the chapter by Christopher Yoo on product differentiation in this
volume to be a helpful complement to this chapter, because sellers sometimes differentiate
their products to discriminate across different market segments.
II. PRICE DISCRIMINATION: OVERVIEW
Successful price discrimination relies on sellers sorting buyers in terms of willingness to

pay; those willing to pay more are charged more. Discrimination will not succeed if the
disfavored buyers can make purchases from, or otherwise gain access to, products that are
sold at a lower price to favored buyers. Thus, a price discriminator must block arbitrage
that could erode the price differential between favored and disfavored buyers. This section
1
Price discrimination is challenging to define. Prohibitions of price discrimination in antitrust
laws provide an implicit definition. Antitrust laws generally operate on the intuitive notion that
price discounts are discriminatory when they are not justified by cost differentials. Such a defini-
tion does not match up well with the economist’s definition of price discrimination (Baker, 1997,
pp. 177 ff.). Economists point out that charging a uniform price can also be discriminatory (e.g.,
when delivery costs vary, a uniform delivered price discriminates in favor of distant customers) and
that product quality or other differences between sales should be accounted for in the definition.

Price discrimination and intellectual property 283
outlines the various forms of price discrimination, using examples drawn from markets in
which intellectual property often plays an important role.
A. Pigou’s Classification
Pigou’s (1932) classic description of price discrimination distinguishes three practices.

In third degree price discrimination, price differentials are tied to a characteristic of a
buyer that is correlated with the buyer’s valuation.2 For instance, movie exhibitors offer
senior discounts and thereby use the age of the buyer to discriminate. They assume that
senior citizens have a weaker demand than other buyers do. This practice requires that
sellers can observe buyers’ characteristics.
In second degree price discrimination, the price differentials are tied to the choices by
the buyer. The pattern of discrimination reflects the seller’s belief that a certain choice will
be made by a low valuation buyer and a different choice will be made by a high valuation
buyer. For example, movie exhibitors offer a discount for Tuesday movies to sort between
movie patrons who are flexible about when they see a movie and those who are not. This
hidden characteristic of flexibility is supposedly correlated with less intense demand for
movies. Second degree price discrimination is used when a seller cannot directly observe
buyers’ characteristics.3
Overall, sellers may care about a mix of observable and hidden buyer characteristics
and may engage in second or third degree price discrimination or a mix of the two.
In first degree price discrimination the seller knows or learns the exact valuation of all
buyers. This of course is an idealized benchmark.
2
Third degree price discrimination does not always depend on variable characteristics like
age. It can also depend on buyer attributes that cannot be changed easily. Textbook publishers
discriminate between college and other bookstores. Movie distributors discriminate based on the
size and location of a theater. These attributes reflect past choices of buyers that will not be altered
just to avoid price discrimination.
3
The standard treatment of second degree price discrimination from information econom-
ics studies the characteristics of an optimal pricing mechanism and product design choice for a
monopolist facing two types of customers. One type of customer has a higher marginal valuation
of quality than the other. The results show that the socially optimal quality is offered to the high
valuation customer. The low valuation customer is provided with quality that is less than the social
optimum. Furthermore, the low valuation customer is left indifferent between participating in the
market or not. The high valuation customer gains positive surplus. Since the seller cannot distin-
guish high and low valuation customers he or she has to offer one price and quality combination
that attracts one type of customer and another combination that will attract the other type. An
artful choice of quality and price pairs solves both the arbitrage and measurement problem. The
high valuation customers prefer higher quality despite the higher price; the low valuation customers
prefer lower quality and price. If the seller could distinguish high and low valuation customers and
practice third degree price discrimination, then the seller would choose the socially optimal quality
level for both types of customers. Under second degree price discrimination the quality is degraded
to low valuation customers to make it easier to sort the two types. When the quality gap is large,
the seller can raise the price differential and gain a higher mark-up from the more profitable market
niche containing the high valuation customers.

B. Product Differentiation and Second Degree Price Discrimination
The movie exhibitor in our senior discount example of price discrimination offered every
buyer an opportunity to see the same movie, but at different prices. Implicit in Stigler’s
definition of price discrimination is the possibility of a price discriminator that offers two
or more versions of its product for sale. Offering multiple product versions may increase
the profitability of second degree price discrimination. Price discrimination occurs when
the seller marks up the price of one version (over marginal cost) more than the other.
Thus, sellers may combine price discrimination and product differentiation. The price
discriminating seller allows customers to choose the high quality or low quality version
of the product and sort themselves into a high value segment and a low value segment (or
possibly multiple segments if there are more than two products offered).
Products can be differentiated in many ways. The term vertical product differentiation
is used to describe products that range from low to high quality, for example, limited and
full-feature software. The term horizontal product differentiation is used to describe other
types of product differentiation in which buyers have different preferences over different
products’ attributes. It is easy to see that vertical product differentiation may be helpful
for sorting buyers into high and low value segments. Horizontal product differentiation
may likewise be helpful in sorting if higher valuation buyers happen to prefer a particular
set of attributes. Movies exhibited at noon may be more valuable to some customers than
movies exhibited in the evening, but most customers prefer the latter, and form the high
value segment that typically bears the higher price.
Different physical attributes can be used to differentiate products, but products can also
be differentiated in terms of delivery date, quantity, and contractual restrictions.
Delivery date differentiation is used frequently in markets for copyrighted works, such
as movies and books. A movie viewer can choose between a first-run or second-run show-
ing in the theater, DVD sales, pay-per-view, premium cable, streaming services, free cable,
or free broadcast. Similarly, the price of novels declines over time. Eager readers pay a
higher price for hard cover books, and more patient readers wait for the later publication
of the cheaper soft cover version. For both movies and novels there are quality as well
as timing differences that roughly correspond with the price and release date. Viewing
quality is higher in a theater than on television, and first-run theaters are usually more
pleasant than second-run theaters. Pay-per-view, premium cable, and videotapes do not
have commercials. Hard cover books are more durable, more attractive, and have larger
print. The price of these choices usually declines with the viewing date. No arbitrage is
possible. Buyers cannot buy cheap and travel back in time to sell the item at a premium.
Similarly, software is often offered at a lower price when subjected to consumer or
educational use restrictions, while identical software without those restrictions is sold at
a higher price. Patent licenses often entail price differences depending on the type of use,
include geographic use restrictions, single or multiple uses, or with or without resale rights.
Note that arbitrage is not possible if the products are differentiated by physical attrib-
utes. Arbitrage is possible, however, against quantity restrictions if a portion of a purchase
can be resold. Arbitrage is possible also when differentiation is created through contract
terms—by violation of the restrictions. For example, a buyer might violate an educational
use restriction on software by using the software for business tasks. Alternatively, such a
buyer might resell it to a business user.

C. Differences between Second and Third Degree Price Discrimination
Three main principles are paramount to an optimal second degree price discrimination
strategy. First, prices respond to marginal cost. If the product with the more attractive
attributes is costlier, then its price should be higher. This relationship between price and
marginal cost applies whether or not a seller price discriminates. Second, market power
allows a seller to mark up prices above marginal cost. A price discriminating seller is
especially interested in marking up the price to the high valuation segment of the market.
Third, prices are subject to a sorting constraint. If a seller is too aggressive and selects an
excessive mark-up for the more attractive product, then high valuation buyers will switch
to the other product. The sorting constraint keeps the difference between the two prices
small enough so that buyers will sort themselves.
Note that there is a tension between the second and third principle. The profitability
of second degree price discrimination is limited because the sorting condition restricts
the mark-up that can be levied against the high valuation buyers. The sorting constraint
does not apply to third degree price discrimination, however, since the buyer characteristic
used for sorting is observable to the seller. Thus, third degree price discrimination is more
profitable given the superior information available to the seller, all else being equal.4
Besides the sorting constraint, the two forms of price discrimination also differ in terms
of their relationship to arbitrage. Resale can be used to arbitrage against both types of
price discrimination.5 But with second degree price discrimination there is an additional
method of arbitrage; a low-price buyer can engage in personal arbitrage. Personal arbi-
trage means that the buyer recovers some or all of the benefit of the high-priced product
by ‘modifying’ the low-priced product. For instance, sophisticated buyers may circumvent
technology that limits how a product is used—for example, refilling printer cartridges or
modifying software to unlock restricted features. Such uses and modifications by buyers
may violate license terms or intellectual property rights. If the violation is undetected,
however, then the buyer gets the benefit of a low price and unrestricted use.
D. Tying, Merchandising, and Bundling
Sellers can also implement second degree price discrimination by bundling or tying
products together, or simply through the exclusive right to sell complementary products.
Specifically, tying arises when a seller conditions the sale of one product or service (the
4
This comparison holds if arbitrage is not possible. The comparison may be reversed if
arbitrage is more difficult to control under third degree price discrimination compared to second
degree.
5
Many copyrighted products subject to second and third degree price discrimination are
services. Resale is difficult or impossible if the product is a service. For example, a buyer cannot
purchase an extra viewing of a movie and resell it. It may be possible to resell the ticket, but
inexpensive enforcement methods can limit that type of arbitrage. If the product is a good, then
resale is more of a problem. Resale is often used to arbitrage against geographic (third degree) price
discrimination. Resale can arbitrage against second degree price discrimination based on quantity
or contract restrictions. A high volume buyer can arbitrage quantity discounts by purchasing extra
units at a discounted price and reselling to low volume buyers. A buyer can arbitrage contractual
restrictions by purchasing a low price product and flouting the resale restriction.

tying good or service) on the purchase of a different product or service (the tied sale).
Some of the most prominent tying disputes in antitrust law involve patented products.6
Also, much of the case law involving the doctrine of patent misuse involves tying.
Bundling arises when a seller sells products X and Y as a bundle but not separately (at least
for certain customers). Bundling of content by media companies and multiple software
products by software publishers is common and sometimes controversial. Merchandising
refers to the sale of relatively low value products that are complementary to a high value
copyrighted work, franchise, or a high value trademark. Sports teams sell hats, T-shirts,
and so on. Movie copyright owners sell toys, posters, and the like.
Tying, bundling, and merchandising might be used to implement price discrimination,
but could serve other goals as well, which could either have pro-competitive or anti-
competitive effects. Tied sales may be used to measure (often the term ‘meter’ is used
instead of measure) the intensity of use of the tying product by observing the volume
of purchases of the tied product (Elhauge and Nalebuff, 2017). The mark-up on the
tied product results in discrimination against high intensity users of the tying product.
Similarly, merchandising can be used to meter intensity of support for a sports team,
music group, or movie. Mark-ups on merchandise can implement price discrimination
against loyal fans. Pricing bundles is relatively complicated. Bundle pricing helps sellers
extract greater consumer surplus when consumers disagree over which product in a bundle
is more valuable. Sellers can keep customers in the market for products they do not value
as much, and at the same time price more aggressively on products they do value highly.
III. WHEN IS PRICE DISCRIMINATION PROFITABLE?
Casual intuition might suggest that price discrimination is always profitable because
discriminatory prices allow a seller to capture a greater share of a consumer surplus than
uniform prices. This intuition ignores two factors that could make price discrimination
unprofitable. First, when there are multiple sellers in a market it is possible for price dis-
crimination to intensify competition between sellers and drive down profits. Second, it is
costly to segment customers and block arbitrage, and these costs may exceed the benefits
of price discrimination (Leeson and Sobel, 2008).7
The profitability of price discrimination changes over time as technologies evolve. The
impact of new technologies can either strengthen or weaken a seller’s opportunity to
engage in price discrimination. For instance, virtual private network (VPN) services hide
the location of internet users, making it more difficult for sellers to engage in geographic
price discrimination. For the most part, however, since the 1990s, digital technologies have
6
See, e.g., Illinois Tool Works Inc. v. Independent Ink, Inc., 547 U.S. 28 (2006); U.S. Steel
Corp. v. Fortner Enterprises, Inc., 429 U.S. 610 (1977); Jefferson Parish Hosp. Dist. No. 2 v. Hyde,
466 U.S. 2, 14 (1984).
7
Additionally, experimental work suggests that a fairness constraint limits the potential scope
of third degree price discrimination (Englmaier et al., 2012). More research is needed to examine
the robustness of this finding, however. Also there is an open question whether fairness is less
salient, and therefore less constraining when discrimination is hidden, for example in the case of
metering via tied sales.

advanced the opportunities for sellers to engage in price discrimination. For instance, by
measuring the frequency and duration of the use of a work by consumers, metering and
on-line licensing technologies enable publishers to obtain much more precise measures
of the demand for their products (Goldstein, 1994, p. 200). The availability of this
information opens the door to highly refined price discrimination (Netanel, 1996, p. 295;
Bell, 1998; Meurer, 1997, p. 880; Manta and Olson, 2015). Furthermore, technological
measures such as digital rights management (DRM) may limit the alteration, copying, or
transfer of digital products (Hughes, 2016). These technologies restrict the opportunity
for buyers to circumvent price discrimination practices. Similarly, the development of
sterile seeds can prevent seeds from sprouting in consecutive years (Jasanoff, 2013),
limiting reproduction and arbitrage opportunities in the market for genetically modified
seeds. Mortimer (2008b) finds that bar code scanners and the internet increased price
discrimination in the video rental industry. Also, the increased collection of personal data
and the emergence of big data may facilitate improved customer sorting and increase the
potential profits from discrimination—unless privacy regulation blocks this approach to
price discrimination.
The profitability of price discrimination also changes as the strength of intellectual
property law waxes and wanes. At its zenith, patent law gives an inventor a monopoly in
a market—best illustrated by patented blockbuster drugs. During the term of the patent,
the inventor can exclude competing drug makers who could disrupt price discrimination
by catering to discriminated-against customers. It is important to recognize that most
patents, and other forms of intellectual property, do not lead to monopoly and might
not even create much market power. But monopoly is certainly not a precondition to
profitable price discrimination. Demsetz (1970) explores the use of price discrimination in
competitive markets.8 Thisse and Vives (1988), Corts (1998), Holmes (1989), Armstrong
and Vickers (2001), and others model various forms of price discrimination in oligopoly
settings.
Rather than market power, intellectual property law more often influences the profit-
ability of price discrimination by helping or hindering sellers to block arbitrage and sort
customers. Intellectual property law may allow sellers to control supply chains in ways that
minimize arbitrage, and directly control uses and resale by end users. We provide many
examples in the sections that follow.9
IV. SOCIAL WELFARE ISSUES
Economists usually examine the social welfare effects of price discrimination by compar-
ing it to uniform pricing. The main focus is on: static effects on output and misallocation
8
Levine (2000) analyses use of price discrimination to recover sunk costs in a competitive
environment, and Hovenkamp (1999) discusses the limited market power required to engage in
9
Market power itself may also increase a seller’s ability to limit arbitrage. One avenue to limit
arbitrage is self-help, whereby a seller punishes users and distributors engaged in arbitrage. Because
measures to deter arbitrage are unpopular among distributors and customers, a seller with market
power, especially a monopolist, is more likely to be willing and able to incur this cost.

of goods across consumers; dynamic effects on product design, entry, and innovative
investment; and distributional effects.
Third degree price discrimination may increase efficiency by increasing output, espe-
cially when price discrimination results in new buyers entering a market who would be
foreclosed from the market by a high uniform price (Varian, 1985). In these circumstances,
discrimination can leave output in the ‘stronger’ market unchanged while expanding
output in the ‘weaker’ market that would otherwise not be served (Robinson, 1933).10
However, output expansion is only a necessary but not a sufficient condition for welfare
improvement; output must expand more in the weak market than it contracts in the strong
one because each unit is more highly valued in the latter (Schmalensee, 1981).
To elaborate, price discrimination causes the marginal rates of substitution to differ
across buyers, and disfavored buyers have a higher marginal valuation than favored buyers.
Disfavored buyers would get a greater marginal benefit from one more unit than favored
buyers would lose from giving up one unit. This presents an opportunity to increase total
surplus via trade, but this opportunity is lost because arbitrage is blocked. This foregone
opportunity represents a source of allocative loss. Overall then, any efficiency gains
from an increase in output may be offset by efficiency losses caused by differences in the
marginal rates of substitution between favored and disfavored buyers (Viscusi et al., 1992,
pp. 279–83; Tirole, 1988, pp. 137–9).
The static welfare effects of second degree price discrimination are likewise unclear,
but perfect price discrimination provides unambiguous gains in allocative efficiency. All
consumers can be served to the point that marginal utility equals marginal cost, and thus,
marginal rates of substitution are equalized and no consumers are rationed. Uniform
monopoly pricing might still have a social welfare advantage over perfect price discrimina-
tion, but that depends on the dynamic and distributional effects that we turn to now.
The additional profits that may be created by price discrimination may induce socially
valuable investment or socially harmful rent dissipation. Many commentators have noted
that price discrimination may augment the reward to authors and inventors and induce
a valuable boost in creative or innovative investment (for example, Demsetz, 1970).
Though of course it is possible for this reward to be too large and induce overinvestment.
Furthermore, by increasing rents, price discrimination may induce socially wasteful
rent-seeking investments. Firms may engage in socially wasteful lobbying, litigation, and
entry deterring practices that work to preserve their right and ability to engage in price
discrimination (Posner, 1986; Meurer, 1997).
Patents or copyrights may create entry barriers and allow firms to earn positive eco-
nomic profits, but this is not always the case.11 When entry barriers are absent, increased
rents from price discrimination will drive entry into the market until profits fall to zero.
This entry may be socially valuable because of added product variety, or socially harmful
if entrants make duplicative investments to enter the market without providing significant
social benefits in the form of greater variety or quality improvement (Katz, 1984).
10
Additionally, discounts and other forms of two-part pricing can increase social welfare
(Tirole, 1988, pp. 145–6).
11
Intellectual property laws raise the cost of creation and innovation when a new work depends
on older works in such a way that permission is required to create or exploit the new work (Landes
and Posner, 1989, p. 332; Lemley, 1997; Scotchmer, 1991; Gordon, 1992).

Finally, price discrimination may give rise to various social costs from measuring
consumer attributes and blocking arbitrage. Direct social losses arise from the cost of
identifying customer attributes and choices, writing and enforcing contracts that prevent
arbitrage, and designing special distribution systems.12 Indirect costs arise from the sort-
ing constraints associated with second degree price discrimination13 by which a seller may
sort buyers by offering a menu of delivery dates, qualities, quantities, and permissible uses
of the product (Meurer, 1997, 2001). If delivery is the choice variable, then the sorting cost
arises from the delivery delay. For many products the optimal delivery policy is immediate
availability. If quality is the choice variable, then the sorting cost arises from degraded
quality for lower valuation buyers14 (for instance, software with code that disables certain
features; see Deneckere and McAfee, 1996). If quantity is the choice variable, then the
sorting cost comes from rationing of low valuation buyers (Maskin and Riley, 1984). The
seller designs quantity discounts in such a way that low valuation buyers purchase less
than they would under third degree price discrimination. The rationing is introduced to
ease the sorting constraint and to enable the seller to charge a higher price to the high
valuation segment of consumers. If use restriction is the choice variable, then the sorting
cost is the restriction itself. Consumer surplus is lost to low valuation buyers who are
constrained in their use of a product. The sorting constraint imposes an implicit cost on
the seller because it restricts the freedom of the seller to select the optimal attributes. With
third degree price discrimination the seller can avoid sorting costs and select the optimal
attributes for the two classes of consumers independently.
Besides efficiency, social welfare depends on the distributional effects of price dis-
crimination. Generally, the impact of price discrimination is felt differently across various
potential buyers. Buyers with higher valuations (or more precisely lower elasticity) tend
to lose and buyers with lower valuations (higher elasticity) tend to gain. High elasticity
customers, often the poorer segments of the consumer base, are especially likely to benefit
from third degree price discrimination (Tirole, 1988, pp. 137–9).
Two special cases are often discussed in the literature. First, perfect price discrimination
transfers all consumer surplus to a monopoly seller. Thus, efficiency is in tension with
a pro-consumer distributional goal, and social welfare may be higher under a uniform
monopoly price if the social welfare function places sufficient weight on consumer surplus
as compared to profit. Second, there are instances of imperfect price discrimination that
are Pareto-improvements over a uniform monopoly price.15 For example, if a uniform
monopoly price excludes a low valuation segment from a market, and optimal third
degree price discrimination keeps the original monopoly price in place for the high value
12
Implementation costs are most problematic when there are entry barriers and few sellers.
Implementation costs are apt to decline as the number of discriminating firms in a market grows
(Stole, 1995).
13
By contrast, under third degree price discrimination, a monopolist seller treats each market
niche separately and no product design or sale distortions are caused. The seller extracts rents solely
on the basis of pricing differences.
14
Economic theory shows that a seller should set the optimal attribute for the high valuation
consumers and a suboptimal attribute for the low valuation consumers. This inefficiency creates a
hidden social cost of sorting.
15
Conditions such that all prices fall under discrimination or rise only for the relatively well-
off are discussed in Stole (2003) and Schwartz (1990).

segment of the market and offers a lower price that induces purchases by the low value
segment, then high valuation consumers are indifferent between the two schemes, low
valuation consumers are better off with price discrimination, the seller is better off with
price discrimination, and allocative efficiency is improved.
The social welfare comparison between perfect price discrimination and uniform
monopoly pricing generates interesting insights but other comparisons are also useful.
An intellectual property scholar may be interested in a welfare comparison between two
different forms of price discrimination that are practiced before and after a substantive
change in intellectual property law. In the following sections we consider several examples
of how a change in the law may lead a seller to switch the form of price discrimination it
practices because arbitrage or metering has become harder or easier.
V. A SOCIAL WELFARE ANALYSIS OF PRICE DISCRIMINATION

IN THE IP CONTEXT
As discussed in detail in the following section, copyright and patent law contain a wide
range of doctrines that facilitate or impede price discrimination.16 In this section we
review the law and economics literature that abstracts from doctrinal details and consid-
ers broad normative questions about the interaction between price discrimination and
copyright or patent protection.
A. Copyright
One strand of literature takes a generally positive view toward price discrimination.
These authors focus on the power of digital technologies to better measure preferences
and limit arbitrage and thereby make price discrimination more attractive to sellers of
copyrighted works.17 The price discrimination optimists emphasized two social benefits
that might flow from finer grained price discrimination in digital markets. First, increased
product variety and a more detailed menu of prices could improve access of poor or
other underserved consumers (Fisher, 1997).18 Second, assuming the price discrimination
enabled by digital technology would increase profit,19 this would bolster the incentives
to create (Hardy, 1996; Besen and Raskind, 1991, p. 5). Fisher noted that ‘we are getting
much more bang for our buck—a much larger incentive for creative activity per unit of
16
The first studies to recognize the link between copyright doctrine and price discrimination
include Fisher (1998), Besen and Kirby (1989), Besen et al., (1992), and Lunney (1996). For a
comprehensive analysis of the relationship between copyright law and price discrimination, see
Meurer (2001).
17
The pioneering writing by Goldstein envisioned a ‘celestial jukebox’ from which consumers
could listen to any song in exchange for a micropayment (1994).
18
On the increase of output and new buyers entering a market that would otherwise be
foreclosed by a high uniform price, see generally, Viscusi et al., 1992, pp. 282–3. In the context of
software sales, see ProCD, Inc. v. Zeidenberg, 86 F.3d 1447, 1449 (7th Cir. 1996); Manta and Olson
(2015). On the positive impact of price discrimination on developing countries that publishers
might otherwise ignore, see Netanel (1998, p. 224).
19
Empirical evidence can be found in Leslie (2004) and Mortimer (2008a, 2008b).

social cost. Such a system of rules, applied to the Internet, should move us faster than
a copyright-based system in the direction of an informational society and rich artistic
tradition’ (Fisher, 1997, p. 1240).20
Besides making price discrimination more profitable, certain new technologies reduced
transaction costs and diminished one of the justifications for the fair use defense and
certain statutory limits on the rights of copyright owners. New technologies also made
it easier to limit resale of copyrighted content, and copyright owners bolstered technical
barriers to resale with contractual limits.21 Several scholars applauded these developments
and argued that copyright law should accommodate efforts by sellers to block arbitrage
in digital markets (Bell, 1998; Friedman, 1998; Fisher, 1997; Merges, 1997; O’Rourke,
1997; Kitch, 1999), including scaling back the scope of fair use (Bell, 1998; Kitch, 1999).22
Critics contend the alleged benefits of price discrimination arise in market conditions
very different from those that apply to creative and informational works (Gordon, 1998,
p. 1389; Boyle, 2000) or apply only to specific markets (Lunney, 2008). A monopolist
with the capacity to price discriminate might be preferable to a monopolist that charges
a single, uniform price. But perhaps the correct comparison would be with a system that
permits lawful free copying and a resulting range of prices. In other words, highlighting
the benefits of a monopoly with price discrimination compared to monopoly without
price discrimination is misleading (Gordon, 1998, pp. 1383–4).
This second strand of literature takes a generally skeptical view of the social benefits of
price discrimination.23 Digital technology, especially the growth of the internet, enabled
massive amounts of unauthorized copying and dissemination of copyrighted content.
Courts and legislatures responded by expanding the scope of copyright to combat
piracy. Critics complained that anti-piracy reforms also promoted socially harmful price
discrimination.
Cohen (1998, 2000) emphasizes copyright law’s limitations and exemptions generate
public benefits that likely would be underproduced by a system of centralized, strictly
market-based control. She argues that copyright law is balanced to serve nonmonetizable
and distributional concerns that are central to the creative and social progress (Cohen,
1998, p. 1128; Merges, 1997).
Full-fledged price discrimination might upset the existing balance of rights and the
ecosystem of tolerated uses that have been identified in various writings (Boyle, 2000;
Tehranian, 2007; Wu, 2008; Balganesh, 2013). In doing so, cumulative creation might
slow down if a legally sanctioned price discrimination regime, bolstered by technology
and contract rights, reduces access to information that was previously available due to the
limited enforceability of contracts of adhesion, first sale, fair use, and so on (Boyle, 2000,
p. 2032; Cohen, 1998, p. 1809).
20
Fisher also suggests a series of reforms designed to preserve the public benefits of specified
types of access and/or use (Fisher, 1998). See also Merges (1997).
21
For a critique of these tactics, see Perzanowski and Schultz (2001) (misalignment with goals
of copyright law); Perzanowski and Hoofnagle (2016) (consumer protection).
22
Others point out, however, that such technologies do not eliminate bilateral monopoly
(Merges, 1997) or anticommons issues (Depoorter and Parisi, 2002).
23
For a discussion of the diverging views on price discrimination among copyright optimists
and pessimists, see Netanel (1996).

Another social welfare loss of enhanced price discrimination that should be considered
is the elimination of free access to previously non-excludable aspects of information works.
This might reduce the variety and diversity of works produced (Cohen, 1998; Benkler,
1999). Overall, the loss of free access ‘may or may not outweigh the welfare gained from
access to newly excludable aspects of the work at a lower price than previously available’
(Benkler, 2000, p. 2076). Free access might be particularly relevant for (1) works with uses
for which it is hard to predict the value ex ante; (2) high positive externality productive
uses; and (3) public domain works produced for free distribution (Benkler, 2000).
Moreover, increasing the reward to copyright holders by promoting price discrimina-
tion may distort the pattern of investments in creative products. For instance, if price
discrimination is easier for copyrighted works than for other products it may lead to
overinvestment in copyrighted works (Lunney, 1996, p. 633; 2008) or the types of works
that enable price discrimination or allow a producer to cover the fixed costs of price
discrimination (Baker, 1997, p. 344).24 Also, since price discrimination may lead a firm to
produce multiple products to implement discrimination through quality differentiation,
bundling, and tying, multiproduct firms will do better when legal obstacles to price
discrimination are removed (Baker, 1997, p. 346).
On a more granular level, Meurer (2001) distinguishes situations where price discrimi-
nation may be socially desirable (and copyright law should promote it), from instances
where price discrimination is undesirable (and copyright law should discourage it).25
Sound policy must consider whether price discrimination can be controlled. Whenever
IP law enables arbitrage, there is a risk that doing so displaces benign price discrimina-
tion into other more pernicious forms. For instance, if a seller cannot rely on copyright
law to block arbitrage it might rely instead on DRM technologies, product design,
marketing methods, and vertical integration. Similarly, a seller may degrade product
quality to sort customers (Deneckere and McAfee, 1996), or a seller may over-rely on
leasing or conversion of durable goods into services to block arbitrage. Also, a firm
that sells intermediate goods might inefficiently integrate downstream displacing low
value consumers so that it can set a high price to high value consumers with no concern
about arbitrage.
Finally, price discrimination may engender additional social costs, such as loss of
privacy, personal autonomy, and foregone ethical or distributional commitments (Boyle,
2000, p. 2027) unless effective regulation of personal data retention and analysis is put in
place (Bar-Gill, 2018).
B. Patents
In the context of innovation and patents, discussion of price discrimination is also polar-
ized across optimists and pessimists.
Optimists point to increased profitability for the patentee (Sidak, 1981) and the
increased incentive to innovate (Bowman, 1973, pp. 56, 112; Klein and Wiley, 2003a), or to
24
On product diversity more generally, see Dixit and Stiglitz (1977).
25
Katz (2014) argues that post-sale restraints are beneficial when coproducing or collaborat-
ing firms are imperfectly vertically integrated.

increase product variety (Wright, 2005). Grennan (2013) documents increased competi-
tion resulting from price variations.
Hausman and MacKie-Mason (1998) emphasize the potential for price discriminators
to serve new markets and achieve scale and learning economies. They also make the
important point that, if the optimal policy solution is to apply the least costly manner of
rewarding innovators, if patentees are allowed to benefit from price discrimination, this
would enable other socially desirable adjustments, such as the reduction of the patent
term (Hausman and MacKie-Mason, 1998, pp. 263–4). Kaplow makes a similar argument
(1985, p. 524).
Pessimists highlight output restrictions (Sullivan, 1977) and other inefficiencies caused
by charging different prices (Baxter, 1966; Kaplow, 1984). Others point to how the rents
from price discrimination may induce socially wasteful rent-seeking (Posner, 1975) and
lead to excessive investments (Sykes, 2002) as well as socially wasteful patent races that
may include potentially duplicative R&D costs (Rai, 2001, p. 199; Sykes, 2002; Scotchmer,
2004, p. 98; Abramowicz, 2003, p. 129). Rent-seeking is aggravated by the enhanced risk
of frivolous or anti-competitive litigation created by patent and also copyright rights. The
difference between a simple contract claim, on the one hand, and a patent or copyright
infringement claim, on the other hand, is that the latter provides the IP owner significant
strategic advantages because of the threat of preliminary and permanent injunction,
fee-shifting, and treble damages for willful infringement. Furthermore, IP rights can be
asserted against innocent strangers (perhaps importers) who might be vulnerable to an
opportunistic IP suit. These rent-seeking costs need to be balanced against any incentive
benefit before IP rights are expanded to support price discrimination (Meurer, 2003b,
pp. 1881–2).
As reflected in our previous discussion of the social welfare effects of price discrimina-
tion in copyright dependent markets, normative conclusions in the patent context depend
on assumptions regarding the appropriate scope and social desirability of the patent
reward for optimal innovation. Since there is a variety of opinions on the dynamic welfare
effects of patent rewards, there are likewise a variety of opinions about the desirability of
price discrimination practices that increase the reward for patentees.
VI. THE MANY FORMS OF PRICE DISCRIMINATION BY WAY

OF IP RIGHTS
A. Geographic Price Discrimination
Sellers commonly establish exclusive territories in order to facilitate third degree geographic
price discrimination but also to encourage investment by distributors in local goodwill and
service. In response, gray markets emerge to arbitrage away price differentials that are
caused by exclusive national territories. Empirical evidence establishes that geographic
price discrimination is common, and probably the most important cause of gray market
transactions (Gallini and Hollis, 1999, p. 6; Malueg and Schwartz, 1994, pp. 173–4).
Because transportation costs are relatively low for most copyrighted works and many
patented inventions, sellers may need to rely on the force of law to prevent arbitrage.
Legislative bans on parallel importation as well as contractual restrictions are some

common instruments that reduce arbitrage and help implement price discrimination.
These trade restrictions run counter to some intellectual property doctrines, however.
In copyright law, the ‘first sale’ doctrine (17 U.S.C. § 109(a)) provides the owner of a
lawfully made copy of a work the right to sell it without the copyright holder’s authoriza-
tion. This doctrine seems at odds with section 602(a) which prohibits the unauthorized
importation of a copyrighted work.
In Quality King Distributors, Inc. v. L’anza Research Int’l, Inc., 523 U.S. 135, 118
S. Ct. 1125; 140 L. Ed. 2d 254 (1998) the United States adopted a partial exclusion of
international price discrimination, permitting exclusion of gray market goods that are
manufactured abroad. The application of the exhaustion principle only to goods pro-
duced domestically provided a perverse incentive to shift domestic production abroad
(Goldberg, 2012). In Kirtsaeng v. Wiley, 568 U.S. 519, 133 S. Ct. 1351 (2013) the US
Supreme Court did away with this distinction based on the location of manufacture
and held that the first sale doctrine applied to all legally produced goods, regardless
of origin of manufacture. This precedent effectively eliminated the opportunity for
sellers to use copyright law to block gray market goods in support of geographic price
discrimination.
Kirtsaeng probably had a particularly significant impact on the music and movie indus-
tries. It is hard to use product differentiation to support price discrimination for these
products. Additionally, customer service and warranties are not relevant to serve as a basis
of price discrimination. One possibility at the disposal of producers is to dub movies and
shows into foreign languages. Another possibility is to encode movies or music, so they
can only be played on devices manufactured for a particular country or region. The movie
industry has taken steps in that direction with country codes embedded in DVD movies.
Such measures are costly of course both with regard to the costs of implementation and
the loss of consumer satisfaction (Meurer, 2001).
Similarly, in the area of patents, sellers of patented products commonly charge
higher prices in the United States than abroad. To maintain these price differences,
such sellers prohibit the resale of their goods in the United States. The U.S. Patent Act
defines infringement in 35 U.S.C. § 271(a) by imposing liability on ‘whoever without
authority makes, uses, offers to sell, or sells any patented invention, within the United
States or imports into the United States any patented invention during the term of
the patent . . .’
The doctrine of patent exhaustion, however, offers a defense to importers, users, and
sellers of a patented invention if the patent holder has either authorized a first sale of a
patented item or licensed its use or sale. Accordingly, whenever a patentee sells an item, it
‘is no longer within the limits of the [patent] monopoly’ and instead becomes the ‘private,
individual property’ of the purchaser (Bloomer v. McQuewan, 55 U.S. 539 1852).
The hostility toward restraints on alienation is highlighted in Impression Products v.
Lexmark, 137 S. Ct. 1523 (2017). The Supreme Court held that a purchaser has the right
to use, sell, or import an item because those are the rights that come along with ownership,
regardless of a patentee’s desire to expressly limit the purchaser’s rights in this regard.
In the dispute, Impression Products imported patented printer cartridges that Lexmark
sold abroad. Impression Products refurbished the cartridges and undercut the price that
Lexmark was charging for the cartridges in the US. This arbitrage was permitted under
the exhaustion doctrine. Interestingly, Lexmark may have been using cartridge sales to

measure the frequency of use of Lexmark printers, in pursuit of a form of second degree
price discrimination that we will discuss more in Section VI.C.
For commentators who embrace a global social welfare perspective, geographic price
discrimination may have desirable distributional effects, for example, facilitating lower
pharmaceutical prices in poor countries. Though, sellers might lower prices of patented
goods in the United States and raise prices abroad (Hemel and Ouellette, 2016). The
benefit from low prices in poor countries might be eroded if sellers design lower quality
products to market in countries with weaker demand (Meurer, 2001, n. 397).
Finally, in the area of trademarks, goods bearing identical trademarks are regularly
sold at different prices in different geographical regions. Trademark owners seek to
protect against arbitrage by arguing that use of the trademark on the imported (gray
market) goods is unauthorized and infringes the exclusive right to use of a trademark on
a particular product in a specific geographic area (Lansing and Gabriella, 1993).26
Supporters of geographic price discrimination by trademark holders argue that gray
marketeers unfairly free-ride on the advertising and goodwill developed by trademark
owners and authorized distributors (Liebeler, 1987, pp. 756–7). Gray markets may also be
harmful to consumers because goods are not sold with the same warranties and quality
assurances as products sold through authorized channels (Higgins and Rubin, 1986,
pp. 228–9; Landes and Posner, 1987, pp. 308–9; Liebeler, 1987, p. 755; Lipner, 1990). They
claim that territorial divisions enable producers to provide consumers with improved
services and assurances of quality. Excess profits resulting from territorial division
and restricted competition can be used to improve the quality of goods purchased by
consumers (Philips, 1981; Tirole, 1988). Moreover, price differentials might be driven by
differences in tastes, technologies, and government regulations across regions (Peterman,
1993; Lansing and Gabriella, 1993).
Critics of geographic price discrimination point to consumer benefits resulting from
lower prices (Rubin, 1992, pp. 618–22), free trade (Lipner, 1990), and intrabrand competi-
tion (Gallini and Hollis, 1999) enabled by gray markets. To the extent that geographic
price differentials reflect monopoly power, gray markets help erode barriers to trade
(Dam, 1964, pp. 53–60). Ghosh (1994) provides a formal model to suggest that the most
efficient result is to permit gray market goods that have alternative labels. Additionally,
prohibiting gray markets imposes considerable public enforcement costs (Ghosh, 1994,
pp. 378–9).
B. Restrictions on Type of Use
Software sellers often charge different prices for commercial versus personal or academic
use. Contract law can be used to enforce the restriction limiting low-priced software pur-
chasers to non-commercial use (ProCD, Inc. v. Zeidenberg, 86 F.3d 1447, 1449 (7th Cir.
1996)). But these sorts of restrictions are more robust when intellectual property law can
26
A US trademark holder may bar the importation of goods bearing the same trademark
when manufactured by a foreign manufacturer but cannot stop importation of goods made under
the control of the domestic trademark holder. See K-Mart Corp. v. Cartier, Inc., 486 U.S. 281, 292
(1987).

be used to enforce the restrictions. Intellectual property remedies are often stronger than
contract remedies, and intellectual property claims are available against arbitrageurs who
are strangers to the seller, thereby overcoming the privity limitation on contract claims.
Patent law broadly facilitates restrictions on type of use. A patent owner has the right
to exclude others from use of a patented invention. The predominant view in patent law
states that because the patent owner can exclude all use, the statute gives an implied right
to grant permission for some uses and still sue the licensee for infringement if she engages
in an unauthorized use (Bowman, 1973, pp. 140–42).27 For example, DuPont imposed a
field of use restriction and charged different prices for a patented synthetic fiber depend-
ing on the end use intended by the customers (Akzo v. Int’l Trade Comm., 808 F.2d 1471
(Fed. Cir. 1986)).
Copyright gives somewhat limited support for restrictions on the type of use.28 Much
of the value from movie and music copyrights comes from the public performance right
which allows copyright owners to control public performances of their works (17 U.S.C.
§ 106(4)). This right facilitates price discrimination in the movie and music markets
between home users and buyers who want to engage in a public performance, for example,
exhibiting a movie in a theater or broadcasting music on the radio (Besen and Kirby, 1989;
1992). It also facilitates the fine-grained price discrimination practiced by ASCAP, BMI,
and SESAC (collective rights organizations that administer blanket public performance
licenses on behalf of music composition copyright owners). The royalties associated with
the blanket licenses vary according to the size and revenue of the establishment using the
license.
The potent threat created by a possible copyright suit helps assure compliance with the
price discrimination schemes used for the public performance of music and movies. In
this manner, copyright law channels sellers into choosing a relatively socially beneficial
form of price discrimination rather than a more socially harmful form. If the public
performance right were deleted from the statute, music and movie producers would need
to find another, likely costlier, way to discriminate between buyers intending to publicly
perform the work, and buyers intending only private use. One possibility would be a very
high initial sales price followed, after a significant delay, with a lower sales price targeted
at home users. Another possibility would be vertical integration into movie exhibition or
radio broadcast. The public performance right allows discrimination and avoids the high
implementation costs associated with the other strategies (Meurer, 2003b, pp. 1883–4).
Finally, we note that the performance right is limited by certain exemptions that
have the effect of limiting the scope of price discrimination. For instance, 17 U.S.C.
§ 110 exempts educational and nonprofit performances from the reach of the public
performance right. These exemptions might be justified by relatively high transaction
costs compared to the private value of these sorts of public performances to purchasers.
In effect, through these exemptions copyright brings about price discrimination for users
intending a public performance, while sheltering certain users who might generate positive
externalities (Meurer, 2003b, p. 1883).
27
But see Kaplow (1984, p. 1846). The implied right is made explicit regarding restrictions on
location of use. See 35 U.S.C. § 261.
28
Lunney (1996) analyses how the right to create derivative works enables price discrimination.

C. Intensity of Use: Tying, Merchandising, and Bundling
Price discriminating sellers often link prices to a direct or indirect measure of the intensity
of use of a durable product. Ideally for the seller, buyers would truthfully respond to
queries about how frequently they will use a product and how many people will benefit
from the use. Most often the high intensity users would be willing to pay more for the
product, and prices would be set accordingly.
Typically, price discrimination relies on indirect measures of intensity of use because
direct measures are hard to implement. Usually it is costly to monitor use, easy to evade
restrictions, and usage-based measures and monitoring sometimes annoy consumers.
Nevertheless, sellers of software often set prices for enterprise or site licenses based on the
number of users or machines that are licensed. Arbitrage may be discouraged by contract
terms that allow the software vendors to audit customers’ facilities.
In some cases, technology intrinsic to the product or added to the product can be used to
facilitate monitoring. The odometer on a rented car or a usage meter on office equipment
can be used to set rental prices based on intensity of usage. Today, sellers can monitor
usage of any equipment connected to the internet, for example, mobile phone data plans.
Additionally, patent and copyright law can be used to bolster contract-based approaches
to usage-based pricing. As we discussed in the previous section, IP infringement claims
against strangers and strong IP remedies strengthen the enforcement power of sellers
against arbitrage. Infringement claims are well grounded in patent law because the patent
owner has broad control over use. To illustrate, in Brulotte v. Thys Co., 379 U.S. 29 (1964),
the patent owner sold patented farm equipment and included a license term that required
royalty payments based on the bushels of harvested crop.
Copyright law does not offer a comparably broad use right but in some important set-
tings unauthorized use is infringing. Computers (and other consumer electronic devices)
usually make a temporary copy of digital content or software during use. Even though
temporary, such a copy may be infringing. Thus, a digital copyright owner can sue a buyer
who violates a frequency of use restriction for breach of contract, and also for copyright
infringement because of the unauthorized temporary reproductions. Copyright law
imposes two important limits on these infringement claims. Section 117 gives software
owners the right to make copies as an essential step in using a program, and the copyright
misuse doctrine may restrict sellers from using this strategy in certain settings.
Sellers use a variety of tactics to indirectly measure intensity of use by measuring
consumption of some complementary product that is used with the durable product.
For example, a patent owner leased a patented canning machine and required lessors to
purchase the salt that they needed for canning from the patent owner (International Salt
Co. v. United States, 332 U.S. 392 (1947)). Tying rental of the machine to sales of salt
offered the patent owner a method to measure intensity of use—assuming those who
used the machine more needed to use more salt. Besides tying agreements, a seller might
design its durable products in such a way that it has an advantage selling a complementary
product because of a proprietary interface between the durable product and the comple-
mentary product. Additionally, a seller might use the threat of an IP lawsuit against the
maker or seller of the complementary product to achieve exclusivity in the market for
the complementary sales. Some sellers use all three of these tactics to control sales of the
complementary product.

In a tying arrangement, the seller charges a supra-competitive price for the tied product
and then measures the intensity of use of the tying product on the basis of the sales of
the tied product. Because the price for the tied product is usually marked up beyond
the competitive price, high-volume users are effectively charged a higher price for the
tying product (Klein and Wiley, 2003a, pp. 604–5; Butler et al., 1984, p. 190; Hansen and
Roberts, 1980). This creates a social loss as all buyers purchase too little of the tied good.
This loss might be offset by a reduction in the price of the tying product that induces new
buyers to enter the market (Lichtman, 2000).
Classic cases include IBM Corp. v. United States, 298 U.S. 131 (1936) where IBM
leased patented tabulator machines on the condition that the lessee purchase all of the
punch cards needed for use in the machines from IBM. Punch card purchases measured
frequency of use. Rather than charging a rental rate that varied directly with frequency
of use, IBM charged a premium over the competitive price for punch cards, and thereby
indirectly collected a rental rate that increased with the frequency of use. In Motion
Pictures Patents Co. v. Universal Film Mfg., 243 U.S. 502 (1917) movie projectors were
tied to film, and the patent owner derived most of its profit from the sale of film. In
Morton Salt Co. v. G.S. Suppiger Co., 314 U.S. 488, 489–90 (1942), the lease of a patented
canning machine was tied to the sale of salt tablets and salt sales metered the intensity of
use of the canning machine.
Rather than a tying agreement, a seller may use the design of a product interface to
gain control over the sale of a complementary product. In Sega v. Accolade, 977 F.2d
1510 (9th Cir. 1992), Sega sold video game consoles that were used in conjunction with
game cartridges. Sega controlled the cartridge market via software required to make the
cartridges function in the console. By earning a licensing fee on authorized cartridges,
Sega could indirectly distinguish between purchasers of the console who bought only a
few cartridges and those who bought many, and it could use this system to implement
usage-based price discrimination. Ultimately this strategy failed because Accolade was
able to reverse engineer the interface and make and sell compatible cartridges that were
not authorized by Sega. Sega sued Accolade for copyright and trademark infringement
and lost on both grounds. In other contexts, a seller may have better luck if the interface
is difficult and costly to reverse engineer, or if the interface is protected by a patent. Sellers
can block all use of a patented interface, they have had mixed success enforcing contracts
that preclude reverse engineering, but reverse engineering is typically permitted under
trade secret and copyright law.
A third approach to controlling frequency of use relies on the threat of IP suits against
competing suppliers of the complementary product. For example, in MAI Systems v. Peak
Computer, 991 F.2d 511 (9th Cir. 1993), MAI threatened copyright litigation to discourage
third parties like Peak from providing maintenance services to buyers of MAI computers.
MAI computers relied on an operating system copyrighted by MAI and they contended that
when Peak turned on the computer to perform a maintenance service, a copy of at least a
part of the operating system software was necessarily made in the computer’s random access
memory (RAM). The court found that copying infringed on the reproduction right.29 Until
29
Specifically, the court held that a computer’s RAM satisfies the fixation requirement and a
temporary RAM copy is a reproduction within the meaning of section 106(1).

Congress adopted legislation to reverse this outcome,30 the plaintiff was able to use the ruling
to exclude the defendant from the maintenance market. Similarly, in certain cases, patent law
may facilitate control over a complementary product by allowing contributory infringement
suits against competing suppliers of the complementary product (Meurer, 2003b).
In other contexts, the threat of copyright and trademark lawsuits can be used against
makers and sellers of merchandise that is consumed as a complement to a copyrighted
work. Movies and television series producers have succeeded in ‘merchandising’ products
derived from their audiovisual works. Merchandising implements usage-based price dis-
crimination if the highest valuation consumers are the ones who buy the most merchan-
dise. Trademark law may protect the title and characters from movie or television series,
and copyright law protects the use of images from these audiovisual works. The threat
of IP suits gives the movie and television producers the chance to control the markets for
posters, clothing, toys, games, and other merchandise based on the copyrighted work.
In general, the social welfare effects of usage-based pricing can be positive or negative.
Output-based pricing tends to draw new customers into a market, specifically, customers
who are infrequent users who are attracted by the relatively low price charged for
infrequent use. At the same time, consumption tends to fall among current customers
who formerly consumed as much as they wanted, and now face a positive price for each
additional use.
The social welfare effects of tying arrangements, in particular, have been widely
debated. Pessimists fear that tie-ins restrict competition in the tied product market. Tying
can be used to leverage monopoly power from the primary market (tying product) to
a secondary market (tied product) (Carlton and Waldman, 2002, p. 194; Leslie, 2004).
Tying arrangements may distort competition by deterring entry into the market for the
tied product. In markets with incumbents that sell tied products, competitors may face
an uphill battle in order to achieve sufficient scale to compete or even cover fixed costs
(Nalebuff, 2004; Whinston, 1990; Elhauge, 2009, pp. 413–14). Optimists are skeptical that
monopolists can increase their total market power by way of tying, since price increases in
one market are likely to lead to lost profits in the other market (Bork, 1978, pp. 366–7, 372;
Posner, 2001, p. 201). Others maintain that tie-in metering protects goodwill, promotes
quality control (Meese, 1997), can increase product quality (Dana and Spier, 2015), and
may bolster research and development by enabling patent holders to recover more of the
social value of their inventions (Wright, 2005; Grill, 2006), and that improved information
may increase producer and consumer surplus (Hylton and Salinger, 2001). Metered tying
has also been associated with social welfare benefits if it leads to improved sales of durable
goods (Elhauge and Nalebuff, 2017).
Similar to the issue of tying, any normative assessment of merchandising and the
appropriate scope of copyright and trademark support of merchandising will depend
on one’s view on the optimal reward for the copyright owner. Skeptics argue, however,
that broad exclusive rights on merchandising distort creative activities, causing excessive
investments in story lines and productions with an eye on toys and other merchandise
30
Congress reversed this result in the Digital Millennium Copyright Act. See 17 U.S.C.A. §§
117(a), (c)–(d) (West 1999) (these sections reflect the amendments to 17 U.S.C.A. by § 301 of the
Digital Millennium Copyright Act).

that can be derived from the movie (Lunney, 1996, pp. 640–41; Sterk, 1996, p. 121; Meurer,
2001, pp. 128–9; Wyatt, 1994, p. 149; Litwak, 1997, p. xii).
An interesting question exists whether a patentee of a tying product has an incentive
to stifle or bolster innovation in the tied product market. On the one hand, a tying
seller would benefit from innovation in the tied product market since improvements
of a complementary product likely increase demand for the patentee’s product. On the
other hand, innovation in the tied product market could potentially undermine the pat-
entee’s dominant position and profitability in the tying product market (Feldman, 1999,
pp. 2091–3; Leslie, 2011, pp. 834–5; Choi, 2004). For that reason, patent tying can reduce
the incentives for innovation in the tied product market (Hovenkamp, 2007; Choi and
Stefanadis, 2001, 2006). Additionally, by forcing rivals to enter two markets concurrently,
tying can stifle competition and innovation in the tied product market (Eastman Kodak
Co. v. Image Tech. Servs., Inc., 504 U.S. 451, 485 (1992) (citing Jefferson Parish Hosp.
Dist. No. 2 v. Hyde, 466 U.S. 2, 14 (1984)).
We close this section by discussing a related method of price discrimination known as
bundling. We distinguish tying agreements that bundle a complementary product that
helps a seller measure the intensity of demand for some durable product, with other bun-
dles that serve to smooth demand by heterogeneous consumers over multiple products.
Bundling naturally aligns with the goal of copyright law to promote broad diffusion of
creative works.
Bundling promotes diffusion because bundles are easier to price. Averaging consumer
demand over multiple products reduces the variance in demand. This means there are
fewer buyers in the ‘tail’ of the demand curve who get excluded.
Bundling creates several other possible benefits. The clearest benefit is a reduction in
transaction and enforcement costs. The blanket licensing practice of the copyright col-
lectives provides the best illustration. The US Supreme Court suspended the per se rule
against price fixing in an antitrust case against BMI due to the difficulty of enforcing the
public performance right (Broadcast Music, Inc. v. Columbia Broadcast. Sys. Inc., 441
U.S. 1 (1979)). Bundling may also avoid wasteful investment in measuring the value of the
components of a bundle (Kenney and Klein, 1983), to lower distribution costs (Hanssen,
2000) and problems associated with anticommons fragmentation (Depoorter and Parisi,
2003).
A bleaker perspective is that bundling may reduce consumer surplus and reduce entry
into digital information goods markets by competitors (Bakos and Brynjolfsson, 2000a).
Nalebuff (2004) claims that bundling enables a company with market power in two goods
to make it harder for a rival with only one of these goods to enter the market.31
Bakos and Brynjolfsson (2000b) predicted correctly that digital technologies and dis-
tribution would amplify both disaggregation-based pricing strategies, such as Itunes, and
aggregation strategies, whereby information goods will be offered in bundles, site licenses,
and subscriptions, such as Spotify, Hulu, Netflix, and so on (Roin, 2014; Liebowitz and
Margolis, 2009).
31
For a critique of theoretical ligature on bundling and, specifically, the lack of empirical
grounding, see Kobayashi (2005).

D. Sharing
A highly contentious issue in copyright is the type and extent of sharing by users without
permission from copyright owners. Especially in the digital era, private copying and
sharing of movies, music, and software has become commonplace. The music industry in
particular aggressively pursued the sharing of copyrighted works on file-sharing networks
(Depoorter et al., 2011).
Sharing can be defined as any activity such that (1) a single copy of a work provides
utility to more than one end-user, and (2) the number of sharing users is relatively small.
Copyright markets feature three common types of sharing: joint use through perfor-
mance; reproduction leading to simultaneous use; and consecutive use through lending or
resale. Consumers often buy, sell, and lend used books, movies, and music. Public libraries
lend books and other copyrighted works to the public at no charge.
Sharing impacts the feasibility and profitability of price discrimination. It can be used
to circumvent price discrimination. Software sharing, for instance, is a common route for
arbitrage. It may bring together two different classes of buyers that the seller would like to
keep separate for the purpose of price discrimination. Software sellers often discriminate
between the academic and business markets, or between the home and business markets.
This sort of discrimination is less effective if business users routinely share with academic
or home users (Meurer, 2003b).
Though sharing may cause losses to sellers, a number of factors help mitigate these
losses. With consumers doing some of the work, sharing copyrighted works may provide
a cost-effective method of production, distribution (Besen and Kirby, 1989, p. 255;
Novos and Waldman, 1984), and marketing. Additionally, sharing may induce network
effects that raise the value to users (Conner and Rumelt, 1991; Takeyama, 1994; Shy and
Thisse, 1999).32 These benefits from sharing can be deliberately attained if the copyright
owner authorizes sharing and sets prices to capture some of the benefits. Sharing poses
a particular threat to profit when sharing decreases distribution costs (Bakos et al., 1999)
or when authorized purchasers do not appropriate (or otherwise account for) much of
the value derived by other users (Meurer, 2001, pp. 133–40).
Formally, copyright law allows users to engage in certain sharing activities without
permission from copyright owners. Consider, for example, the type of sharing enabled by
the VCR. The Supreme Court permitted private videotaping of televised movies under the
fair use doctrine in Sony Corp. of America v. Universal City Studios, Inc. (464 U.S. 417
(1984)). Additionally, Congress refused to prohibit unauthorized commercial rental of
videotapes. Informally, copyright owners tolerate many private, non-commercial sharing
activities (Tehranian, 2007; Wu, 2008).
Restrictions on sharing may be socially harmful if (1) the owner blocks socially valuable
sharing because it is unprofitable, or (2) the owner inefficiently distorts the nature of shar-
ing to gain more profit, especially if profits exceed levels necessary to stimulate the c reation
32
Varian (2000) identifies three situations where sharing will prompt a content producer to
sell a smaller amount at higher prices and see an increase in profits: (1) when the transactions cost
of sharing is less than the marginal cost of production; (2) when content is viewed only a few times
and transactions costs of sharing are low; and (3) when a sharing market provides a way to segment
high value and low value users.

and distribution of works (Meurer, 1997, pp. 1183–9; 2001, pp. 132–40). Optimally,

copyright law should permit sharing when the profit-based incentives of copyright owners
are misaligned with the social incentive in maximizing ex post total surplus, provided the
social cost in terms of lost productive incentive is not too great (Meurer, 2004, pp. 910–12).
VII. CONCLUSION
The literature we have reviewed has allowed us to compile an intriguing compendium of

connections between intellectual property law and price discrimination. We think that
new technologies are expanding the range of such connections and present new topics that
should be addressed in more depth by intellectual property law scholars. In particular, we
expect to see growing commentary on copyright and trademark liability of e-commerce
platforms and how that connects to arbitrage and price discrimination. Further, we expect
to see growing commentary on the connection between intellectual property, privacy, and
antitrust laws and incentives to build and use databases and algorithms in support of
To conclude we note a severe imbalance between empirical and theoretical work on
intellectual property and price discrimination. The extensive normative analysis we have
reviewed is usually inconclusive because commentators lack the evidence to balance the
social costs and benefits that theory tells us are created by price discrimination. More
troubling, there is relatively little empirical evidence demonstrating that changes in
intellectual property cause significant changes in the practice or profitability of price
discrimination. Intuition and anecdote point to strong connections but rigorous empirical
work needs to be done to confirm our intuitions.
BIBLIOGRAPHY
Abramowicz, Michael. 2003. ‘Perfecting Patent Prizes,’ 56 Vanderbilt Law Review 115–236.
Aguirre, Inaki, Simon Cowan, and John Vickers. 2010. ‘Monopoly Price Discrimination and Demand
Curvature,’ 100 American Economic Review 1601–15.
Alger, Ingela. 1999. ‘Consumer Strategies Limiting the Monopolist’s Power: Multiple and Joint Purchases,’ 30
The RAND Journal of Economics 736–57.
Anderson, Eric T., and James D. Dana, Jr. 2009. ‘When Is Price Discrimination Profitable?,’ 55 Management
Science 980–89.
Anderson, Simon P., and Andre De Palma. 1988. ‘Spatial Price Discrimination with Heterogeneous Products,’
55 The Review of Economic Studies 573–92.
Anderson, Simon P., and Victor A. Ginsburgh. 1994. ‘Price Discrimination via Second-Hand Markets,’ 38
European Economic Review 23–44.
Armstrong, Mark. 1999. ‘Price Discrimination by a Many-Product Firm,’ 66 The Review of Economic Studies
151–68.
Armstrong, Mark, and John Vickers. 2001. ‘Competitive Price Discrimination,’ 32 The RAND Journal of
Baker, Edwin C. 1997. ‘Giving the Audience What It Wants,’ 58 Ohio State Law Journal 311–417.
Baker, Jonathan B. 1997. ‘Product Differentiation through Space and Time: Some Antitrust Policy Issues,’ 42
Antitrust Bulletin 177–96.
Bakos, Yannis, and Eric Brynjolfsson. 1999. ‘Bundling Information Goods: Pricing, Profits, and Efficiency,’ 45
Bakos, Yannis, and Erik Brynjolfsson. 2000a. ‘Aggregation and Disaggregation of Information Goods:
Implications for Bundling, Site Licensing and Micropayment Systems,’ in Hal Varian and Brian Kahin, eds.,

Internet Publishing and Beyond: The Economics of Digital Information and Intellectual Property, Cambridge,
MA: MIT Press.
Bakos, Yannis, and Erik Brynjolfsson. 2000b. ‘Bundling and Competition on the Internet,’ 19 Marketing Science
Informs 63–82.
Bakos, Yannis, Eric Brynjolfsson, and Douglas Lichtman. 1999. ‘Shared Information Goods,’ 42 The Journal
of Law and Economics 117–56.
Balganesh, Shyamkrishna. 2013. ‘The Uneasy Case against Copyright Trolls,’ 86 Southern California Law
Review 723–82.
Baxter, William F. 1966. ‘Legal Restrictions and Exploitations of Patent Monopoly: An Economic Analysis,’
76 Yale Law Journal 267–310.
Bar-Gill, Oren. 2018. ‘Algorithmic Price Discrimination: When Demand Is a Function of Both Preferences and
(Mis)Perceptions,’ The Harvard John M. Olin Discussion Paper Series, No. 05/2018.
Baumol, William J., and Daniel G. Swanson. 2003. ‘The New Economy and Ubiquitous Competitive Price
Discrimination: Identifying Defensible Criteria of Market Power,’ 70 Antitrust Law Journal 661–85.
Bell, Tom W. 1998. ‘Fair Use vs. Fared Use: The Impact of Automated Rights Management on Copyright’s Fair
Use Doctrine,’ 76 North Carolina Law Review 557–619.
Benkler, Yochai. 1999. ‘Free as the Air to Common Use: First Amendment Constraints on Enclosure of the
Public Domain,’ 74 New York University Law Review 354–446.
Benkler, Yochai. 2000. ‘An Unhurried View of Private Ordering in Information Transactions,’ 53 Vanderbilt
Bergemann, Dirk, Benjamin Brooks, and Stephen Morris. 2015. ‘The Limits of Price Discrimination,’ 105
Besen, Stanley M., and Sheila N. Kirby. 1989. ‘Private Copying, Appropriability, and Optimal Copyright
Royalties,’ 32 The Journal of Law and Economics 255–80.
Besen, Stanley M., and Leo J. Raskind. 1991. ‘An Introduction to the Law and Economics of Intellectual
Property,’ 5 Journal of Economic Perspectives 3–27.
Besen, Stanley M., Sheila N. Kirby, and Steven C. Salop. 1992. ‘An Economic Analysis of Copyright Collectives,’
78 Virginia Law Review 383–411.
Borenstein, Severin. 1985. ‘Price Discrimination in Free-Entry Markets,’ 16 The RAND Journal of Economics
380–97.
Bork, Robert. 1978. The Antitrust Paradox: A Policy at War with Itself, New York, NY: Free Press.
Bowman, Ward S. Jr. 1957. ‘Tying Arrangements and the Leverage Problem,’ 67 Yale Law Journal 19–36.
Bowman, Ward S. Jr. 1973. Patent and Antitrust Law, Chicago, IL: University of Chicago Press.
Boyle, James. 2000. ‘Cruel, Mean, or Lavish? Economic Analysis, Price Discrimination and Digital Intellectual
Property,’ 53 Vanderbilt Law Review 2007–39.
Butler, Henry N., W.J. Lane, and Owen R. Phillips. 1984. ‘The Futility of Antitrust Attacks on Tie-In Sales: An
Economic and Legal Analysis,’ 36 Hastings Law Journal 173–213.
Carlton, Dennis W., and Michael Waldman. 2002. ‘The Strategic Use of Tying to Preserve and Create Market
Power in Evolving Industries,’ 33 RAND Journal of Economics 194–220.
Choi, Jay Pil. 2004. ‘Tying and Innovation: A Dynamic Analysis of Tying Arrangements,’ 114 The Economic
Journal 83–101.
Choi, Jay Pil, and Christodoulos Stefanadis. 2001. ‘Tying, Investment, and the Dynamic Leverage Theory,’ 32
Choi, Jay Pil, and Christodoulos Stefanadis. 2006. ‘Bundling, Entry Deterrence, and Specialist Innovators,’ 79
Journal of Business 2575–94.
Cohen, Julie E. 1998. ‘Copyright and the Jurisprudence of Self-Help,’ 13 Berkeley Technology Law Journal
1089–143.
Cohen, Julie E. 2000. ‘Copyright and the Perfect Curve,’ 53 Vanderbilt Law Review 1799–819.
Conner, Kathleen R., and Richard P. Rumelt. 1991. ‘Software Piracy: An Analysis of Protection Strategies,’ 37
Cornish, William R. 1998. ‘Trade Marks: Portcullis for the EEA,’ 20 European Intellectual Property Review 172–7.
Corts, Kenneth S. 1998. ‘Third-Degree Price Discrimination in Oligopoly: All-Out Competition and Strategic
Commitment,’ 29 The RAND Journal of Economics 306–23.
Cowan, Simon. 2012. ‘Third-Degree Price Discrimination and Consumer Surplus,’ 60 The Journal of Industrial
Economics 333–45.
Crane, Daniel A. 2008. ‘Patent Pools, RAND Commitments, and the Problematics of Price Discrimination,’
Cardozo Legal Studies Research Paper, No. 232.
Dam, Kenneth W. 1964. ‘Trademarks, Price Discrimination and the Bureau of Customs,’ 7 Journal of Law and
Economics 45–60.
Dana, James D., and Kathryn E. Spier. 2015. ‘Bundling and Quality Assurance,’ 49 RAND Journal of Economics
128–54.

Danzon, Patricia M., and Adrian Towse. 2003. ‘Differential Pricing for Pharmaceuticals: Reconciling Access,
R&D and Patents,’ 3 International Journal of Health Care Finance and Economics 183–205.
Demsetz, Harold. 1970. ‘The Private Production of Public Goods,’ 13 The Journal of Law and Economics
293–306.
Deneckere, Raymond J., and R. Preston McAfee. 1996. ‘Damaged Goods,’ 5 Journal of Economics and
Management Strategy 149–74.
Depoorter, Ben, and Francesco Parisi. 2002. ‘Fair Use and Copyright Protection: A Price Theory Explanation,’
21 International Review of Law & Economics 453–73.
Depoorter, Ben, and Francesco Parisi. 2003. ‘The Market for Intellectual Property: The Case of Complementary
Oligopoly,’ in Wendy Gordon and Richard Watts, eds., The Economics of Copyright: Developments in Research
and Analysis, Cheltenham, UK and Northampton, MA, USA: Edward Elgar Publishing.
Depoorter, Ben, Alain Van Hiel, and Sven Vanneste. 2011. ‘Copyright Backlash’ 84 Southern California Law
Review 1251–92.
Dixit, Avinash, and Joseph Stiglitz. 1977. ‘Monopolistic Competition and Optimum Product Diversity,’ 67
Elhauge, Einer. 2003. ‘Defining Better Monopolization Standards,’ 56 Stanford Law Review 253–344.
Elhauge, Einer. 2009. ‘Tying, Bundled Discounts, and the Death of the Single Monopoly Profit Theory,’ 123
Elhauge, Einer, and Barry Nalebuff. 2017. ‘The Welfare Effects of Metering Ties,’ 33 Journal of Law, Economics,
and Organization 68–104.
Englmaier, Florian, Linda Gratz, and Markus Reisinger. 2012. ‘Price Discrimination and Fairness Concerns’
(February 5, 2012), accessed April 3, 2019 at http://ssrn.com/abstract=1999757.
Feldman, Robin Cooper. 1999. ‘Defensive Leveraging in Antitrust,’ 87 Georgetown Law Journal 2079–115.
Fisher III, William W. 1988. ‘Reconstructing the Fair Use Doctrine,’ 101 Harvard Law Review 1659–795.
Fisher III, William W. 1997. ‘Property and Contract on the Internet,’ 73 Chicago-Kent Law Review 1203–56.
Fisher III, William W. 2007. ‘When Should We Permit Differential Pricing of Information,’ 55 UCLA Law
Review 1–38.
Friedman, David. 1998. ‘In Defense of Private Orderings: Comments on Julie Cohen’s “Copyright and the
Jurisprudence of Self-Help,”’ 13 Berkeley Technology Law Journal 1151–72.
Gallini, Nancy T., and Aidan Hollis. 1999. ‘A Contractual Approach to the Gray Market,’ 19 International
Review of Law and Economics 1–21.
Ghosh, Shubha. 1994. ‘An Economic Analysis of the Common Control Exception to Gray Market Exclusion,’
15 University of Pennsylvania Journal of International Business Law 373–439.
Goldberg, Melissa. 2012. ‘A Textbook Dilemma: Should the First Sale Doctrine Provide a Valid Defense for
Foreign-Made Goods?,’ 80 Fordham Law Review 3057–92.
Goldstein, Paul. 1994. Copyright’s Highway: From Gutenberg to the Celestial Jukebox, NYC, NY: Farrar, Straus
& Giroux.
Gordon, Wendy J. 1992. ‘On Owning Information: Intellectual Property and the Restitutionary Impulse,’ 78
Gordon, Wendy J. 1998. ‘Intellectual Property as Price Discrimination: Implications for Contract,’ 73
Chicago‑Kent Law Review 1367–90.
Grennan, Matthew. 2013. ‘Price Discrimination and Bargaining: Empirical Evidence from Medical Devices,’
Grill, Brian T. 2006. ‘The Treatment of Metering in Antitrust Law: The Supreme Court’s Apparent Abolition
of the Per Se Rule against Metering in Illinois Tool Works, Inc. v. Independent Ink, Inc.,’ 2006 Wisconsin
Hansen, Robert S., and R. Blaine Roberts. 1980. ‘Metered Tying Arrangements, Allocative Efficiency and Price
Discrimination,’ 47 Southern Economic Journal 73–83.
Hanssen, F. Andrew. 2000. ‘The Block Booking of Films Re-Examined,’ 43 Journal of Law and Economics 395–426.
Hardy, Trotter. 1996. ‘Property (and Copyright) in Cyberspace,’ 1996 University of Chicago Legal Forum 217–60.
Hausman, Jerry A., and Jeffrey K. MacKie-Mason. 1998, ‘Price Discrimination and Patent Policy,’ 19 The
Hemel, Daniel J., and Lisa L. Ouellette. 2016. ‘Trade and Tradeoffs: The Case of International Patent
Exhaustion,’ 116 Columbia Law Review Sidebar 17–31
Hendel, Igal, and Aviv Nevo. 2013. ‘Intertemporal Price Discrimination in Storable Goods Markets,’ 103
Higgins, Richard S., and Paul H. Rubin. 1986. ‘Counterfeit Goods,’ 29 Journal of Law and Economics 211–30.
Holmes, Thomas J. 1989. ‘The Effects of Third-Degree Price Discrimination in Oligopoly,’ 79 American
Hovenkamp, Herbert. 1999. Federal Antitrust Policy: The Law of Competition and Its Practice, New York, NY:
West Publishing Co.

Hovenkamp, Herbert. 2007. ‘Restraints on Innovation,’ 29 Cardozo Law Review 247–60.

Hovenkamp, Herbert, Mark D. Janis, Christopher R. Leslie, and Mark A. Lemley. 2009. IP and Antitrust: An
Analysis of Antitrust Principles Applied to Intellectual Property Law, vol. II, New York, NY: Wolters Kluwer
Law and Business.
Hughes, Justin. 2016. ‘Motion Pictures, Markets, and Copylocks,’ 23 George Mason Law Review 941–66.
Hylton, Keith N., and Michael Salinger. 2001. ‘Tying Law and Policy: A Decision-Theoretic Approach,’ 69
Antitrust Law Journal 469–526.
Jasanoff, S. 2013. ‘Epistemic Subsidiarity: Coexistence, Cosmopolitanism, Constitutionalism,’ 4 European
Journal of Risk Regulation 133–41. 
Kaplow, Louis. 1984. ‘The Patent Antitrust Intersection: A Reappraisal,’ 97 Harvard Law Review 1813–92.
Kaplow, Louis. 1985. ‘Extension of Monopoly Power through Leverage,’ 85 Columbia Law Review 515–56.
Katz, Ariel. 2014. ‘The First Sale Doctrine and the Economics of Post-Sale Restraints,’ 2014 Brigham Young
Katz, Michael L. 1984. ‘Price Discrimination and Monopolistic Competition,’ 52 Econometrica: Journal of the
Econometric Society 1453–71.
Katz, Michael L. 1987. ‘The Welfare Effects of Third-Degree Price Discrimination in Intermediate Good
Markets,’ 77 American Economic Review 154–67.
Kenney, Roy W., and Benjamin Klein. 1983. ‘The Economics of Block Booking,’ 26 Journal of Law and
Kitch, Edmund. 1999. ‘Can the Internet Shrink Fair Use?,’ 78 Nebraska Law Review 880–90.
Klein, Benjamin, and Lester F. Saft. 1985. ‘The Law and Economics of Franchise Tying Contracts,’ 28 The
Journal of Law and Economics 345–61.
Klein, Benjamin, and John Shepard Wiley Jr. 2003a. ‘Competitive Price Discrimination as an Antitrust
Justification for Intellectual Property Refusals to Deal,’ 70 Antitrust Law Journal 599–642.
Klein, Benjamin, and John Shepard Wiley Jr. 2003b. ‘Market Power in Economics and in Antitrust: Reply to
Baker,’ 70 Antitrust Law Journal 655–9.
Kobayashi, Bruce. H. 2005. ‘Does Economics Provide a Reliable Guide to Regulating Commodity Bundling by
Firms? A Survey of the Economic Literature,’ 1 Journal of Competition Law & Economics 707–46.
Landes, William M., and Richard A. Posner. 1987. ‘Trademark Law: An Economic Perspective,’ 30 Journal of
Landes, William M., and Richard A. Posner. 1989. ‘An Economic Analysis of Copyright Law,’ 18 Journal of
Lansing, Paul, and Joseph Gabriella. 1993. ‘Clarifying Gray Market Gray Areas,’ 31 American Business Law
Journal 313–37.
Leeson, Peter T., and Russell S. Sobel. 2008. ‘Costly Price Discrimination,’ 99 Economics Letters 206–8.
Lemley, Mark A. 1997. ‘The Economics of Improvement in Intellectual Property Law,’ 75 Texas Law Review
989–1084.
Leslie, Christopher R. 2011. ‘Patent Tying, Price Discrimination, and Innovation,’ 77 Antitrust Law Journal
811–54.
Leslie, Philip. 2004. ‘Price Discrimination in Broadway Theater,’ 35 The RAND Journal of Economics 520–41.
Levine, Michael E. 2000. ‘Price Discrimination Without Market Power,’ Harvard John M. Olin Discussion
Paper Series, No. 276.
Lewis, Tracy R., and David E.M. Sappington. 1994. ‘Supplying Information to Facilitate Price Discrimination,’
35 International Economic Review 309–27.
Li, Xing, Megan MacGarvie, and Petra Moser. 2018. ‘Dead Poets’ Property: How Does Copyright Influence
Price?,’ 49 The RAND Journal of Economics 181–205.
Lichtman, Douglas. 2000. ‘Property Rights in Emerging Platform Technologies,’ 29 Journal of Legal Studies
615–44.
Liebeler, Lars H. 1987. ‘Trademark Law, Economics and Grey Market Policy,’ 62 Indiana Law Journal 753–77.
Liebowitz, Stanley J. 1985. ‘Copying and Indirect Appropriability: Photocopying of Journals,’ 93 Journal of
Political Economy 945–57.
Liebowitz, Stanley J. 1986. ‘Copyright Law, Photocopying, and Price Discrimination,’ in John Palmer and
Richard O. Zerbe, Jr., eds., Research in Law and Economics: The Economics of Patents and Copyrights,
Stamford, CT: JAI Press Inc.
Liebowitz, Stanley J., and Stephen E. Margolis. 2009. ‘Bundles of Joy: The Ubiquity and Efficiency of Bundles
in New Technology Markets,’ 5 Journal of Competition Law and Economics 1–47.
Lipner, Seth E. 1990. The Legal and Economic Aspects of Gray Market Goods, Westport, CT: Praeger Publishing.
Litwak, Mark. 1997. ‘Forward,’ in Tiiu Lukk, Movie Marketing: Opening the Picture and Giving it Legs,
Hollywood, CA: Silman-James Press, Inc.

Liu, Qihong, and Konstantinos Serfes. 2013. ‘Price Discrimination in Two-Sided Markets,’ 22 Journal of
Lunney, Glynn S. Jr. 1996. ‘Reexamining Copyright’s Incentives-Access Paradigm,’ 49 Vanderbilt Law Review
483–656.
Lunney, Jr., Glynn S. 2008. ‘Copyright’s Price Discrimination Panacea,’ 21 Harvard Journal of Law and
Technology 387–456.
Malueg, David A. 1993. ‘Bounding the Welfare Effects of Third-Degree Price Discrimination,’ 83 American
Malueg, David A., and Marius Schwartz. 1994. ‘Parallel Imports, Demand Dispersion, and International Price
Discrimination,’ 37 Journal of International Economics 167–95.
Manta, Irina D., and David S. Olson. 2015. ‘Hello Barbie: First They Will Monitor You, then They Will
Discriminate against You. Perfectly,’ 67 Alabama Law Review 135–87.
Maskin, Eric, and John Riley. 1984. ‘Monopoly with Incomplete Information,’ 15 The RAND Journal of
Economics 171–96.
Maskus, Keith E. 2001. ‘Parallel Imports in Pharmaceuticals: Implications for Competition and Prices in
Developing Countries,’ Final Report to World Intellectual Property Organization, accessed December 20, 2018
at http://www.wipo.int/export/sites/www/about-ip/en/studies/pdf/ssa_maskus_pi.pdf.
Meese, Alan J. 1997. ‘Tying Meets the New Institutional Economics: Farewell to the Chimera of Forcing,’ 146
University of Pennsylvania Law Review 1–98.
Merges, Robert P. 1997. ‘The End of Friction? Property Rights and Contract in the “Newtonian” World of
On-Line Commerce,’ 12 Berkeley Technology Law Journal 115–36.
Meurer, Michael J. 1997. ‘Price, Personal Use and Piracy: Copyright Protection of Digital Works,’ 45 Buffalo
Meurer, Michael J. 2003a. ‘Too Many Markets or Too Few: Copyright Policy towards Shared Works,’ 77
Southern California Law Review 903–74.
Meurer, Michael J. 2003b. ‘Vertical Restraints and Intellectual Property Law: Beyond Antitrust,’ 87 Minnesota
Mortimer, Julie Holland. 2008a. ‘Price Discrimination, Copyright Law, and Technological Innovation: Evidence
from the Introduction of DVDs,’ 122 Quarterly Journal of Economics 1307–50.
Mortimer, Julie H. 2008b. ‘Vertical Contracts in the Video Rental Industry,’ 75 The Review of Economic Studies
165–99.
Nahata, Babu, Krzysztof Ostaszewski, and Prasanna K. Sahoo. 1990. ‘Direction of Price Changes in Third-
Degree Price Discrimination,’ 80 American Economic Review 1254–8.
Nalebuff, Barry. 2004. ‘Bundling as an Entry Barrier,’ 119 Quarterly Journal of Economics 159–87.
Netanel, Neil Weinstock. 1996. ‘Copyright and a Democratic Civil Society,’ 106 Yale Law Journal 283–387.
Netanel, Neil Weinstock. 1998. ‘Asserting Copyright’s Democratic Principles in the Global Arena,’ 51 Vanderbilt
Ning, Lizhi, Shubha Ghosh, and Wei Zhou. 2015. ‘Price Discrimination in Patent Licensing and the Application
of FRAND,’ 3 Journal of Antitrust Enforcement i207–27.
Novos, Ian E., and Michael Waldman. 1984. ‘The Effects of Increased Copyright Protection: An Analytic
O’Rourke, Maureen A. 1997. ‘Copyright Preemption after the ProCD Case: A Market-Based Approach,’ 12
Perzanowski, Aaron, and Chris J. Hoofnagle. 2016. ‘What We Buy When We “Buy Now,”’ UC Berkeley Public
Law Research Paper, No. 2778072.
Perzanowski, Aaron, and Jason Schultz. 2010. ‘Digital Exhaustion,’ 58 UCLA Law Review 889–946.
Peterman, Brian W. 1993. ‘The Gray Market Solution: An Allocation of Economic Rights,’ 28 Texas
International Law Journal 159–90.
Philips, Louis. 1981. The Economics of Price Discrimination, Cambridge: Cambridge University Press.
Pigou, Arthur, C. 1932. Economics of Welfare, 4th edn, London: MacMillan and Co., Limited.
Posner, Richard A. 1975. ‘The Social Costs of Monopoly and Regulation,’ 83 Journal of Political Economy
807–28.
Posner, Richard A. 1986. Economic Analysis of Law, 3rd edn, Boston, MA: Little, Brown, and Co.
Posner, Richard A. 2001. Antitrust Law, 2nd edn, Chicago, IL: University of Chicago Press.
Posner, Richard A. 2005. ‘Vertical Restraints and Antitrust Policy,’ 72 The University of Chicago Law Review
229–41.
Raff, Horst, and Nicolas Schmitt. 2007. ‘Why Parallel Trade May Raise Producers’ Profits,’ 71 Journal of
International Economics 434–47.

Rai, Arti K. 2001. ‘The Information Revolution Reaches Pharmaceuticals: Balancing Innovation Incentives,
Cost, and Access in the Post-Genomics Era,’ 2001 University of Illinois Law Review 173–210.
Reichman, Jerome H., and Jonathan A. Franklin. 1999. ‘Privately Legislated Intellectual Property Rights:
Reconciling Freedom of Contract with Public Good Uses of Information,’ 147 University of Pennsylvania
Robinson, Joan. 1933. The Economics of Imperfect Competition, London: Macmillan.
Roin, Benjamin N. 2014. ‘Intellectual Property versus Prizes: Reframing the Debate,’ 81 The University of
Chicago Law Review 999–1078.
Rubin, Harry. 1992. ‘Destined to Remain Grey: The Eternal Recurrence of Parallel Imports,’ 26 International
Law 597–622.
Schmalensee, Richard. 1981. ‘Output and Welfare Implications of Monopolistic Third-Degree Price
Discrimination,’ 71 American Economic Review 242–7.
Schwartz, Marius. 1990. ‘Third-Degree Price Discrimination and Output: Generalizing a Welfare Result, 80
Scotchmer, Suzanne. 1991. ‘Standing on the Shoulders of Giants: Cumulative Research and Patent Law,’ 5
Scotchmer, Suzanne. 2004. Innovation and Incentives, Cambridge, MA: MIT Press.
Shapiro, Carl, and Hal R. Varian. 1999. Information Rules: A Strategic Guide to the Network Economy, Brighton,
MA: Harvard Business Review.
Shepard, Andrea. 1991. ‘Price Discrimination and Retail Configuration,’ 99 Journal of Political Economy 30–53.
Shiller, Benjamin Reed. 2014. ‘First-Degree Price Discrimination Using Big Data,’ Working Papers 58, Brandeis
University, Department of Economics and International Business School.
Shy, Oz, and Jacques-Francois Thisse. 1999. ‘A Strategic Approach to Software Protection,’ 8 Journal of
Sidak, J. Gregory. 1981. ‘Rethinking Antitrust Damages,’ 33 Stanford Law Review 329–52.
Spinks, S.O. 2000. ‘Exclusive Dealing, Discrimination, and Discounts under EC Competition Law,’ 67 Antitrust
Sterk, Stewart E. 1996. ‘Rhetoric and Reality in Copyright Law,’ 94 Michigan Law Review 1197–249.
Stigler, George. 1987. A Theory of Price, New York, NY: Macmillan.
Stokey, Nancy L. 1979. ‘Intertemporal Price Discrimination,’ 93 The Quarterly Journal of Economics 355–71.
Stole, Lars A. 1995. ‘Nonlinear Pricing and Oligopoly,’ 4 Journal of Economics and Management Strategy
529–62.
Stole, Lars A. 2003. ‘Price Discrimination and Imperfect Competition,’ in Richard Schmalensee and Robert
Willig, eds., Handbook of Industrial Organization, vol. I, Amsterdam, Netherlands: North Holland Publishing.
Stole, Lars A. 2007. ‘Price Discrimination and Competition,’ 3 Handbook of Industrial Organization 2221–99.
Sullivan, Lawrence A. 1977. Handbook of the Law of Antitrust, St. Paul, MN: West Publishing Co.
Sykes, Alan O. 2002. ‘TRIPS Pharmaceuticals, Developing Countries, and the Doha “Solution,”’ 3 Chicago
Journal of International Law 47–68.
Takeyama, Lisa. 1994. ‘The Welfare Implications of Unauthorized Reproduction of Intellectual Property in the
Presence of Network Externalities,’ 62 The Journal of Industrial Economics 155–66.
Takeyama, Lisa N. 1997. ‘The Intertemporal Consequences of Unauthorized Reproduction of Intellectual
Property,’ 40 The Journal of Law and Economics 511–22.
Tehranian, John. ‘Infringement Nation: Copyright Reform and the Law/Norm Gap,’ 2007 Utah Law Review
537–50.
Thisse, Jacques, and Xavier Vives. 1988 ‘On the Strategic Choice of Spatial Price Policy,’ 78 American Economic
Review 122–37.
Tirole, Jean. 1988. The Theory of Industrial Organization, 1st edn, Cambridge, MA: MIT Press.
Varian, Hal R. 1985. ‘Price Discrimination and Social Welfare,’ 75 American Economic Review 870–75.
Varian, Hal R. 1989. ‘Price Discrimination,’ 1 Handbook of Industrial Organization 597–654.
Varian, Hal R. 2000. ‘Buying, Sharing, and Renting Information Goods,’ 8 Journal of Industrial Economics
473–88.
Viscusi, W. Kip, Joseph E. Harrington, and John M. Vernon. 1992. Economics of Regulation and Antitrust,
Lexington, MA: D.C. Health and Co.
Whinston, Michael D. 1990. ‘Tying, Foreclosure, and Exclusion,’ 80 American Economic Review 837–59.
Wolinsky, Asher. 1987. ‘Brand Names and Price Discrimination,’ 35 The Journal of Industrial Economics 255–68.
Wright, Joshua D. 2005. ‘Missed Opportunities in Independent Ink,’ 2005–06 Cato Supreme Court Review
333–59.
Wu, Tim. 2008. ‘Tolerated Use,’ 31 Columbia Journal of Law & the Arts 617–36.
Wyatt, Justin. 1994. High Concept: Movies and Marketing in Hollywood, Austin, TX: University of Texas Press.
Yoo, Christopher S. 2014. ‘Public Good Economics and Standard Essential Patents,’ Faculty Scholarship Paper,
No. 1371, accessed on June 14, 2018 at http://scholarship.law.upenn.edu/faculty_scholarship/1371.

Legislation
17 U.S.C. § 106(4), 109(a), 110, 602(a).

35 U.S.C. § 154(a), 261, 271(a).
17 U.S.C.A. § 117(a), (c)–(d) (West 1999).
Case Law
Akzo v. Int’l Trade Comm., 808 F.2d 1471 (Fed. Cir. 1986).
Bloomer v. McQuewan, 55 U.S. 539 (1852).
Broadcast Music, Inc. v. Columbia Broadcast. Sys. Inc., 441 U.S. 1 (1979).
Brulotte v. Thys Co., 379 U.S. 29 (1964).
Eastman Kodak Co. v. Image Tech. Servs., Inc., 504 U.S. 451 (1992).
IBM Corp. v. United States, 298 U.S. 131 (1936).
Illinois Tool Works Inc. v. Independent Ink, Inc., 547 U.S. 1281 (2006).
Impression Products, Inc. v. Lexmark International, Inc., 137 S. Ct. 1523 (2017).
International Salt Co. v. United States, 332 U.S. 392 (1947).
Jefferson Parish Hosp. Dist. No. 2 v. Hyde, 466 U.S. 2 (1984).
K-Mart Corp. v. Cartier, Inc., 486 U.S. 281 (1987).
Kirtsaeng v. John Wiley & Sons, 568 U.S. 519 (2013).
MAI Systems v. Peak Computer, 991 F.2d 511 (9th Cir. 1993).
Motion Pictures Patents Co. v. Universal Film Mfg. Co., 243 U.S. 502 (1917).
ProCD, Inc. v. Zeidenberg, 86 F.3d 1447 (7th Cir. 1996).
Quality King Distributors, Inc. v. L’anza Research International, Inc., 523 U.S. 135 (1998).
Sega v. Accolade, 977 F.2d 1510 (9th Cir. 1992).
Sony Corp. of America v. Universal City Studios, Inc., 464 U.S. 417 (1984).
U.S. Steel Corp. v. Fortner Enterprises, Inc., 429 U.S. 610 (1977).

11. When are IP rights necessary? Evidence from
innovation in IP’s negative space
Kal Raustiala* and Christopher Jon Sprigman**
3 34
Contents
I. Introduction
II. The Concept of Negative Space
III. Mapping the Negative Space of IP
A. Social Norms
B. First-Mover Advantage
1. Football
2. Computer databases
3. Software
C. Products versus Performances
D. Creativity-Enhancing Copying
E. Market Power Unrelated to IP
IV. Implications and New Directions
A. The Negative Space Literature and the Law’s Unintended Consequences
B. Unexplored Negative Space—Industrial Design
V. Conclusion
References
I. INTRODUCTION
The law and economics of intellectual property has long rested on a foundational, if
implicit, premise: that intellectual property (IP) law is best understood by studying how
legal rules operate in actual markets for creative work. To understand copyright, for
instance, scholars have explored copyright-dependent fields such as music, film, and pub-
lishing, seeking to understand how copyright law shapes innovation in these particular
contexts. In other words, IP scholars have generally studied creative fields that rely on IP.
The assumption that we best understand a body of legal rules by looking at the fields
they directly address is eminently reasonable and seemingly so obvious that, until recently,
** Professor and Director of the UCLA Ronald W. Burkle Center for International Relations,
UCLA Law School.
** Professor, New York University School of Law and Co-Director of the Engelberg Center
on Innovation Law and Policy. Thanks to our research assistants, Sydney Sherman and Jaclyn A.
Hall, and to the Filomen D’Agostino and Max E. Greenberg Research Fund and the University of
California Faculty Senate for grants that supported this work.
309

it was almost never questioned. Yet this way of analysing the effects of IP is incomplete. IP
rules are purposive; the core goal and rationale of IP is to incentivize creative and innovative
production. Put differently, IP rights are granted and enforced by the state because they are
believed to have significant effects on behavior, spurring innovation that would not occur in
their absence. Thought of this way, looking only at the creative fields that IP law addresses
is misguided.1 It leaves open, or at least incomplete, a host of important and fascinating
questions. Can innovation flourish in the absence of IP protection? Can market incentives,
psychological factors, social norms, first-mover advantages, or any number of other causes,
including path-dependency or even happenstance, serve as whole or partial substitutes for
IP rights? And is it possible that, under some conditions and in some industries, IP protec-
tion is counterproductive—that is, it inhibits more innovation than it promotes?
To answer these questions, we must look beyond the space affirmatively covered by IP
law. We must look instead at what we call the ‘negative space’ of IP. By ‘negative space’
we mean those creative and innovative fields that, for historical, doctrinal, or other
reasons, are not addressed by IP law (Raustiala and Sprigman, 2006). In fashion, cuisine,
tattoo artistry, professional magic, financial services, sports, and many other innovative
endeavors, IP rights are absent or highly limited. IP as a causal variable either is out of
the picture or is marginalized, even if IP rights sometimes apply as a formal matter. As a
result, these fields provide an illuminating and sometimes arresting take on the relation-
ship between IP and innovation.
In this chapter, we explore the concept of IP’s negative space and the scholarship
that has begun to grow around it. This literature is barely more than a decade old, but
in important ways it has shaped how scholars today think about IP. The negative space
literature is predominantly focused on copyright, but there is no reason it cannot apply
to at least some industries that are within the domain of patent, which shares copyright’s
incentive justification.2 Because the underlying rationales for trademark are so different,
insights for trademark law are more limited. But the reverse is not true. One implication
of much of the negative space scholarship is that the role of brands and marks in shaping
innovation and innovative industries may be quite powerful, especially in those fields
where copyright and patent are absent or weak.
In the remainder of this chapter we first briefly define the concept of negative space.
We then explore what has been learned thus far, surveying the growing range of research
into ‘intellectual production without intellectual property’ (Dreyfuss, 2010). We conclude
with an assessment of this line of inquiry and offer some key questions for future research.
II. THE CONCEPT OF NEGATIVE SPACE
IP rights, particularly in the American context, are fundamentally incentive-based. They

are predicated on the theory that restraints on copying are necessary to motivate c reativity,
1
In social science terms, it is selecting on the dependent variable. Gary King, Robert
Keohane, and Sidney Verba, Designing Social Inquiry: Scientific Inference in Qualitative Research,
section 4.3.1 (1994).
2
Indeed, a number of works have done so. See, e.g., Rosenblatt (2013), discussing negative
space as applied to the creation of innovative medical techniques.

When are IP rights necessary? 311
lest copyists seize the returns from creation and destroy the incentive to create in the first
place. The degree to which this theory is true is a question of tremendous importance for
both law and policy. Innovation is a substantial part of any flourishing contemporary
economy; it is arguably the key source of growth in advanced economies today. And it is
widely believed that innovation rests upon—and indeed requires—strong IP protection.
Given this, it is surprising how little empirical evidence supports this theory. Indeed,
some prominent IP scholars have gone so far as to call IP’s rationale fundamentally ‘faith-
based’ (Lemley, 2016). The most important task that law and economics research can
undertake in the IP field is to shed light on how well IP law’s core incentives model holds
up in the real world. The study of IP’s negative space is part of this endeavor. If healthy
innovation is observed in IP’s negative space, understanding why is essential.
IP’s negative space is perhaps most simply identified by contrasting it with IP’s positive
space. The positive space encompasses all those creative activities that IP law addresses,
such as novels, poems, films, television shows, music, software, painting, and video
games.3 The negative space of IP, by contrast, encompasses any other creative art, craft,
or act that does not enjoy or at least does not ordinarily rely on IP rights against copyists,
either because IP is formally inapplicable or because something—perhaps a social norm
against IP enforcement, or a legal or economic barrier that discourages resort to formal
IP—limits its salience.
Jessica Litman (1994) was characteristically prescient in identifying such areas of
creativity as potentially valuable objects of study. In her article The Exclusive Right to
Read, Litman suggested an interesting counterfactual:
Imagine for a moment that some upstart revolutionary proposed that we eliminate all intellectual
property protection for fashion design. No longer could a designer secure federal copyright
protection for the cut of a dress or the sleeve of a blouse. Unscrupulous mass-marketers
could run off thousands of knock-off copies of any designer’s evening ensemble, and flood
the marketplace with cheap imitations of haute couture . . . The dynamic American fashion
industry would wither, and its most talented designers would forsake clothing design for some
more remunerative calling like litigation. And all of us would be forced either to wear last year’s
garments year in and year out, or to import our clothing from abroad.
Then Litman brought the argument home: ‘Of course, we don’t give copyright protection
to fashions . . . We never have.’4 Despite this, the fashion industry remains creative and
economically vibrant: nothing like Litman’s dystopian and satirical scenario.
Doctrinally, the reason fashion is unprotected by IP is that clothing is viewed as
functional, and functional items, under IP law, cannot be protected by copyright.5 Yet
3
For the purposes of this chapter, we will broadly define this realm consistent with American
IP law and generally focus on copyright. But, of course, there are some important variations
elsewhere in the world, and we will mention one or two along the way.
4
In the United States. Elsewhere in the world, fashion design is often covered by copyright or
other IP laws, but the net effect of those laws is very hard to discern, and the American industry is
certainly just as vibrant and successful as its major foreign counterparts.
5
Patent, by contrast, is intended to cover functional inventions. But the bar for patentability
is so high, and the fashion industry’s cycle of innovation so fast (partly endogenously; see Raustiala
and Sprigman (2006) for more) that patent is almost never used for apparel outside of a few special-
ized items like waterproof fabrics. The one small exception, perhaps growing, is the use of design

the field features sustained and high levels of creativity and investment in the creation of
new works. The ability of the fashion industry to continually produce creative work runs
counter, as Litman noted, to the conventional wisdom that IP rights are essential to spur
investment in the creation of new works.6
In years past others too have observed that some areas of creative endeavor—areas such
as cuisine, that are arguably just as dynamic and creative as music or motion pictures—
were not protected by copyright (Das, 2000; Jesien, 2007; Pollack, 1991). For the most
part this fact was simply mentioned; occasionally it was bemoaned. Apparel design in
particular was the subject of several articles in the 1990s and 2000s that decried the lack
of protection for fashion designs and generally proposed some doctrinal fix (Mencken,
1997; Sanchez-Roig, 1989). This work had value in that it identified, sometimes implicitly,
a gap in our understanding regarding how IP operates empirically. Yet this early literature
generally failed to take seriously the idea that these innovative and unusual fields might
offer important evidence about the strength of IP’s incentive justification. IP’s negative
space was largely treated as a curiosity or a problem to be fixed through doctrinal change;
not as a unique window on foundational questions.
What distinguishes the newer line of research we focus on in this chapter is precisely
this orientation: a concern not only with delineating how hitherto un- or under-explored
creative fields operate, but also with examining the larger implications of negative space
for our understanding of IP theory, doctrine, and practice. By looking at a sufficient
number of low-IP creative areas, we can perhaps begin to understand the various ways in
which innovation incentives can succeed or fail. And ultimately, it may be possible to say
something deeper about the strength and breadth of IP’s incentives justification. While
this may not be the goal of all the scholarship we gather here under the rubric of negative
space, we think it is important to most—and is the primary reason this research has drawn
so much attention.
At a foundational level this line of scholarship draws deeply from the well of ideas
associated with Robert Ellickson (1991) and his influential book, Order without Law.
Ellickson famously studied the cattle ranchers of Shasta County, California, a community
that one would predict would be aware of and would rely upon formal rules of property.
(Cattle stray, and, when they do, they damage fences and crops.) Ellickson found that
the Shasta County ranchers often behaved as if the legal rules were irrelevant to their
disputes. The ranchers instead developed and enforced a set of social norms regarding
responsibility for straying cattle.
Some of these norms looked efficient relative to the formal property rules they dis-
placed; some did not. The central point was that particular communities can and often
patents. Design patents differ from ordinary utility patents in that they cover ornamental designs.
We have seen increased use in areas such as athletic and ‘athleisure’ wear, but to date very little use
in conventional apparel.
6
A few years later, Lloyd Weinreb (1998, p. 1235) similarly noted that:
It may be difficult to imagine how, in an industry in which copyright is taken for granted,
authors and the commercial marketers of their works could survive if copyright were eliminated.
But, without more to go on, one may question whether that does not mistake the familiar for the
necessary. The fashion industry has thrived despite the absence of protection for designs and the
prevalence of ‘knock-offs.’

do achieve order using social norms rather than formal rules. In Ellickson’s study, law
certainly existed, and was sometimes even employed (or served as a backdrop, casting
a shadow over informal norms) (Mnookin and Kornhauser, 1979). But for the most
part, legal rules were not significant to the settlement of actual disputes. Much like the
relational contracting literature of the 1960s, Ellickson found that the informal norms
developed within a given community mattered far more (Macaulay, 1963).
Ellickson’s work, and much of the now-substantial literature on law and social norms
that he helped spawn, had nothing directly to do with IP (Bernstein, 1992; Posner, 2002).
Yet it has guided much of the growing negative space scholarship. Scholarship on negative
space highlights the limitations of law, the frequent disconnect between formal law and
actual social practice, and the importance of careful empirical and even anthropological
research into how markets and communities work. It focuses directly on the creative
ways communities, industries, and individuals structure relations without relying on legal
rules or institutions. Most importantly, negative space scholarship reverses the lens of
traditional IP scholarship. Rather than study how IP law works (or does not), it looks
where IP is not and tries to understand how innovation occurs nonetheless.
III. MAPPING THE NEGATIVE SPACE OF IP
What falls within and what falls without IP’s domain remains largely unexplained. In
this sense one of the most fundamental questions about IP’s negative space—what
explains its contours—remains ripe for future research. (We say more about this below.)
While some theories of the scope of IP’s negative space have been advanced, empirical
support remains, as of this writing, unfortunately scant (Raustiala and Sprigman, 2012;
Rosenblatt, 2011). That fundamental questions remain unanswered is not, of course,
unusual in IP. Whether, when, or even where IP rules should be tightened or relaxed is
similarly unclear. One could say the same about IP’s entire incentive justification, though,
as we describe below, the recent wave of negative space scholarship has certainly shed
some light on this question.
Whatever larger insights have emerged from studies of negative space, this literature has
undeniably generated a clutch of fascinating case studies. At a minimum, these studies
substantially enhance our understanding of the empirics of innovation across a strikingly
wide range of human endeavor. Studies of the fashion industry (Barnett, 2005; Barnett,
et al., 2010; Hemphill and Suk, 2009, 2014; Raustiala and Sprigman, 2006, 2009), cuisine
(Buccafusco, 2007; Fauchart and von Hippel, 2008), fan fiction (Tushnet, 2009), por-
nography (Darling, 2014), nineteenth-century US commercial publishing (Spoo, 2013),
video games featuring significant user-generated content (Lastowka, 2014), stand-up
comedy (Oliar and Sprigman, 2008), roller derby (Fagundes, 2012; Magliocca, 2009),7
software (Benkler, 2007; Fisk, 2006; Garon, 2010; Lerner and Tirole, 2005), jam bands
(Schultz, 2006), tattoos (Perzanowski, 2013), magic (Loshin, 2010), and the flu vaccine
(Kapczynski, 2016) detail an extraordinary variety of creative and innovative work, and
7
See also Magliocca (2009), discussing industry norms against patenting and arguing that
business method patents should not be expanded to cover industries where such norms exist.

show the ways in which creative production can flourish with relatively little or no IP
protection.8 Related studies of scientific innovation document communal practices that
emphasize sharing, and resist the full potential for propertization of research (Murray et
al., 2009; Strandburg, 2005). And as Eric von Hippel (2005) and others have shown, a lot
of innovation is generated by users, in contexts as varied as extreme sports, surgery, library
science, and commercial high-tech manufacturing, who work mostly in the absence of
IP incentives, and who often willingly and openly share the fruits of their creativity with
others (de Jong and von Hippel, 2009; Gault and von Hippel, 2009).
Relevant markets, incentives, participants, practices, and norms vary enormously
across the negative space literature. In many respects the only shared quality is that crea-
tive production occurs in an environment that can be characterized as at least low-IP, and
sometimes no-IP. Still, some broad features across cases can be discerned, and a certain
taxonomic order can be imposed. In what follows, we briefly describe some of this work
and highlight its broad themes, commonalities, and contrasts.
A. Social Norms
Many negative space studies have documented the powerful role social norms play in
stimulating innovation and constraining appropriation. In many of these cases IP law is
formally relevant but for disparate reasons is often displaced, in an Ellicksonian fashion,
by the social norms of a particular creative community. These social norms are almost
exclusively producer norms, and typically reflect a shared sense of professional or artistic
identity that allows such norms to develop and become entrenched.9 A large body
of work has developed over the past decade analysing creative industries that feature
relatively robust social norms.
For example, an early study by one of us (Sprigman) and Dotan Oliar (2008) on
copying among stand-up comedians is in this vein. Technically copyright law covers
comedy. The issue then, as now, is that copyright protection only adheres to a specific
formulation of a joke or routine, and not to the general premise or structure. In practice
this means that a joke can easily be rewritten; hence infringement claims are thus difficult
to bring and rarely cost-effective. Oliar and Sprigman (2008) describe the development
of social norms among comics through two eras: the post-vaudeville era of joke slingers
like Henny Youngman, Milton Berle, and Phyllis Diller, and the modern age of personal-
ized comedy which started with people like Lenny Bruce and Mort Sahl and is today
represented by comedians such as Sarah Silverman, Amy Schumer, and Louis C.K.
The post-vaudeville era was marked by two features. The currency of comedy was the
joke (understood as a one-liner, or maybe a simple premise + punchline). And comedians
of the day freely appropriated whatever jokes they desired. Jokes were part of a commons
that comedians could access. A famed Milton Berle quip summed up this state of affairs:
Berle, known as ‘the Thief of Bad Gags,’ would come up on stage and say about the
8
See also the collection of essays in Making and Unmaking Intellectual Property (Mario
Biagioli, Peter Jaszi and Martha Woodmansee, eds., Chicago: University of Chicago, 2011).
9
An interesting exception is in the jam-band community associated with acts such as Phish
and the Grateful Dead (Schultz, 2006).

previous act, ‘I laughed so hard I nearly dropped my pencil.’ In short, post-vaudeville era
comics operated on a norm of open copying.
By the 1960s, however, comedy was undergoing major change. Whereas the first era
was largely about the rapid-fire slinging of jokes, the second era featured longer, indi-
vidualized narratives that were tied to a particular comedic persona. And as the nature of
comedy changed, so too did the social norms. Comedians still do not sue one another over
copying. Instead, contemporary comics operate via a set of well-developed and widely
accepted social norms governing appropriation. Oliar and Sprigman describe norms that
address copying, ownership, co-authorship, and joke transfer. These norms are backed
by a surprisingly stiff regime of community-imposed, extra-legal sanctions. For the most
part, these sanctions are reputational, but they appear to be effective in policing joke-
stealing. On rare occasions, norm-breakers are subject to group boycotts (i.e., cooperative
refusal by comedians to work with a perceived joke thief) and even violence. The overall
result is a fairly robust, but thoroughly private and extra-legal, system of rules governing
the appropriation of creative work.10 And comedy remains a highly creative and very
productive field.
In a similar vein, Aaron Perzanowski (2013) has studied tattoo artists and the owner-
ship of tattoo designs. As with comedy, nominally tattoo designs fall within copyright’s
positive space. But in practice, they are rarely if ever governed by the laws of IP.11 Instead,
a complex set of social norms operates to govern tattoo artists and their designs. There
are, Perzanowski argues, five core norms:
First, tattooers as a rule recognize the autonomy interests of their clients both in the design of
custom tattoos and their subsequent display and use. Second, tattooers collectively refrain from
reusing custom designs—that is, a tattooer who designs an image for a client will not apply that
same image on another client. Third, tattooers discourage the copying of custom designs—that
is, a tattooer generally will not apply another tattooer’s custom images to a willing client.
Fourth, tattooers create and use pre-designed tattoo imagery, or ‘flash’, with the understanding
that it will be freely reproduced. Finally, tattooers generally embrace the copying of works that
originate outside of the tattoo industry, such as paintings, photos, or illustrations. In some ways,
these norms unintentionally echo familiar concepts from copyright law, but they differ from
formal law in important respects as well.
As with comics, tattoo artists around the United States are generally familiar with and
subscribe to these norms, despite the fact that there is no legal sanction for failing to do
so.12
Looking further back in history, in his book Without Copyrights, Robert Spoo (2013)
tells the fascinating story of the nineteenth-century American publishing industry, which
operated in a legal regime where foreign works were largely unprotected by US law and
free to be pirated. Conventional copyright theory predicts that the low-IP environment
10
As this suggests, social norms systems are subject to serious problems of due process;
scholars in this tradition do not necessarily view them as normatively superior.
11
Litigation is rare but not nonexistent (Whitmill v. Warner Bros. Entertainment, 4:11-cv-752
(E.D. Mo. 2013)). The case, which settled, involved the appearance of a tattoo look-alike, copied
from boxer Mike Tyson, on the face of actor Ed Helms.
12
An interesting question is whether there is a difference in the power or scope of such norms
since comics travel widely and tattoo artists are generally local.

Spoo describes will lead to chaos: a free-for-all that provides no return to foreign authors
and publishers. What Spoo finds instead is quite a bit of order without law. He describes a
detailed and fluid system of ‘trade courtesy’ under which American publishers made what
looked like purely gratuitous payments to foreign publishers and foreign authors, and
agreed among themselves to a sort of ‘first-among-pirates’ rule governing distribution of
foreign works in the US. It is by no means certain that the trade courtesy system Spoo
describes provided the level of return to foreign authors that the expanded US copyright
system eventually would. Nonetheless, trade courtesy provided substantial recompense to
foreign authors, illustrating that the alternative to copyright is not unmitigated piracy—
nor the collapse of creative work.
High-end chefs also exhibit substantial adherence to social norms governing appropria-
tion. Whereas tattoo designs and comedy routines are formally covered by copyright, if
in practice rarely addressed by it, cuisine is much like fashion: it is unprotected in its two
core features: the recipe and the finished dish. A creative chef who develops an original
and delicious recipe can try to keep it from rivals as a trade secret. But since talented chefs
can often reverse-engineer a dish they have eaten, the utility of this strategy is limited.
Moreover, for many top chefs, cookbooks provide substantial income and are part of a
larger brand-building effort. And by definition, cookbooks supply recipes for any reader
to recreate. In short, recipes are legally open to appropriation by others.
In practice, however, chefs—at least at the higher levels—have developed a set of norms
and practices to constrain appropriation and reward ingenuity. Both Chris Buccafusco
(2007) and the team of Emmanuelle Fauchart and Eric von Hippel (2008) have docu-
mented these practices and the ways they sustain high levels of culinary innovation. In
particular, they note the strong norm of attribution, in which a given dish may be copied
freely as long as the originator is noted (sometimes directly on the menu; even more often
in a cookbook recipe’s notes). For example, Figure 11.1 shows famed New York chef
David Chang citing directly on his menu the pioneering New York chef Wylie Dufresne
of the now-defunct wd-50 restaurant.
Many diners would surely miss the reference; it is a dog whistle to a special few. But
for the intended audience—other chefs, culinary professionals, food fanatics—it is an
important signal that Chang, himself a renowned innovator, has copied a terrific dish
from Dufresne, and that he acknowledges the innovation it contains. Attribution is the
central value here, not exclusivity, and chefs, unlike comedians, are far more comfortable
with appropriation—as long as it is acknowledged as such.
In short, social norms have been found to play a powerful, innovation-facilitating role
in a host of creative industries in the negative space, from chefs and comedians to roller
derby, magicians, and jam bands (Fagundes, 2012; Schultz, 2006). To be sure, many of
these industries are small, though not all are. Modest size appears to help create the sort
of community of interest and sense of shared professional identity that can originate
and perpetuate robust social norms. And it is important to underscore that in most of
these studies these norms develop and are sustained among producers of creative content.
To the degree content producers lack a shared professional identity norms may tend
to break down, or perhaps work only in a local or regional context. Or, as with chefs,
norms may segment by category. Truly world-class chefs, for example, are loathe to copy
because much of their reputation rests on their innovative capacity. And when they do,
as the Chang menu above illustrates, they tend to attribute creations to originators. But

Figure 11.1 Menu, Momofuku Daisho, Toronto
further down the food chain, these norms are less apparent and copying becomes much
more prevalent. If you have ever eaten a ‘molten chocolate cake,’ you have observed this
process in action.
One more important limitation must be stressed. In much of the research into social
norms in IP’s negative space the industries involved tend to feature relatively low costs of
innovation. Low-cost innovation is important because it is likely to facilitate the effective
operation of social norms. If innovation were extremely expensive—as is the case, for exam-
ple, in pharmaceutical development or blockbuster filmmaking—the power of producer-
policed, reputation-based social norms might be wholly insufficient to constrain copying
to an appreciable degree. In short, the extant literature suggests, if only by the topics most
widely covered, that social norms are more robust and meaningful in creative fields where
the field is small and ideally tight-knit and investment in creation is relatively low.
B. First-Mover Advantage
Sometimes industries rely on first-mover advantage to create and preserve innovation

incentives. This segment of the negative space literature is embryonic; much work remains
to be done to investigate the range of creative industries in which first-mover advantage
may conceivably play a role in incentivizing or otherwise protecting and spurring innova-
tion. But at very least, the few case studies that we have thus far suggest that the area
merits exploration.
The concept of first-mover advantage is often understood to mean the period of de
facto exclusivity that an innovator enjoys due to the practical difficulties of copying a

particular innovation. In other words, if it takes time for copyists to successfully copy a
creation, the creator may have a first-mover advantage—particularly if being first can
provide enough of a head start to lock in markets or at least make it hard for latecomers to
compete. In some cases, first-mover advantage can offer a sufficient incentive to engage in
meaningful innovation, even without the prospect of IP rights to protect that innovation.
This should not be surprising. First-mover advantage is at the core of IP rights: the
central feature of IP rights is that they extend any period of de facto exclusivity by making
it illegal to copy an innovation for a set period of time. Patent and copyright are not per-
petual; anyone can copy the work of another innovator eventually, but the law is meant to
be calibrated so that the gain to the first mover (i.e., the creator/inventor) is large enough
that it will incentivize continued innovation. Under IP law, the state, in short, creates and
enforces a specified period of first-mover advantage.
The interesting question here is how much a creator can benefit from ‘natural’ (that
is, nonlegal) barriers to copying, what we will simply refer to as first-mover advantage,
as opposed to those barriers created by legal restraints. In previous work, we described
several examples of first-mover advantage incentivizing significant innovation in the
absence of IP (Raustiala and Sprigman, 2012). There are likely other examples, and future
work may seek to identify and assess creative work for which first-mover advantage is an
important contributor to innovation incentives. Here we very briefly summarize a few
prominent first-mover examples.
1. Football
American football features substantial innovation, and first-mover advantage, coupled
to strong competitive incentives, seems to be a primary driver. New formations and
plays can offer meaningful advantages to creative coaches and teams, even though
nothing in American law stops other teams from copying those innovations. In fact,
some football coaches openly teach other coaches their plays and approach. The precise
mechanisms here are not well established. But it appears that coaches keep innovating
despite the prospect of copying because they face short-term incentives to win a game
every week and because winning now trumps the possibility of losing over the longer
term as (hypothetically) their idea spreads. But there is another reason that copying does
not deter innovation in football. Football formations and plays sometimes depend on a
certain kind of team and player, and teams cannot be reconstituted quickly. Given this,
an innovative coach can achieve substantial success in the near term with a new formation
even if opponents ultimately adopt it later. The window in which the innovating team is
the only one using the new system—or at least, using it well—may be large enough to
make continued innovation worthwhile.
2. Computer databases

Databases also exhibit some degree of first-mover advantage. Like an innovative football
team deploying a new formation, a successful database can remain competitive due to the
need to train users in the new interface. We can explain this with an example from our own
professional lives. As law professors, we rely heavily on legal databases such as Westlaw.
These databases charge paying customers a substantial fee, and they require extensive
training to learn to use well. That training typically begins in law school, and the big
database companies allow students to use their products for free as a way to get them to

learn—and to become hooked. Once a law student becomes comfortable with Westlaw (or
its primary competitor, Lexis-Nexis) he or she is unlikely to shift to another database. The
result is that even if we create a new database tomorrow with all the federal and state cases
and other materials contained in Westlaw—and lower prices—we will have a hard time
competing with the incumbent firms, who know that lawyers who have spent years, if not
decades, using one system are unlikely to start over unless the savings are very substantial.
That does not mean, of course, that law students who have not yet invested in the learning
required to become skilled in the use of a particular database are not up for grabs. But the
difficulty of penetrating the existing market substantially raises the costs of entry.
3. Software
Probably the most common example of first-mover advantage is found in certain software
markets that exhibit network effects. Being first—and creating a network of users that all
rely on the same program and, as a result, can easily share files or data or documents—can
give decisive and durable advantages to the first mover’s product. And that can lead to
substantial market power and lasting profits.
Only a few industries exhibit such positive network externalities; that is, benefits that
accrue to users from the fact that others are using the same the network. The simplest
example of network externalities is a telephone: a single phone is useless, two phones on
a network are nice, but thousands of connected phones are much, much better. Each
additional phone on the network makes the other phones more valuable. First-mover
advantages can certainly accrue in the absence of positive network externalities. But when
these externalities exist, the power of first-mover advantage is even greater. The ability to
lock consumers in a network that they do not want to leave makes it easier to defeat new
entrants into a market, even those that mimic or improve an existing product.
Think of the short history of social networks. Perhaps Facebook, so dominant today,
will give way to another social network. But Facebook has already seen off a challenge
from Google +, which tweaked the Facebook approach and arguably improved it. People
were hesitant to shift to Google + because their friends are all on Facebook. It is not
impossible to dislodge a leading product even when network externalities exist. Friendster
and Myspace, after all, were pathbreaking firms that were ultimately buried by Facebook.
But it is more difficult.
In sum, first-mover advantage is an important area of future study for scholarship. The
fundamental purpose of copyright and patent is to create first-mover advantage: IP laws
regulate second movers so the first mover has ample time to make money. A goal of future
research should be to document instances where first-mover advantage exists even when
IP law is absent or ineffective, and to determine whether and when first-mover advantage
is powerful enough to sustain a meaningful level of innovation.
C. Products versus Performances
Sometimes industries preserve creative incentives by shifting away from forms of creativ-
ity that are easily copied, refocusing on forms of creativity that are more resistant to
appropriation. The online pornography industry is instructive of this dynamic. Adult
entertainment is currently protected by copyright, though (under American law at least)
prior to the Fifth Circuit’s 1979 decision in Mitchell Bros Film Group v. Cinema Adult

Theater, 604 F.2d 852, 854–5, 858 (5th Cir. 1979), its IP status was unclear (holding that
the Copyright Act neither explicitly nor implicitly prohibits protection of ‘obscene mate-
rials,’ such as the films at issue there, and rejecting the defendant’s affirmative defense of
‘unclean hands’). Perhaps because pornography for so long resided in the negative space
of IP, the industry is a comparatively light user of IP litigation today and lives with very
high rates of free and pirated content.
As Kate Darling (2014) described, ready access to free online content, most notably
through the ubiquitous ‘porntube’ sites,13 has affected the industry’s output of new con-
tent. Darling argues that production in the industry has shifted away from pornographic
feature films and toward cheaper scenes (i.e., shorter bits of recorded pornography, not
embedded in any larger story). These are designed to be viewed, for free, on the porntube
sites, which have entered into deals with many producers to split associated ad revenue.
Darling also documents the rise of ‘cam girls’—women (and men) who perform live over
the internet using webcams. Clients pay to watch these performances, and sometimes
pay more to essentially direct them. The revenue stream that results is resistant to piracy
for much the same reason that live music performances are resistant to piracy—what
is valuable is the immediacy of the live (streamed) performance. This is true even when
the performance is made over an internet connection, because a feature of these perfor-
mances is interactivity—ask (and pay) for the performer to engage in a particular sex act,
and you might receive it.
In earlier work, we argued that a similar phenomenon explained the continuing creativ-
ity of contemporary bartenders, despite the fact that drink recipes, like food recipes, are
wholly unprotected from copying:
Bars and high-end cocktails epitomize this phenomenon of performance over narrowly defined
creative product. Why else do people pay upward of $15 for a drink that may cost less than $2
to make? As a sage bartender once said, you are not really buying a drink, you are renting a bar
stool. And the rent varies, as you would expect, with the quality of the experience. In short, the
high-end bar is a live performance venue. The drink is the ticket to the show. Anything that is a
live performance must be experienced to be appreciated, and that experience can shelter creativ-
ity from the pernicious effects of copying. Why? Because copying all the facets of the experience
is very difficult and often extremely costly—and sometimes impossible, as many would-be
restaurateurs and bar owners have discovered to their peril. (Raustiala and Sprigman, 2012)
When the performance is a significant element of the total offering, the product can
be copied with fewer negative ramifications. And in cocktails, as in cuisine, the level of
innovation is quite high despite the absence of any legal barrier to appropriation. As dis-
cussed earlier, social norms among producers often play an important role in constraining
appropriation, or mitigating its putative negative effects. The emphasis on performance
over product attacks the problem of appropriation in a different way: reducing the ability
of the consumer to find a true copy, rather than (as with social norms) constraining
producers’ willingness to copy one another.
The product-performance continuum has been highlighted by work focused on the
negative space of IP, but the insights that result are not limited to it. As just noted, the
13
Websites, such as pornhub.com, redtube.com, and xvideos.com, that offer clips of porno-
graphic content in a format similar to the way non-pornographic content is offered by YouTube.

music industry demonstrates the same tendencies. It is undeniable that music sales are
down more than 50 percent from their peak in 1999 of approximately $14.5 billion—in
fact, adjusted for inflation, almost 67 percent lower. It is also clear that, despite the appeal
of legitimate streaming services such as Spotify and Apple Music, many consumers
continue to pirate (or share) music. And yet recorded music continues to thrive. Online
music piracy has not reduced the quantity of music produced, or indeed its quality, as
a series of studies by Joel Waldfogel (2011, 2012) demonstrates. And, importantly, even
as the recorded music business has withered, revenues from live music have boomed—in
2014, North American revenues from concert ticket sales grew to approximately $6.3 bil-
lion, rising from less than $1.5 billion in 1999—more than a threefold increase, adjusted
for inflation.14 The music industry is re-configuring to emphasize performance and
experience over product—a shift in the industry’s output that is provoked by piracy, but
that does not appear to have blunted incentives to create new music. Indeed, recorded
music is increasingly seen as advertising for live performance, rather than the other way
around.15 This inversion of the traditional relationship renders copying far less harmful,
and, arguably, even beneficial.
The centrality of performance and experience helps explain the coexistence of some
otherwise-contradictory trends in a number of industries. Consider the willingness of
customers to pay ever higher prices for movie tickets in some theaters, even as streaming
video in the comfort of one’s home grows ever more common. Why pay to go out to a
movie theater when you can watch the exact same film on your widescreen high-definition
television, thanks to one of the many torrent websites that feature illegal content? One
answer is that the experience is quite different, and many smart theater owners have been
rapidly moving to accentuate that difference as dramatically as they can.
D. Creativity-Enhancing Copying
Sometimes creative incentives co-exist relatively easily with copying, because copying
sets trends that accelerate consumption of creative goods and, in turn, their production.
The fashion industry, for example, is rife with copies. They are part of the ecosystem in
fashion. Take Forever 21, a multibillion-dollar retailer that is a major presence in the US
and now also in Europe. Forever 21’s entire business model is based on appropriating
others’ designs and selling them, perhaps slightly tweaked, for far less.
In the US, this practice is entirely legal; fashion is a paradigmatic negative space
industry. Knockoffs are ubiquitous in the industry, and expected if not demanded by
consumers. What is striking, however, is how even rampant copying fails to drive down
the level of creativity and creative production in the apparel industry. Fashion seems to
dramatically violate the fundamental premise of IP rights.
In earlier work, we argued that this result was not due to the role of social norms,
14
http://www.statista.com/statistics/306065/concert-ticket-sales-revenue-in-north-america/
(accessed March 18, 2019).
15
E.g., a 2015 review of the EP ‘Big Grams’ in Rolling Stone argued that even star guest musi-
cians could not make it ‘feel like anything more than an attempt at landing a better festival slot.’
David Turner, ‘Review of Big Grams,’ September 25, 2015, accessed March 18, 2019 at http://www.
rollingstone.com/music/albumreviews/big-grams-big-grams-20150925.

first-mover advantages, or other factors. Rather, the freedom to copy not only did not
destroy the creative impulse; it actually incentivized and accelerated it (Raustiala and
Sprigman, 2006). Fashion is a status-laden good, whose value, for many, is dependent on
its scarcity and novelty. When a design is fresh and new it is desirable. But as it spreads into
the marketplace, and becomes more common, the early adopters no longer value it and
seek to move on to newer designs. By allowing designs to be widely copied, and thereby
stimulating demand for new designs, the regime of free appropriation that exists in the
apparel world actually spurs innovation in design.
Copying is thus a basic element of the industry’s trend-driven business model. This
is implicit in the cycle just described; designs are adopted by others once there is some
(often very early) evidence of their market appeal. The result, if copying is widespread
enough, is a trend, and trends are the centerpiece of the fashion world. Copying first
helps to set or identify trends, and then to anchor consumers’ expectations about
what is in style at a given moment. This benefits the fashion industry by lowering
consumers’ information costs about what is and is not currently in style and easing
the decision about what to wear—which, in turn, encourages more apparel purchases
by lowering the risk of purchase. And as copying spreads still further, the regime of
free appropriation helps to kill the trend that it birthed. As a design becomes very
widely copied, its cachet typically falls. Copying is, in short, the engine that drives the
fashion cycle faster.
Are there other creative fields that sustain incentives to innovate based on trends?
Certainly, trends feature in a number of industries. For example, trends have been identi-
fied in the design of new typefaces (Raustiala and Sprigman, 2012). Likewise, there are
trends in the music industry, although it is unclear how much they are connected to either
consumers’ consumption decisions or incentives to innovate. As in the case of first-mover
advantage, more work remains to be done for the negative space literature to offer a more
complete understanding of when trend-driven consumption might be relevant to extra-IP
incentives to create.
E. Market Power Unrelated to IP
One hypothesis that has received only limited attention—but we believe should receive
more—is that market power unrelated to IP might help to create and maintain innova-
tion incentives in the absence of IP protections. What we mean here is market power in
a conventional sense, that is, that may arise in a market that features weak competition
paired with barriers to entry. One plausible example is the financial services industry,
which has been the source of many innovations despite the fact that for much of its
history IP protections were either difficult or impossible to obtain or enforce (Raustiala
and Sprigman, 2012).
The financial service industry’s creative output has included thousands of varieties of
derivatives, bonds, credit and currency swaps, collateralized debt obligations, the Black-
Scholes option pricing formula, the formation of index mutual funds, the use of high-
yield or ‘junk’ bonds as means of financing mergers and acquisitions, and much more.
And for much of the industry’s history the most plausible form of legal protection, patent,
was unavailable. Nor could innovators rely on trade secret law for financial innovations
that related to publicly traded securities. Because virtually all the details of a new security

become public once the offering is filed with the Securities and Exchange Commission,
secrecy is typically impossible.16
In 1998, however, State Street Bank and Trust Co. v. Signature Financial Group Inc.,
149 F.3d 1368 (Fed. Cir. 1998) established for the first time that novel methods of doing
business were patentable. State Street had some important effects for the industry. But
whether it transformed the level or type of innovation is unclear. For example, the National
Science Foundation, which tracks research and development (R&D) spending, measured
no significant increase in financial industry R&D investment in the wake of State Street
(Raustiala and Sprigman, 2012). A survey of data from the US Bureau of Labor Statistics
also revealed no trend in the financial services industry of hiring more R&D workers, as
we might expect if the availability of those patents was making a significant difference in
resources devoted to innovation (Hunt, 2010, pp. 322–52).
So how do we explain intellectual production without intellectual property in the
financial services industry? Previous research posits a number of interacting reasons.
Financial firms may innovate to satisfy the unmet needs of particular customers; to lower
transaction costs; to avoid taxes and regulation; to take advantage of rating agencies’ rules
for assessing the quality of debt; and to take advantage of opportunities offered by new
technologies (Tufano, 2002). And for many of these kinds of innovations, patent protec-
tion would be counterproductive, because sharing with rivals is helpful or even necessary
to grow markets to the size at which they become efficient and lucrative.
To understand this point, consider the market for a new type of investment security.
In most cases, new securities are likely to be most lucrative if they trade in markets big
enough to become standardized and deeply capitalized. In practice, this typically requires
a number of firms to enter the market. Patents, however, can act as a barrier to entry. If
the innovator patents the new security, potential market participants may hesitate to enter
the market for fear that the patent might be used against them. This fear might persist
even if the innovator is willing to license the patent to its rivals.17
The bottom line is that IP rights do not seem to matter much to success in this market
or to incentives to innovate. Financial firms that introduce a new and unpatented type of
security typically retain a dominant market share for several years, even though rival firms
rapidly copy the innovation (Tufano, 2002).
Why? One hypothesis looks to in-house expertise developed in the process of innova-
tion. Like a football team that has trained and recruited to run a particular offense—and
is thereby better able to implement that offense than are its rivals—the innovating firm is
more likely (at least until rival firms catch up or hire away key personnel) to have special-
ized in-house expertise in using the security that will advantage it over rivals.
But perhaps a deeper explanation relates to the market power of large banks. Financial
services are dominated by a small number of very large firms. These firms control large
shares in particular lines of business. Investment banking is also driven by relationships,
and many clients have long-term ties to their bankers that span a variety of product areas.
16
Trade secrecy is more viable for other types of financial investments, such as pricing models,
but even in these cases, financial firms often are better off sharing information than keeping it
secret.
17
Especially if those rivals worry that, as a consequence of the license fees, they will face
higher costs in marketing the security compared to the innovator.

As a result, even if innovations can freely be copied, a large bank can capture a significant
share of the return on its investment in innovation simply by virtue of its dominance
of the particular business at issue and its enduring client relationships (Merges, 2003,
pp. 1–14). Consistent with this hypothesis, the leading innovators in financial services
have been the biggest firms. When a small bank innovates, it has a strong incentive to
partner with a large bank—the larger institution is able to capture a greater share of the
returns from the innovation, which the smaller institution will share. In some instances,
banks will be incentivized to sell innovations to the institution that has a leading role in
the particular line of business addressed by the innovation (Battacharyya and Nanda,
2000, pp. 1101–27).
Financial innovation is a complex topic. But the bottom line is relatively simple. Much
of the innovation that we see in the financial services industry has been led by firms
responding to market incentives, rather than the incentives created by IP rules. As these
innovations are introduced, they quickly spread. The prospect that rivals will copy the
inventions does not destroy the incentive to create them in the first place, and indeed
in some situations copying enhances the value of these innovations by creating a larger
market for them. Investment firms locked in a competitive market for clients innovate to
serve clients better, and their rivals imitate those innovations to remain competitive.
IV. IMPLICATIONS AND NEW DIRECTIONS
A. The Negative Space Literature and the Law’s Unintended Consequences
Taken together, the studies we have described suggest that the incentives created by IP
rights are not nearly as central to innovation as conventional wisdom suggests. They
may be sufficient, but they are clearly not necessary. To be sure, there are limitations to
extrapolating larger lessons from forays into IP’s negative space. As noted, many of the
industries studied thus far feature relatively low-cost investments and often fast innova-
tion cycles. Low-IP equilibria also can be unstable, and lead to a variety of non-optimal
outcomes, including inefficient non-IP strategies for maintaining competitive advantage
and exploitation of knowledge workers (Dreyfuss, 2010). These are all serious critiques,
and they point out the difficulty, at least given current research, of drawing strong
normative conclusions from the negative space literature that would lead to concrete
recommendations for policy. The negative space literature has nonetheless succeeded,
we believe, in displacing what previously was far too automatic an association between
innovation incentives and intellectual property rules.
The negative space literature has begun to explore another facet of IP’s effect on real-
world creativity that could turn out to be very important. The principal justification for
IP protection, at least in the US, has been closely linked to the quantity of innovation.
That is, the dominant question has been: how much creative work will be produced by a
particular set of monopoly rights? In a world without restraints on copying, IP theory
tells us to expect too little innovation—that is, an amount less than the social optimum.
Yet much of the negative space literature calls into question whether IP serves as the
exclusive or even primary determinant of the quantity of innovation. Some of the studies
also point to a different and equally interesting effect—an effect on what kind of creative

work is produced. In other words, this research suggests that legal rules can affect the
quality of innovations we see, not merely the quantity.18 We might get certain kinds of
creative work versus other kinds depending on whether we have strong levels of IP protec-
tion, weak levels, or none at all.
To see this, take the example of stand-up comedy. At a general level, comedians’ norms
about appropriation and attribution appear to achieve an important end. They restrain
copying enough that incentives to invest in the creation of new comedic material remain
robust. Yet the history of stand-up comedy does not suggest that there was a dearth of
comedic production before the norm system took hold in comedy. Lots of jokes were
created during the era of the Henny Youngman-type joke-slinger. After the norms system
takes hold, what we see is not more material; what we see is different material. Comics
begin to develop material that is more personalized, that is more narrative in nature, that is
more individually tailored to individual comedians, and that arguably represents a higher
level of creative investment.
The idea that IP protection may impact the quality as well as the quantity of innovation
should not be surprising, but it has profound implications. Seen in this light, IP policy,
especially as it relates to copyright, implicates difficult normative questions about what
kind of culture we want. In the old stand-up community, with its comedic commons
that could be freely raided by any participant, the kinds of jokes that were heard were
perhaps not amazingly original. But they were easy to remember and recount to others. If
you think of comedy as serving as a social lubricant and providing a shared vocabulary,
you probably like the one-liners that the post-vaudeville era produced. If, on the other
hand, you think of comedy as a platform for political statements or individual artistic
exploration—that is, the kind of deeply original, individualized, diverse, richly narrative,
and persona-driven comedy we get today—you probably like the informal regime of
norms that accompanies this sort of comedic work and that helped to cement it into place.
The recent history of the music business likewise raises the question whether the
primary real-world effect of IP rights is the quantity of output or quality of output.
Over the past 15 years the music industry has involuntarily slipped into low-IP status
due to the diffusion of digital technologies. Yet the overall output of recorded music does
not seem to have changed. Indeed, if anything, the total musical output, and certainly
actual consumer access to music, is higher than ever. As the famed musician David Byrne
recently wrote, ‘more of it is being found, made, distributed, and listened to than ever
before’ (Byrne, 2015).
Of course, even as more music is being made and consumed, consumers are paying
less for it. They are paying less in part because the music they pirate is free. But probably
a bigger part of why consumers are paying less is that the market has shifted, massively,
from a la carte purchasing to all-you-can-eat renting. The latter is what one purchases
with a subscription to a streaming service such as Spotify. Streaming services are a great
anti-piracy tool for the same reason Netflix is—once you purchase a subscription, there’s
a lot of content available for zero marginal cost, so why pirate? And, because trying out
new music is free, streaming services are a friendlier environment for new music than are
traditional record stores or even iTunes. And as all this good news (for consumers) has
18
We mean quality in a non-normative way; i.e., not high or low quality but type or kind.

occurred, the industry has shifted its product mix from something that is easily pirated—
recordings—to something that is not: live concerts. The latter is growing, faster than it
probably otherwise would if the former were not comparatively vulnerable to piracy.
The music industry’s shift to a more performance-based business model has been of
course highly disruptive to the traditional market structure of the music industry. There
are winners (the firms that control live performance, such as Live Nation) and losers
(the traditional major record companies). And there is the appearance of crisis, not least
because the losers have a big megaphone, and, from their narrow perspective, the new
reality is a hostile one.
Yet from a broader social perspective, music is thriving despite moving in the direction
of a low-IP industry. We can argue about whether the industry’s renewed focus on live
performances is a good thing—especially if it comes at the expense of investment in
great recordings. Or one might speculate that the rising importance of the live show to
musicians’ incomes might subtly shift the sort of recorded music we get—and shift it
toward music that is more impactful when performed live. But there is little evidence that
recorded music and its live counterpart are opposed in some sort of zero-sum game. Great
recordings fill the seats for live performances, and live shows make money. That is the new
music industry business model.
B. Unexplored Negative Space—Industrial Design
We will end by mentioning a very significant area of potentially fruitful investigation that
the negative space literature has of yet not explored. This is not the only such field, but
we believe it is particularly ripe for new research. Industrial design is a very broad area of
creativity encompassing the design of articles—kitchen appliances, hair dryers, bicycles,
machine tools, and many more product categories—that are not ‘art’ in the pure sense
but are rather made for practical use. Manufacturers often direct considerable effort into
making industrial design articles attractive. And yet the aesthetic content of industrial
design is protected only peripherally by US IP law. The application of US copyright to
industrial design is limited by the useful articles doctrine, which denies protection to
most articles which contain aesthetic content that cannot be separated from the design’s
utility. US design patent law protects only ‘novel’ designs. And US trademark law protects
only designs which can be shown to have obtained ‘secondary meaning’—that is, that
are understood by consumers to designate the source of products, rather than merely
functioning as an attractive element of product appearance (Wal-Mart Stores, Inc. v.
Samara Bros., Inc., 529 U.S. 205 (2000)).
In contrast to the US, Europe has a legal regime of encompassing protection for many
industrial designs, including sui generis protections for both registered and unregistered
designs. Moreover, in some European countries copyright protection is far broader in its
application to industrial design than it is in the US
So with respect to industrial design, we have something of a natural experiment.
Encompassing protection in the EU; peripheral protection in the US. Does the difference
in legal regimes lead to different levels (or types) of innovation across the broad diversity
of industrial design fields? Do we see a higher level of design creativity in Europe versus
in the United States? Do companies that market designs in both jurisdictions understand
the incentives differently in each? Do they behave differently? And what about sequential

innovation in these fields? Do second-comers behave differently in Europe versus the

United States—that is, do we see more activity in the US dedicated to improving existing
designs that the peripheral IP system leaves unprotected?
Of course, fashion is a form of industrial design, and, as noted above, faces very
difference legal regimes in the EU (protective) versus the US (permissive of copying).
We saw no evidence that the EU-based fashion industry is more innovative than its US
counterpart—or indeed, that fashion firms behave differently in the two jurisdictions.
Some of the most productive fashion copyists—fast fashion firms like H&M, Zara,
and Topshop—are based in Europe. And indeed we see in Europe and the US broadly
the same form of copying combined with tweaking that leads to the creation of fashion
trends. The industry’s basic innovation practice appears to be very similar on both sides
of the Atlantic, despite the very different nominal IP rules. In the US, IP is not relevant.
In the EU, it is mostly ignored.
Does the same observation hold for the broad range of other sorts of industrial design?
These industries await study. We have no predictions, except that we are sure the questions
will be worth exploring.
V. CONCLUSION
In his 2014 Nimmer Lecture, Mark Lemley (2016) provocatively referred to intellectual
property law as ‘faith-based.’ Given the continued absence of compelling evidence of its
efficacy, Lemley argued, proponents of stringent IP protection were essentially operating
on belief. And indeed, he suggested, some seemed to have almost given up on the search
for evidence itself. Whether one agrees that Lemley accurately characterized the state of IP
theory or not, there is no question that the empirical underpinnings of IP rights and rules
remain surprisingly unclear and contested. Moreover, this is true even as IP looms larger
in economic and political discourse than ever before, and as innovation and creativity
increasingly serve as the critical drivers of contemporary economies.
The recent move to explore IP’s negative space directly confronts the strong belief that
protections against copying are a necessary feature of robust innovation. By getting closer
to the ground in previously unexplored creative areas, negative space scholarship has done
several important things. It has given us valuable and often closely observed detail about
how vibrant, innovative industries and markets work and how they can often achieve
innovation—and order—without law. It has demonstrated that sustained and high levels
of intellectual production can take place in the absence of strong and effective intellectual
property rights. And in doing both of these things, it has called into question the central
tenet of IP theory, even as IP law is increasingly harmonized and strengthened around
the world via new domestic laws as well as the increasingly common and stringent IP
provisions found in major trade accords.
Negative space scholarship by no means has decided the question of whether IP
rights can be justified on more than faith—or, indeed, on the basis of the deontological
justifications that have been offered as adjuncts to, or indeed sometimes as replacements
for, empirical grounding. While the existing negative space research is important and
suggestive, it arguably remains too scant, and too concentrated in small industries that
often feature what is essentially artisanal production. Yet this line of scholarship has

rovided—and we have some grounds to believe that it will continue to provide—an unu-
p
sual and important lens on the most significant of foundational questions about IP rights.
REFERENCES
Barnett, Jonathan M. 2005. ‘Shopping for Gucci on Canal Street: Reflections on Status Consumption,
Intellectual Property, and the Incentive Thesis,’ 91 Virginia Law Review 1381–423.
Barnett, Jonathan M., Gilles Grolleau, and Sana El Harbi. 2010. ‘The Fashion Lottery: Cooperative Innovation
in Stochastic Markets,’ 39 The Journal of Legal Studies 159–200.
Battacharyya, Sugato, and Vikram Nanda. 2000. ‘Client Discretion, Switching Costs, and Financial Innovation,’
13 Review of Financial Studies 1101–27.
Benkler, Yochai. 2007. The Wealth of Networks: How Social Production Transforms Markets and Freedom. New
Haven, CT: Yale University Press.
Bernstein, Lisa. 1992. ‘Opting Out of the Legal System: Extralegal Contractual Relations in the Diamond
Industry,’ 21 Journal of Legal Studies 115–57.
Buccafusco, Christopher J. 2007. ‘On the Legal Consequences of Sauces: Should Thomas Keller’s Recipes Be
per se Copyrightable?,’ 24 Cardozo Arts & Entertainment Law Journal 1121–56.
Byrne, David. 2015. ‘Open the Music Industry’s Black Box,’ New York Times (July 31).
Darling, Kate. 2014. ‘IP without IP? A Study of the Online Adult Entertainment Industry,’ 17 Stanford
Das, Proloy K. 2000. ‘Offensive Protection: The Potential Application of Intellectual Property Law to Scripted
Sports Plays,’ 75 Indiana Law Journal 1073–101.
De Jong, Jeroen P.J., and Eric von Hippel. 2009. ‘Transfers of User Process Innovations to Process Equipment
Producers: A Study of Dutch High-Tech Firms,’ 38 Research Policy 1181–91.
Dreyfuss, Rochelle Cooper. 2010. ‘Does IP Need IP? Accommodating Intellectual Production outside the
Intellectual Property Paradigm,’ 31 Cardozo Law Review 1437–73.
Ellickson, Robert C. 1991. Order without Law: How Neighbors Settle Disputes. Cambridge, MA: Harvard
University Press.
Fagundes, David. 2012. ‘Talk Derby to Me: Intellectual Property Norms Governing Roller Derby Pseudonyms,’
Fisk, Catherine L. 2006. ‘Credit where It’s Due: The Law and Norms of Attribution,’ 95 Georgia Law Journal
49–117.
Garon, Jon M. 2010. ‘Wiki Authorship, Social Media, and the Curatorial Audience,’ 1 Harvard Journal of Sports
& Entertainment Law 95–144.
Gault, Fred, and Eric von Hippel. 2009. ‘The Prevalence of User Innovation and Free Innovation Transfers:
Implications for Statistical Indicators and Innovation Policy,’ MIT Sloan School of Management Research
Paper No. 4722-09, accessed March 18, 2019 at http://papers.ssrn.com/sol3/papers.cfm?abstract_id=1337232.
Review 1147–99.
Hemphill, C. Scott, and Jeannie Suk. 2014. ‘The Fashion Originators’ Guild of America: Self-Help at the
Edge of IP and Antitrust,’ in Rochelle Dreyfuss and Jane Ginsburg, eds., Intellectual Property at the Edge.
Cambridge: Cambridge University Press.
Hunt, Robert M. 2010. ‘Business Method Patents and U.S. Financial Services,’ 28 Contemporary Economic
Policy 322–52.
Jesien, Karolina. 2007. ‘Don’t Sweat It: Copyright Protection for Yoga . . . Are Exercise Routines Next?,’ 5
Cardozo Public Law, Policy and Ethics Journal 623–54.
Kapczynski, Amy. 2016. ‘Order without Intellectual Property Law: The Flu Network as a Case Study in Open
Science,’ 102 Cornell Law Review 1539–648.
Lastowka, Greg. 2014. ‘Minecraft as Web 2.0: Amateur Creativity & Digital Games,’ accessed March 18, 2019
at http://papers.ssrn.com/sol3/papers.cfm?abstract_id=1939241.
Lemley, Mark A. 2016. ‘Faith-Based Intellectual Property,’ 62 UCLA Law Review 1328–46.
Lerner, Josh, and Jean Tirole. 2005. ‘The Economics of Technology Sharing: Open Source and Beyond,’ 19
Litman, Jessica. 1994. ‘The Exclusive Right to Read,’ 13 Cardozo Arts and Entertainment Law Journal 29–54.
Loshin, Jacob. 2010. ‘Secrets Revealed: Protecting Magicians’ Intellectual Property without Law,’ in Christine
Corcos, ed., Law and Magic: A Collection of Essays. Durham, N.C.: Carolina Academic Press.

Macaulay, Stewart. 1963. ‘Non-Contractual Relations in Business: A Preliminary Study,’ 28 American

Sociological Review 55–67.
Magliocca, Gerard N. 2009. ‘Patenting the Curve Ball: Business Methods and Industry Norms,’ 4 BYU Law
Review 875–904.
Mencken, Jennifer. 1997. ‘A Design for the Copyright of Fashion,’ Boston College Intellectual Property &
Technology Forum (Dec. 12).
Merges, Robert P. 2003. ‘The Uninvited Guest: Patents on Wall Street,’ 88 Economic Review 1–14.
Mnookin, Robert H., and Lewis Kornhauser. 1979. ‘Bargaining in the Shadow of the Law: The Case of
Divorce,’ 88 Yale Law Journal 950–97.
Murray, Fiona, Phillipe Aghion, Mathias Dewatripont, Julian Kolev, and Scott Stern. 2009. ‘Of Mice and
Academics: Examining the Effect of Openness on Innovation,’ National Bureau of Economic Research
Working Paper No. 14819, accessed March 18, 2019 at http://www.nber.org/papers/w14819.
Oliar, Dotan, and Christopher Jon Sprigman. 2008. ‘There’s No Free Laugh (Anymore): The Emergence
of Intellectual Property Norms and the Transformation of Stand-Up Comedy,’ 94 Virginia Law Review
1787–867.
Perzanowski, Aaron. 2013. ‘Tattoos and IP Norms,’ 98 Minnesota Law Review 511–91.
Pollack, Malla. 1991. ‘Intellectual Property Protection for the Creative Chef, or How to Copyright a Cake: A
Modest Proposal,’ 12 Cardozo Law Review 1477–523.
Posner, Eric A. 2002. Law and Social Norms. Cambridge, MA: Harvard University Press.
Raustiala, Kal, and Christopher Jon Sprigman. 2006. ‘The Piracy Paradox: Innovation and Intellectual Property
in Fashion Design,’ 92 Virginia Law Review 1687–777.
Raustiala, Kal, and Christopher Jon Sprigman. 2009. ‘The Piracy Paradox Revisited,’ 61 Stanford Law Review
1201–26.
Raustiala, Kal, and Christopher Jon Sprigman. 2012. The Knockoff Economy: How Imitation Sparks Innovation.
Oxford: Oxford University Press.
Rosenblatt, Elizabeth L. 2011. ‘A Theory of IP’s Negative Space,’ 34 Columbia Journal of Law and Arts 317–65.
Rosenblatt, Elizabeth L. 2013. ‘Intellectual Property’s Negative Space: Beyond the Utilitarian,’ 40 Florida State
Sanchez-Roig, Rebeca. 1989. ‘Putting the Show Together and Taking It on the Road: Copyright, the Appropriate
Protection for Theatrical Scenic and Costume Designs,’ 40 Syracuse Law Review 1089.
Schultz, Mark F. 2006. ‘Fear and Norms and Rock & Roll: What Jambands Can Teach about Persuading People
to Comply with Copyright Law,’ 21 Berkeley Technology Law Journal 651–728.
Spoo, Robert. 2013. Without Copyrights: Piracy, Publishing, and the Public Domain. Oxford: Oxford University
Press.
Strandburg, Katherine J. 2005. ‘Curiosity-Driven Research and University Technology Transfer,’ in Gary D.
Libecap, ed., University Entrepreneurship and Technology Transfer: Process, Design, and Intellectual Property.
New York, NY: Elsevier JAI.
Tufano, Peter. 2002. ‘Financial Innovation,’ in George M. Constantinides, Milton Harris, and Rene Stulz, eds.,
The Handbook of the Economics of Finance. Amsterdam: Elsevier.
Tushnet, Rebecca. 2009. ‘Economics of Desire: Fair Use and Marketplace Assumptions,’ 51 William & Mary
von Hippel, Eric. 2005. Democratizing Innovation. Cambridge, MA: The MIT Press.
Waldfogel, Joel. 2011. ‘Bye, Bye, Miss American Pie? The Supply of New Recorded Music since Napster,’
National Bureau of Economic Research Working Paper No. w16882, accessed March 18, 2019 at http://
papers.ssrn.com/sol3/papers.cfm?abstract_id=1789463.
Waldfogel, Joel. 2012. ‘And the Bands Played On: Digital Disintermediation and the Quality of New Recorded
Music,’ accessed March 18, 2019 at http://papers.ssrn.com/sol3/papers.cfm?abstract_id=2117372.
Weinreb, Lloyd L. 1998. ‘Copyright for Functional Expression,’ 111 Harvard Law Review 1149–254.
Cases
Mitchell Bros Film Group v. Cinema Adult Theater, 604 F.2d 852 (5th Cir. 1979).
State Street Bank and Trust Co. v. Signature Financial Group Inc., 149 F.3d 1368 (Fed. Cir. 1998).
Wal-Mart Stores, Inc. v. Samara Bros., Inc., 529 U.S. 205 (2000).
Whitmill v. Warner Bros. Entertainment, 4:11-cv-752 (E.D. Mo. 2013).

12. Open innovation and ex ante licensing
Michael J. Burstein* 19
Contents
I. Introduction
II. The Centrality of Information Sharing and its Challenges
III. Collaborative Innovation and its Institutions
A. Ex Ante Licensing
B. Open Innovation
C. User Innovation
D. Peer Production
E. Knowledge Commons
IV. Conclusion—Information Exchange and IP Pluralism
References
I. INTRODUCTION
The traditional economic theory of intellectual property (IP) is based on an atomistic

model of innovation. Innovators in that model are solo practitioners—the lone genius
toiling in a garage or the vertically integrated firm with a famous research lab. But that
model describes a limited and increasingly rare innovation ecosystem. Most innovation
today takes place in a more distributed fashion, in which individuals and firms draw upon
innovations that originated elsewhere or actively seek to disseminate their own technolo-
gies for the purpose of developing and commercializing new products and processes. This
activity takes place in a variety of institutional settings. The business world recognizes
it as ‘open innovation’ (Chesbrough, 2003). Innovators as varied as medical doctors,
software engineers, and windsurfers engage in ‘user innovation’ (von Hippel, 2005).
And participants in open source software projects like Linux engage in decentralized,
distributed innovation at tremendous scale (Benkler, 2006).
This chapter surveys the roles that intellectual property plays—both positive and
negative—in innovation environments marked by the importance of sharing and collabo-
ration. In the classic model, intellectual property is thought necessary to solve the problem
of underproduction that arises from the public goods character of information (Menell
and Scotchmer, 2007; Scotchmer, 2004). In collaborative models, this problem often
(though not always) fades into the background; indeed one of the most salient critiques of
the rewards-based theory of intellectual property is that it fails to explain the persistence of
innovation in areas where IP is unavailable or foresworn (Sprigman and Raustiala, 2018).
* Professor of Law, Cardozo School of Law, Yeshiva University, New York.
330

Open innovation and ex ante licensing 331
But sharing information presents a different economic challenge—how to accomplish

technology transfer. Each of the modes of innovation described briefly above requires
parties to come together and exchange technical information and know-how.
This chapter takes an integrated approach to the problem of collaboration. It begins by
articulating the economic challenges of information sharing and technology transfer. It
then explains how various innovation environments overcome these challenges, with a focus
on the roles that intellectual property plays in these arrangements. The literature to date
suggests that whether intellectual property helps or hurts the collaboration necessary to sus-
tain open innovation is highly context- and situation-specific. The presence of significant
non-market mechanisms for information exchange, that are not tied to property or contract,
suggests that we ought to be cautious in our approach to intellectual property policy.
II. THE CENTRALITY OF INFORMATION SHARING AND ITS

CHALLENGES
The usual starting point for economic analyses of innovation is the observation that
inventors and creators often have suboptimal incentives to invest in acts of invention or
creation. Innovation is risky and expensive, but its product—information—is cheap and
easy to copy (Arrow, 1962; Nelson, 1959). That is because information is usually thought
to be both nonrivalrous and non-excludable. After information is produced, it is difficult
to keep others from accessing it and using it, and the marginal cost of copying tends
toward the (usually low) cost of reproduction. A grant of exclusive rights in information
allows those engaged in research and development potentially to recoup their investment
by excluding others from the information for a set period of time (Arrow, 1962). This
is thought, in turn, to provide an ex ante incentive to innovate (Menell and Scotchmer,
2007; Scotchmer, 2004). Patents and copyrights are usually posited as solutions to the
particular problem of underproduction of information goods in an unregulated market.
This justification is commonly referred to as the ‘reward’ theory of intellectual property.
The intellectual property incentive comes with significant social costs (Lemley, 2005).
For one, the ability of the rightsholder to price her good above marginal cost results in
deadweight loss (Scotchmer, 2004). For another, because innovation is cumulative and
information is a critical input into the downstream production of further innovation
(Scotchmer, 1991), increasing the cost of that input also yields dynamic social welfare
losses (Arrow, 1962; Merges and Nelson, 1990). This cost is especially troublesome when
considering the need for collaboration.
The reward theory of intellectual property assumes that individuals or firms engage in
inventive activity by themselves. In this account, a single decision maker faces a choice to
innovate or not, and that choice depends on her ability to recoup her investment. Even
in models of cumulative innovation, the central question is how a subsequent—but still
singular—inventor incorporates previous innovation into her own subsequent work.
These assumptions are highly contestable. It is rare that innovation takes place in a
vacuum, subject to decisions by firms or individuals acting alone (Lemley, 2011).
Some scholars have complicated the atomistic model by focusing not just on the initial
production of information, but also on its further development and commercialization.
The central insight of this line of work, derived in part from Schumpeter (1947), is that

there is a difference between invention and innovation. The former is only the first step
to bringing an intellectual product into the world. The latter is the broader process
of developing, improving, adopting, and commercializing new processes or products.
Bringing an idea to commercial fruition requires many activities beyond invention and,
importantly, many of those activities may not be within the skill or ability of the initial
inventor—things like developing a working prototype, market testing, distribution, and
follow-on improvements (Sichelman, 2010). The challenge of commercializing inventions
has been recognized by scholars and policy makers alike (Kieff, 2001; Abramowicz, 2007;
U.S. Congress Office of Technology Assessment, 1995). And meeting that challenge is
particularly important. Every innovation that fails for lack of commercialization deprives
the public of a potentially beneficial technological advance.1
Commercialization is one particularly prominent example of the benefits of technology
transfer. In order for commercialization to take place, inventors and their development
partners must communicate; they must engage in information exchange sufficient for
further technological development to take place (Burstein, 2012). This is true as well with
respect to any part of the invention or innovation process in which actors with differenti-
ated skills would find it useful to utilize those skills together. It is also true even in the
absence of differentiated skill sets. Cohen and Levinthal (1989), for example, articulate
the benefits of learning to research and development (R&D). In short, exposure to new
technological ideas builds individuals’ and firms’ capacities to innovate themselves. The
diffusion of technological innovations begets further innovation as multiple individuals
or firms utilize nonrivalrous technology in more than one way.
There are two challenges that must be overcome for information transfer to occur.
The first is a particular kind of transaction cost that arises from the public goods
nature of information. Arrow (1962) described a ‘fundamental paradox’ that arises when
two parties try to exchange information (p. 615). The buyer of information must know
enough about the subject of the trade to determine the price she is willing to pay. But
of course, once the seller has disclosed the information, then the cat is out of the bag,
and the buyer no longer has a reason to pay. This ‘disclosure paradox’ or ‘information
paradox’ acts as a kind of transaction cost, interfering with the parties’ ability to reach
a negotiated exchange. It is worth noting in this context that information transactions
are subject to the usual range of ordinary transaction costs—difficulty in finding and
negotiating with potential partners, the possibility of holdup, and so on. The disclosure
paradox compounds these well-understood difficulties.
The second challenge arises because, even if the parties can reach agreement, informa-
tion is costly and difficult to transfer. Economic models of information often treat it
as a discrete product like an apple. To be sure, some kinds of information may behave
as discrete objects—stock tips, for example. But most information does not. Instead,
information tends to be, in von Hippel’s (1994) formulation, ‘sticky.’ Information exists
in the world in artifacts (Baldwin and Clark, 2000). Some of these artifacts are tangible,
like objects that instantiate the information or books and writings that attempt to explain
1
The authors cited above view the commercialization problem as one of incentives, similar to
the underproduction problem that can plague initial invention. My focus here is on the economi-
cally distinct challenge of exchanging information for the purpose of collaboration.

it; others are intangible such as the information embedded in organizations, processes,
and, most importantly, persons. Tangible artifacts themselves sometimes reveal the
information contained therein on their faces—think, for example, of the design of a
pencil, most of which can be determined by careful observation of a pencil. Others are
less ‘self-revealing,’ such as the formula for Coca Cola (Strandburg, 2004). Regardless,
these artifacts are often imperfect. There is much knowledge that cannot be embodied
in something tangible but is instead tacit—it is the ‘know how’ that remains uncodified
and difficult to access (Polanyi, 1966; von Hippel, 1994; Arora et al., 2001). Tacit, sticky
knowledge is costly to transfer. It requires not codification, but social interaction and
learning. Winter (1987) articulates a set of characteristics of information that determine
how easy it is to transfer—whether it is: tacit or articulable; not teachable or teachable; not
articulated or articulated; not observable in use or observable in use; complex or simple;
and whether it comprises elements in a system or is independent (pp. 170–73).
The theory of knowledge production and learning described above is fundamentally
social in nature. As Benkler (2016) writes, ‘[i]nnovation is a collective, not individual
process. It is a process of learning, and therefore depends crucially on communication. . ..
It is therefore sticky, local, and social’ (p. 3). The social nature of innovation has been
demonstrated as an empirical matter as well. In a study of biotechnology, Powell et
al. (1996) show how innovation takes place through repeated interactions in networks
that exist independently of institutional boundaries. And the literature on economic
geography (Saxenian, 1994; Gilson, 1999) takes the cluster as its fundamental unit of
organization rather than the firm, demonstrating that movement among and interactions
between individuals and firms in the cluster are critical drivers of innovation performance.
The next section describes several paradigms of collaborative innovation. Each para-
digm addresses the transaction and information costs described above in different ways
and with different strategies. But what links them together is the need to communicate
information from one party to another.
III. COLLABORATIVE INNOVATION AND ITS INSTITUTIONS
The challenges described above are not insurmountable, of course. Much scholarly
attention in recent years has been paid to the various institutional mechanisms by which
parties come together to exchange information in productive ways. This section draws a
thread through these various literatures—on licensing, open innovation, user innovation,
peer production, and knowledge commons.
A. Ex Ante Licensing
We begin with the simplest model of collaboration: the bilateral license. In this setting, two
firms or individuals seek to enter a contract for the exchange of technological know-how.2
2
The term ‘license’ in this section is used to signify not only the licensing of positive law
intellectual property rights like patents or copyrights, but also licensing of trade secrets, know-how,
and other forms of unprotected or unprotectable knowledge.

This kind of license is often described as an ‘ex ante’ license. It is an agreement entered
into before the purchaser has obtained the technology through other means, such as
independent invention (Federal Trade Commission, 2011). Ex ante licenses are funda-
mentally about technology transfer.3 The discussion below assumes that the firms, when
they become aware of the opportunity to collaborate, want to do so.
In some bilateral circumstances, intellectual property may facilitate the completion of
an agreement, both by helping to overcome transaction costs and by lowering the cost
of information transfer. But intellectual property appears to be neither necessary nor
sufficient in every case.
Most basically, exclusive rights appear to solve the disclosure paradox described
above (Arrow, 1962; Kitch, 1977; Merges, 2005). The disclosure paradox arises because
information is nonexcludable—once the seller reveals the information to the buyer, the
buyer is in possession of the subject of the trade regardless of whether she has paid for
it. Exclusive legal rights like patents or copyrights make the information excludable; the
buyer might know the information, but she cannot make, use, or sell it without the seller’s
consent. The seller therefore can engage in free revealing backed by sanction of law. In
this way, the intellectual property becomes the subject of the trade (Gallini and Winter,
1985). The parties can bargain over a legally delimited sphere of information that can be
freely disclosed.
This account of intellectual property fits neatly into the framework of transaction cost
economics pioneered by Coase (1937) and elaborated by Williamson (1985) and others.
In that framework, the question whether production happens in firms or markets depends
on the transaction costs of market exchange. When those transaction costs exceed the
costs associated with managerial hierarchy, we expect to see production moved inside
the boundary of a firm. The classic example is the threat of opportunism. If this danger
cannot be mitigated through contractual mechanisms, which is often the case since
contracts cannot be theoretically complete, the parties must instead vertically integrate.
But there is a third option—assigning a residual property right in the subject of the
transaction makes up for contractual incompleteness. This explanation of the make or
buy decision maps neatly onto the dynamics described above (Burk and McDonnell, 2007;
Heald, 2005). The disclosure paradox functions as a kind of transaction cost, preventing
parties from engaging in market transactions. Assigning a property right to the subject of
the trade makes it so that the parties understand the residual allocation of rights should
the contract fail.
One consequence of this reasoning is that we should be able to test it empirically.
Strong property rights ought to be correlated with the existence of relatively disaggregated
R&D supply chains. The logic is that if intellectual property makes it easier to transact
in information, then protectable innovation will take place within smaller, specialty firms
that engage in trade with respect to their innovations while unprotectable innovation will
take place only in contexts of vertical integration (Arora and Merges, 2004; Bar-Gill and
Parchomovsky, 2004; Barnett, 2011). There is some evidence that this is the case. Arora et
3
Ex ante licenses stand in contradistinction to ex post licenses. Ex post licenses, usually
expressly of intellectual property rights, require payment of rents to the rightsholder after the licen-
see has already obtained the technology through other means (Federal Trade Commission, 2011).

al. (2001) demonstrate statistically that innovative labor in many high-tech industries such
as chemicals and biotech is increasingly specialized, and Barnett (2011) offers a case study
of fabless semiconductor manufacturing that suggests that intellectual property rights
facilitate the entry of specialized research firms and vertical disintegration of that industry.
Intellectual property could also help alleviate some of the stickiness of information
that makes its exchange difficult. IP laws can perform this function in two related ways.
First, at least with respect to patents, the very process of filing a patent application
requires codification (Burk, 2008). Codifying knowledge is not without economic cost
(Bessen, 2011), so the prospect of exclusivity offered by a patent can serve as an incentive
to overcome that cost. Second, patents can form a kind of focal point around which
negotiation over more tacit or less transferable forms of knowledge can take place.
Bundling patent-excludable assets with the transfer of know-how helps both to overcome
the possibility of holdup over the non-protectable knowledge and provide the parties with
a concrete platform to begin the process of transferring tacit knowledge (Arora et al.,
2001; Gans and Stern, 2003). Because the more- and less-excludable information assets
are complementary, the ability to withdraw the excludable asset provides the parties with
sufficient assurance to contract over the less-excludable asset. Arora et al. (2001) provide
some empirical evidence of this phenomenon, drawing upon a data set of licensing agree-
ments to demonstrate that physical assets and know-how are often transferred as part of
a package. A patent can also serve as a kind of ‘beacon’ to signal available technology
and begin a relationship that might eventually lead to transfer of both tangible and tacit
knowledge (Kieff, 2006; Long, 2002).
All that said, it does not appear that intellectual property is necessary for transactions
in information goods (Burstein, 2012), nor is intellectual property always sufficient for
effective technology transfer. There is evidence from practice of a mix of strategies being
employed in different contexts.
For one thing, the disclosure paradox does not always operate in practice as the
theoretical literature described above predicts. That is because information is not always
perfectly nonexcludable, and because it is layered and multifaceted and communicates
value in complicated ways. Putting these two characteristics together yields a range of
strategies by which parties can manipulate the information that flows to counterparties in
the context of commercial transactions. Burstein (2012) describes two strategies employed
in the biotech industry: staged disclosure with increasingly formal trust and commitment
measures, and wholesale disclosure of replicable information while holding back tacit
knowledge until formal commitments are in place.
Contractual innovations also can, in the right circumstances, decrease the parties’ need
to rely on intellectual property for information transfer. Anand and Khanna (2000), for
example, draw upon a data set of licensing contracts to demonstrate significant variation
in contractual terms by industry, in a manner that roughly correlates with the strength and
relevance of intellectual property protection in that industry. This finding suggests that
contractual adaptability rather than the strength of IP rights may be a key determinant
of success in technology transfer. Indeed, in some cases, IP rights may not be necessary
at all. Gilson et al. (2009) describe ‘contracts for innovation’ found in many disaggregated
supply chains. These contracts amount to governance mechanisms that the parties erect to
build sufficient trust over time to exchange highly appropriable information in the course
of product development.

At the same time, evidence suggests that, in many contexts, intellectual property is
not a sufficient condition for information exchange either. Although the disclosure of
the invention required in a patent application is often thought to be the central quid pro
quo upon which the system relies to promote follow-on invention, in practice it tends to
underperform that function (Fromer, 2009). The legal requirements for disclosure are
easily manipulated so that clever drafters can get away with disclosing very little useful
information. And to the extent that they do disclose such information, it is unlikely to be
read or understood—for a variety of reasons, scientists in many fields (though not all),
do not routinely read patents for technical information (Ouellette, 2012).
Even where disclosure does serve a technically useful function, it often is not enough
to promote real technology transfer. In an empirical study of university technology
licensing, Lee (2012) concludes that much of technology licensing that results in actual
product development couples formal IP licenses under the Bayh-Dole Act of 1980 with
relationship-based exchange of tacit knowledge and, in some cases, organizational
integration between the university researchers and their commercialization partners.
This suggests that to be successful, technology collaborations must often be actual
collaborations—deeply social and local.
In the bilateral context, then, intellectual property is a mixed bag. It can be helpful in
facilitating collaboration, but is neither necessary nor sufficient in many cases. The next
section considers multilateral collaboration.
B. Open Innovation
The simple bilateral model of collaboration described above often is a more accurate
depiction of innovation in practice than the atomistic view of solo inventors. But it too is
incomplete in important ways. This section moves from bilateral collaborations to multi-
lateral and distributed models of innovation. Within the firm production paradigm, this
kind of collaboration is usually referred to as ‘open innovation,’ following Chesbrough
(2003, 2006).4 Outside of the firm production paradigm, collaborative innovation takes
a variety of forms that are discussed in the sections that follow.
At its core, open innovation is a business strategy. It is distinguished from ‘closed’
innovation models in which R&D is vertically integrated within the firm (Chandler,
1977). Famous twentieth-century corporate research centers like Bell Labs and Xerox
PARC are quintessential examples of the closed model. In those institutions, R&D man-
agers within the firm selected promising internal research projects for development, and
shepherded developed projects through product launch. By and large, the ideas chosen
for development came from within the lab, and those ideas that were not selected for
further development were discarded (Chesbrough, 2003). The underlying management
theory was that internal R&D created economies of scale that would lower the cost of
4
Outside the management literature, ‘open innovation’ is a term with many meanings (de
Beer, 2015). It is useful to distinguish ‘open innovation’ as business strategy, which is the use of
the term with which this section is primarily concerned, from ‘open’ intellectual property strategy.
The latter term usually refers to innovations for which ‘all information related to the innovation
is a public good—nonrivalrous and nonexcludable’ (Baldwin and von Hippel, 2011, p. 1400), as in
‘open collaborative innovation’ (p. 1400), ‘open source’ (Raymond, 1999), or ‘open science.’

innovation, and that exchanging ideas outside the firm would only increase costs and
benefit competitors.
Open innovation models, by contrast, start from the premise that information is
widely distributed in society and that a firm’s innovation strategy does not require
that all good ideas be generated from within. Instead, open innovation ‘is the use of
purposive inflows and outflows of knowledge to accelerate internal innovation, and
expand the markets for external use of innovation, respectively’ (Chesbrough, 2006).
As Chesbrough writes, ‘[o]pen innovation is a paradigm that assumes that firms can
and should use external ideas as well as internal ideas, and internal and external paths
to market, as they look to advance their technology’ (p. 1). More recently, Chesbrough
and his collaborators have expanded the definition of open innovation further to
recognize and incorporate the diversity of strategies that firms may utilize. In this
framing, open innovation is ‘a distributed innovation process based on purposively
managed knowledge flows across organizational boundaries, using pecuniary and non-
pecuniary mechanisms in line with the organization’s business model’ (Chesbrough and
Bogers, 2014, p. 17).
Much of the literature on open innovation is devoted to case studies and frameworks
that explain the various ways in which firms implement open strategies (West et al., 2014;
West and Bogers, 2014.) There are, broadly speaking, three modes of open innovation.
The first is ‘outside-in,’ in which companies open their internal innovation processes to
inputs and contributions that originate outside the firm, leveraging the distributed nature
of technological innovation and knowledge to drive their own R&D efforts. Companies
do this with a number of mechanisms, including scouting, in-licensing IP, collaborating
with universities and research institutions, funding startup companies in an industry to
get an early look into new technologies, or working with users,5 suppliers, or customers.
Increasingly popular are mechanisms like crowdsourcing (Howe, 2006) or innovation
prizes (Boudreau et al., 2011), that draw on wide and previously unknown audiences to
solve innovation problems collectively.
By contrast, inside-out open innovation happens when companies take underutilized
ideas developed in-house and pursue external means of development or commercializa-
tion. In other words, a firm allows others to make use of its inventions. As discussed in
further detail below, companies can do this either by selling or revealing, depending on
the choice of business model for appropriating the returns to the initial R&D investment.
Specific mechanisms to enable this sort of open innovation include out-licensing IP,
donating IP, spinning out technological assets into new companies, engaging in corporate
venture capital investing, and all manner of corporate external innovation strategies
including incubators, joint ventures, and the like.
There is a third open innovation strategy, referred to in the management literature as
‘coupled’ open innovation, that essentially links the outside-in and inside-out models
(Piller and West, 2014). In this model, collaborators manage mutual flows of knowledge
across multiple organization boundaries and engage in joint investment and commer-
cialization activities. As Lakhani et al. (2013) argue, managing the boundaries of the firm
5
User innovation in particular is a mode of open innovation but also a phenomenon unto
itself, described in the next section.

amidst increasingly decentralized networks of collaboration is a central challenge for

managers. Indeed, as transaction costs decline and the ease of communicating knowledge
increases, the case for the centrality of the firm as the relevant economic entity, as opposed
to the network, begins to fade (Benkler, 2016).
The role for intellectual property law in this environment is ambiguous. That is because
the management question whether to pursue a closed or an open innovation strategy
is distinct from the set of operational choices around how to do so. Open innovation
strategies—collaboration outside the firm—can be accomplished through different IP
strategies. Early work on the open innovation paradigm emphasized the importance
of appropriability over free revealing (Chesbrough, 2003). This was particularly true
with respect to inside-out strategies, which were defined more-or-less completely in
terms of out-licensing patents (West et al., 2014). The argument in favor of this posi-
tion is grounded in the classic argument for intellectual property rights, that firms need
exclusivity to recoup the cost of their investment in research, even if (or especially if) the
development of that research into a viable product takes place outside the firm (West,
2006). In that circumstance, retaining an exclusive right that gives rise to a royalty stream
enables the originating firm to participate in the upside should the licensee firm succeed
in marketing the product.
Chesbrough and Ghafele (2014) draw an explicit link between open innovation and
markets for technology. They draw from previous work that argues that strong IP rights
enable ‘intermediate’ markets for technology to form because such rights represent loci for
bargaining, on both the buy side and the sell side (Federal Trade Commission, 2011; Gans
and Stern, 2003; Arora et al., 2001). For buyers, a well-functioning market for technology
enables them to readily determine what technologies are available at what prices. For
sellers, such a market provides an easy source of liquidity for technologies that they do
not wish to develop in-house. The link between markets and open innovation is relatively
straightforward—well-functioning markets would ease the way toward open innova-
tion. Chesbrough and Ghafele (2014) therefore call for policy changes to improve the
functioning of IP markets. Burstein (2015), however, calls into question the link between
markets for technology and markets for patents. He argues that patent markets may be
welfare enhancing to the extent that they accurately represent markets for the underlying
technology, but may generate significant social welfare losses when the two diverge, which
may often be the case. While it therefore may be true that IP can facilitate market-based
open innovation strategies, it is also possible that the optimal social policy may diverge
from the private value associated with a firm-centric open innovation paradigm (West
and Lakhani, 2008).
In all events, Teece (1986) has long made the argument that IP is not the only source
of appropriability for innovating firms. Just as important in many cases are first mover
advantages or ownership of complementary assets. Firms may have a variety of mecha-
nisms for appropriating their investments in R&D even while eschewing strong exclusive
rights in the fruits of those investments. Those alternatives mean that ‘[e]ven within
a single organization’s strategy . . . the trade-off between strong intellectual property
and open collaboration is not clear cut’ (Benkler, 2017, p. 237). In particular, firms may
choose free revealing rather than intellectual property in order ‘to build reputation, gain
market share, attract third party contributions, or to grow the market’ (West et al., 2014,
p. 808). Consider, for example, Tesla’s 2014 decision to make all of its patents available

for free.6 The underlying logic of this decision was to use patents for inclusion rather than
exclusion, and to foster the development of a market for the products for which Tesla
believed it had a technological lead (Chien, 2016). There are demand-side economies of
scale in the market for electric cars because more users mean higher demand for the kinds
of services—charging stations and the like—that sustain even further use. By opening
up its patents to competitors, Tesla sought to generate a large enough market that these
network effects could begin to take hold. A similar dynamic is at work where markets are
organized around open standards (Simcoe, 2006). In those cases, interoperability benefits
all of the relevant players in an industry. It is also frequently the case that a firm may
benefit more from reciprocal openness than it will lose in contributing to that openness.
Henkel et al. (2014) demonstrate how computer component manufacturers drove stronger
collaborations by shifting from a high- to a low-appropriability regime; similarly, Laursen
and Salter (2014) draw upon survey data to find that strong appropriability is associated
with greater external search, but that high levels of protection are associated with lower
levels of collaboration.
The choice between appropriability and free revealing therefore is highly context-
specific. It will depend on the relevant industry structure and the firm’s strategic choices
within that structure (de Beer, 2015).
C. User Innovation
The previous two sections explored collaboration under the classic assumptions of
innovation economics: that innovation takes place among producers who then seek to
profit from their innovations by selling them. This classic account does not take account
of innovation by a distinct group of economic actors: users. Eric von Hippel, who has long
pioneered the study of user innovators and their communities, defines user innovators as
‘firms or individual consumers that expect to benefit from using a product or service’ that
they invent (von Hippel, 2005, p. 3).7 Unlike producer innovation, in which inventors
benefit by selling their inventions, user innovation is motivated primarily by use. Examples
of this sort of innovation are legion:
scientific instruments invented by researchers, various kinds of sporting equipment invented by

participants, open source software, medical devices, medical procedures, and off-label drug uses
invented by medical practitioners or patents and caregivers, banking and healthcare services
invented by users, financial services invented by smartphone users in developing countries, and
manufacturing processes invested by manufacturers. (Strandburg, 2018, (draft) pp. *3–4)
Von Hippel and his collaborators have amassed significant evidence about the frequency
and prevalence of user innovation (Baldwin and von Hippel, 2011, p. 1400; von Hippel,
2005, pp. 19–22 and tbl. 2.1). In short, it is pervasive. In a range of fields, many of the
most important inventions originated with users. And the percentage of users in certain
6
For more on the enforceability of such patent pledges, see Contreras (2015).
7
User innovation can and does happen within profit-making firms. For example, machinists
working for a manufacturing company frequently develop new tools for use in industrial produc-
tion. This is user rather than producer innovation because the firm will use the tools in its business
rather than sell the tools for a profit.

fields who report that they themselves engage in innovation is high—10–40 percent across
several fields in which user innovation is thought to play an important role (von Hippel,
2005). User innovation is likely to grow in importance as barriers to its exercise fall. Falling
design and communication costs make it easier for user innovators to reach one another
and to exchange information and ideas, and modular design architectures enable users to
innovate with respect to discrete use cases (Baldwin and von Hippel, 2011; Benkler, 2017).
User innovation also tends to be qualitatively different from producer innovation,
because users’ incentives and abilities are different from producers’ incentives and abili-
ties. More specifically, users typically are motivated to invent when the existing technology
does not meet their particularized needs, whereas producers are motivated to invent where
a market signal suggests significant profit opportunity. User innovators therefore tend
either to be ‘lead users,’ who are at the leading edge of market trends and therefore dis-
satisfied with the state of mainstream technology (von Hippel, 2005), or users who have
non-mainstream needs (Strandburg, 2018). These users therefore develop innovations
that would otherwise go ignored in the producer innovation paradigm.
User innovators appear to rely little on intellectual property. To the contrary, they
often engage in free revealing of their inventions. In part this is because the nature of the
innovations makes them difficult to protect either through formal intellectual property or
through secrecy (Harhoff et al., 2003). But in part this is because the cost-benefit calculus
often favors free revealing. The benefits include ‘network effects, the possibility that others
will test, comment upon, and help to improve the invention, the potential for reciprocal
access to other users’ inventions, benefits from a shared innovation process (such as
learning from others), reputational rewards and altruistic and other hedonic benefits’
(Strandburg, 2018, (draft) p. *20). As was the case with respect to open innovation,
described above, free revealing in the user innovation context may help to unlock network
effects or generate open standards that advantage the initial innovator (von Hippel, 2005).
The costs of free revealing usually are minimal, particularly in noncompetitive settings.
Of course, a user must incur the cost of codification and transmission of information and
these should not be minimized—some user innovations are ‘lost to secrecy’ because users
may not have an incentive to reveal or disseminate (Strandburg, 2018).
That disincentive is reduced in part because user innovation often takes place in discrete
communities (von Hippel, 2005; West and Lakhani, 2008). In these communities, users
create distinct governance mechanisms for the production and sharing of user innova-
tions.8 Free revealing is the norm in these communities in no small part because of an
ethos of reciprocal sharing. Strandburg (2018) collects several case studies that detail the
phenomenon.
This leads to an argument in favor of openness—in the strict sense (Baldwin and von
Hippel, 2011)—when it comes to user innovation (West et al., 2014). In the user innova-
tion mode, intellectual property does not appear to be necessary to motivate innovation in
the first instance. Because the motivation to produce depends on the user’s direct benefit
rather than appropriation through external sales, the possibility of free riding does not
depress innovative activity. Instead, the relevant policy question is whether intellectual
8
See infra Section III.E and the separate contribution of Frischmann in Chapter 21 of this
volume for a deeper discussion.

property hinders or helps the kind of collaboration that marks user innovation, particu-
larly as it takes place within the user communities described above.
Von Hippel (2005) argues that strong appropriability hinders collaborative processes.
He cites a range of strategic behaviors by patent holders that can inhibit the cumulative
innovation that marks user-based collaboration. Anticommons (Heller and Eisenberg,
1998) and patent thickets (Shapiro, 2001) that create opportunities for holdup and
significant risk of infringement litigation across wide fields, respectively, inhibit the free
exchange of information that tends to be critical for user communities to build upon
and improve upon each other’s inventions. These problems may be particularly acute for
user innovators because such innovators tend to be highly distributed with relatively few
innovations per user and little intellectual property of their own. That decreases their
ability to withstand strategic behavior from larger IP holders. Fisher (2010) highlights
the potential legal responses of producers whose goods are improved by users. Those
producers, particularly when they hold significant intellectual property, are in a position
to squelch user innovation.
The user innovation literature is broadly skeptical about the utility of intellectual prop-
erty. Indeed, it more frequently cites the dis-utility that arises from inhibiting downstream
innovation and collaboration.
D. Peer Production
Peer production is another mode of distributed, collaborative innovation. Like user

innovation, it takes place outside the context of firms and managerial hierarchies. Unlike
some forms of user innovation, it also takes place entirely outside the context of market-
mediated exchange. Peer production thus represents an organizational innovation distinct
from markets or hierarchies, focused instead on distributed innovation in networks
and marked by openness (Benkler, 2016; Baldwin and von Hippel, 2011). The study of
commons-based peer production began with free and open source software (Ghosh, 1998;
Lerner and Tirole, 2002) but was generalized most prominently by Benkler (2002, 2006),
who drew from varied additional examples including Wikipedia, open science, and others.
Peer production combines three core characteristics: ‘(a) decentralization of concep-
tion and execution of problems and solutions, (b) harnessing diverse motivations, and (c)
separation of governance and management from property and contract’ (Benkler, 2016).
The difficulty of information transfer in certain circumstances is central to Benkler’s
(2002, 2016) account of the rise of peer production. Indeed, peer production may best be
understood as an alternative institutional mechanism for achieving intellectual produc-
tion in circumstances where markets and hierarchies are suboptimal. In order to preserve
this mechanism, moreover, openness is key. Many peer production projects preserve open-
ness through the creative use of intellectual property to preserve the ‘commons’ nature
of innovation in the relevant ecosystems (Reichman and Uhlir, 2003; Schultz and Urban,
2012). Peer production thus presents a paradox with respect to IP. It is a viable—and
increasingly successful—alternative means to ensure information exchange and collabora-
tion among highly distributed actors working toward a common innovation goal. But in
some circumstances, it is built upon the very IP it purports to eschew.
Peer produced projects are highly distributed and decentralized. There is no central
authority allocating tasks to different people, or even determining what the right tasks to

perform might be. Instead, the distribution and completion of tasks happens within the
network, supplemented by governance mechanisms that fall short of market or hierarchy.
Consider Wikipedia. The world’s largest encyclopedia is staffed by a dispersed group of
thousands of volunteers who draft and edit articles using an open source software tool.
Open source software itself is similarly structured: communities of coders collaborate to
improve upon software products like Linux or Apache that supplant proprietary competi-
tor products (Benkler, 2002).
The rise of this organizational form can be explained in part with a gloss on the tradi-
tional transaction cost theory of organization described in Section II above. Baldwin and
von Hippel (2011), for example, argue that organizational form in innovation is a function
of the cost of communication and the cost of design. Where both are high, production
within firm-based hierarchies tends to dominate because there are significant capital
requirements and information aggregation and decision-making costs are internalized
through managerial behavior. On the contrary, when design costs are high and communi-
cations costs are low, peer production becomes a highly efficient means of organization.
That is because low communications costs enable easy modularization and collaboration
among disparate parties, and the high costs of design can be spread over a large popula-
tion of collaborators.
To this model, Benkler (2016) adds the insight that relevant information for innova-
tion can be highly distributed and sticky. With respect to the former characteristic,
low communication costs allow the holders of distributed information to disseminate
and link that information with ease. Peer production also allows tacit information
to be used efficiently without incurring the losses associated with attempts to codify
it. Tacit knowledge is generally not well utilized either by markets or by hierarchies,
each of which must formalize such knowledge in order to price it (for market alloca-
tion) or to specify it (for managerial allocation). But tacit, sticky knowledge resists
such formalization. The advantage of peer production for the exchange and use of
such knowledge is that formalization is not required. Instead, distributed innovation
allows the holders of tacit knowledge to self-identify and bring the full scope of their
knowledge to bear. This is especially important in innovation environments marked
by complexity and uncertainty. In those cases, distributed forms of production enable
diverse actors to ‘experiment, learn, and iterate on solutions and their refinement with-
out requiring intermediate formalizations to permit and fund the process’ (Benkler,
2016).
The result is that as the networked information economy—with its associated low cost
of transmission and minimal capital structure—has come to dominate economic activity,
commons-based peer production has become a viable alternative means of production
in many key areas of economic activity and an important mechanism for solving the
problem of collaborating for innovation.9
9
The literature on peer production is also focused on what motivates participants in distributed
innovation. Because price signals are not generally employed in these projects, participants tend to
be motivated by a variety of non-pecuniary benefits, including the value associated with using the
product, the hedonic pleasure of participating in the project, reputational gains or accumulated
human capital, and social status within the peer group of participants (Lerner and Tirole, 2002;
Benkler, 2002). Though not the subject of this review, the prevalence of these diverse motivations

As the above account suggests, information exchange is critical for peer production
to succeed. Production mediated without price or managerial signals requires that all
participants be able to access and share the relevant information and fruits of the innova-
tive process, to build upon, edit, experiment, and iterate. Yet most of the products of
peer production are protected by intellectual property as a default. Copyright arises in
software, for example, upon its fixation in a tangible medium. To ensure continued open-
ness, peer production communities have relied upon contractual mechanisms. There is a
wide variety of these open licenses (Lerner and Tirole, 2005), but the most common and
most studied is the GNU General Public License (GPL) developed by Richard Stallman
for free and open source software. This license grants users the right to copy, alter, and
distribute the software source code in its original or modified form. But it also obligates
downstream users—makers of derivative works, in copyright parlance—to grant these
same rights to users of the derivative software (Weber, 2004). In other words, it obligates
users to license their own works on the same terms. This arrangement is often called
‘copyleft,’ in contradistinction to copyright, because it uses the tools of copyright law to
ensure that software remains open as it develops.
The success of this licensing regime has led to its expansion to domains other than
software. The Creative Commons organization makes available a number of licenses
that creators can use to set the terms for use of their own work. These licenses tend
to favor open use, but allow creators to restrict use in a variety of ways, including to
noncommercial purposes. Similar efforts have been made to develop ‘open source’-style
patent licenses (Schultz and Urban, 2012). These licenses would require the original user
to take the affirmative steps necessary to patent her invention but then to license it on
open terms with a copyleft-style requirement. Although experience with these licenses has
been mixed, they have enjoyed success in some patent-based scientific fields, including
synthetic biology (Torrance, 2017).
E. Knowledge Commons10
Open innovation, user innovation, and peer production may sometimes be thought of as
instantiations of the broader phenomenon of governing the production of knowledge
through a commons, rather than through either market or government production.
‘Knowledge commons’ is ‘shorthand for the institutionalized community governance of
the sharing of and, in some cases, creation of information, science, knowledge, [or] data’
(Madison et al., 2010). These are self-governing arrangements in which information is
shared among participants for the purpose of innovation, usually without IP.
In recent years, a number of scholars have demonstrated that Ostrom’s (1990) classic
work on self-governing natural resource commons can be adapted to study the produc-
tion of information-based goods. Instead of the tragedy of the commons, the problem
to be overcome with respect to managing information resources is the underproduction
and the empirical fact that they are sufficient in many cases to sustain innovative activity calls into
doubt the classic incentive theory of intellectual property.
10
This section draws from Burstein (2016). For additional detail, see Chapter 21 by
Frischmann (2019) in this volume.

problem described above: the nonrivalrous and nonexcludable nature of information

presents a disincentive to innovate because the fruits of investment in research and
development can be taken by others freely. And just as in the natural resources context,
the conventional wisdom holds that this underproduction problem must be remedied
either through privatization—in the form of intellectual property rights—or through
government provision, like research grants.
Noting that ‘cultural production is an inherently social phenomenon, taking place over
a wide range of scales and within a complex, overlapping variety of formal and informal
institutional structures,’ Madison et al. (2010) posit that there are institutional structures
intermediate between exclusive rights and government provision that allow for collective
action to produce intellectual goods even in the face of the economic challenges described
above. Examples abound: in addition to the already-discussed examples of free and
open source software and Wikipedia, patent pools, the Associated Press, jamband fan
communities, and others can all be conceived of as knowledge commons. Each of these
communities utilizes a set of institutional rules—sometimes intertwined with formal law,
sometimes not—to manage the production and dissemination of information.
To be sure, Ostrom’s (1990) framework for research into collective action to manage
natural resource communities does not map perfectly onto innovation environments. This
is particularly so because such environments involve not only management of resources,
but the production of intellectual goods. Nevertheless, the analogy remains a good one.
Collective action problems in the natural environment may lead to overuse of resources
in the absence of a governance structure; so too in the cultural or innovative environment
may collective action problems lead to underproduction. Barriers to collective action for
the production of innovation can be lowered through governance mechanisms.
To investigate systematically the nature of those governance structures, Madison et al.
(2018) set forth a series of questions to answer about the relevant communities, actors,
and activities. Broadly speaking, they include questions about: the background environ-
ment, such as the legal context and the ‘default’ role of intellectual property in respect
of the resources to be produced; the various attributes of the community, including the
characteristics of the resource sought to be produced and managed, the profiles of the
relevant communities members and their varying roles, and the goals and objectives of
the commons and its members; the governance mechanisms, formal and informal, of
the commons, including things like institutional and technological architectures, legal
structures, and decision rules; and finally the patterns and outcomes of the commons
including the benefits and costs to various members.
These attributes combine in various arrangements to enable collaborative innovation
usually outside the mechanisms of intellectual property. Determining more specifically
how they work is, as an empirical matter, highly context-specific. The methodology that
accompanies study of knowledge commons is therefore primarily case study-based.
* * *
In each of these institutional contexts, participants in an innovation ecosystem exchange

information for the purpose of collaboration. It is important to note that none of these
mechanisms for collaboration is exclusive of the others, and hybrids often exist. User
innovations migrate into firms practicing open innovation, where they are c ommercialized

by producers. For-profit firms often engage as participants in open source software

communities for a variety of business and strategic purposes. And so on. The point of
articulating these separate phenomena is to draw attention to the variety of contexts
in which collaborative innovation takes place, and the variety of roles that intellectual
property can play in that process.
IV. CONCLUSION—INFORMATION EXCHANGE AND IP

PLURALISM
Innovation is a fundamentally social activity. It is rarely carried out by solo inventors

or firms. Focusing on intellectual property law’s incentive effects therefore neglects
half of the story. The common thread that runs through the institutional arrangements
described above is that they all enable collaboration among inventors or creators. Some
are facilitated through IP; others eschew IP entirely; and yet others fall somewhere in
between, utilizing IP in some but not all stages of the innovation process or in some but
not all industry structures.
To the extent that IP law ought to be configured to encourage collaboration in all of
its varying forms, a one-size-fits-all approach would seem to be problematic. After all,
strong IP could interfere with non-IP-based information exchange and vice versa. The
literature has evolved several suggestions or approaches to this problem. Most broadly,
Dreyfuss (2010) argues that traditional IP law ought to be ‘accommodating’ of these new
modes of intellectual production. In this view, there is no necessary reason why IP-based
collaboration cannot exist side-by-side with non-IP-based collaboration. That is certainly
correct so far as it goes. But law still can privilege one type of innovation over another.
Because it is difficult to evaluate the social utility of different innovation regimes, the key
legal recommendations in the open innovation literature(s) revolve around leveling the
playing field (von Hippel, 2005).
There are two ways in which IP law can be changed to take account of and even pro-
mote collaboration. First, IP law ought to be structured in a manner that does not unduly
inhibit non-IP-based information exchange. Second, IP law can be structured actively to
promote information exchange in those models in which it plays a significant role.
As to the first, several authors have suggested changes to the existing positive law
of IP that can smooth the path for non-IP-based collaboration. These changes usually
involve scaling back or slowing what has been a significant expansion of the coverage of
IP within the last couple of decades. For example, the expansion of patentable subject
matter to encompass processes that appear not to require IP for either initial production
or downstream development ought to be reconsidered. Other tweaks to existing patent
law could take account of users in setting the standard for obviousness or in crafting an
experimental use exemption that covers innovation in user communities (Dreyfuss, 2010;
Strandburg, 2008).
In the other direction, some scholars have explained how the IP system could better
serve the goal of facilitating information exchange. Chien (2016) focuses on patent law’s
‘diffusionary levers’—the ways in which patents can be used actively to spread informa-
tion. She suggests (a) changes to patent law’s disclosure requirements to make the patent
document itself a better vehicle for the transmission of information, (b) improvements to

the infrastructure of patent licensing to make it easier for parties that want to transact to
find each other and do so, and (c) changes to the law of waiver that make patent pledges to
the public domain and open licensing schemes easier to enforce. Yelderman (2016) focuses
on the coordination function of patents and suggests a series of legal changes that would
better enable that function.
The role of intellectual property in collaborative innovation is highly context-specific.
At the very least, therefore, high-quality empirical research is needed to continue to
identify the environments in which IP is helpful or harmful, and policy makers ought to
be cautious about and sensitive to these different contexts.
REFERENCES
Abramowicz, Michael. 2007. ‘The Danger of Underdeveloped Patent Prospects,’ 92 Cornell Law Review 1065–121.
Anand, Bharat N., and Tarun Khanna. 2000. ‘The Structure of Licensing Agreements,’ 48 Journal of Industrial
Economics 103–35.
Arora, Ashish, and Robert Merges. 2004. ‘Specialized Supply Firms, Property Rights, and Firm Boundaries,’ 13
Industrial and Corporate Change 451–75.
Arora, Ashish, Andrea Fosfuri, and Alfonso Gambardella. 2001. Markets for Technology: The Economics of
Innovation and Corporate Strategy. Cambridge, MA: MIT Press.
Arrow, Kenneth. 1962. ‘Economic Welfare and the Allocation of Resources for Invention,’ in National Bureau
of Economic Research, ed., The Rate and Direction of Inventive Activity: Economic and Social Factors.
Princeton, NJ: Princeton University Press.
Baldwin, Carliss Y., and Kim B. Clark. 2000. Design Rules. Cambridge, MA: MIT Press.
Baldwin, Carliss Y., and Eric von Hippel. 2011. ‘Modeling a Paradigm Shift: From Producer Innovation to User
and Open Collaborative Innovation,’ 22 Organization Science 1399–417.
Bar-Gill, Oren, and Gideon Parchomovsky. 2004. ‘Intellectual Property Law and the Boundaries of the Firm,’
Harvard Law School John M. Olin Center for Law, Economics & Business Discussion Paper Series, No. 480,
accessed March 21, 2019 at https://papers.ssrn.com/sol3/papers.cfm?abstract_id=559195.
Barnett, Jonathan. 2011. ‘Intellectual Property as a Law of Organization,’ 84 Southern California Law Review
785–857.
Benkler, Yochai. 2016. ‘Peer Production and Cooperation,’ in Johannes M. Bauer and Michael Latzer, eds.,
Handbook on the Economics of the Internet. Cheltenham, UK and Northampton, MA, USA: Edward Elgar
Publishing.
Benkler, Yochai. 2017. ‘Law, Innovation, and Collaboration in Networked Economy and Society,’ 13 Annual
Review of Law and Social Science 231–50.
Bessen, James. 2011. ‘From Knowledge to Ideas: The Two Faces of Innovation,’ Boston University School of Law
Working Paper Series, No. 10-35, accessed March 21, 2019 at https://papers.ssrn.com/sol3/papers.cfm?abstract_
id=1698802&rec=1&srcabs=1742380&alg=7&pos=2.
Boudreau, Kevin J., Nicola Lacetera, and Karim R. Lakhani. 2011. ‘Incentives and Problem Uncertainty in
Innovation Contests: An Empirical Analysis,’ 57 Management Science 843–63.
Burk, Dan L. 2008. ‘The Role of Patent Law in Knowledge Codification,’ 23 Berkeley Technology Law Journal
1009–34.
Burk, Dan L., and Brett H. McDonnell. 2007. ‘The Goldilocks Hypothesis: Balancing Intellectual Property
Rights at the Boundary of the Firm,’ 2007 University of Illinois Law Review 575–636.
Burstein, Michael J. 2012. ‘Exchanging Information without Intellectual Property,’ 91 Texas Law Review 227–82.
Burstein, Michael J. 2015. ‘Patent Markets: A Framework for Evaluation,’ 47 Arizona State Law Journal 507–42.
Burstein, Michael J. 2016. ‘The Entrepreneurial Commons: Reframing the Relationship between Intellectual
Property and Entrepreneurship,’ 2016 Utah Law Review 611–30.
Chandler, A.D. 1977. The Visible Hand: The Managerial Revolution in American Business. Cambridge, MA:
Chesbrough, Henry. 2003. Open Innovation: The New Imperative for Creating and Profiting from Technology.
Cambridge, MA: Harvard Business Review Press.

Chesbrough, Henry. 2006. ‘Open Innovation: A New Paradigm for Understanding Industrial Innovation,’ in
Henry Chesbrough, Wim Vanhaverbeke, and Joel West, eds., Open Innovation: Researching a New Paradigm.
Chesbrough, Henry, and Marcek Bogers. 2014. ‘Explicating Open Innovation: Clarifying an Emerging
Paradigm for Understanding Innovation,’ in Henry Chesbrough, Wim Vanhaverbeke, and Joel West, eds.,
New Frontiers in Open Innovation. Oxford: Oxford University Press.
Chesbrough, Henry, and Roya Ghafele. 2014. ‘Open Innovation and Intellectual Property: A Two-Sided Market
Perspective,’ in Henry Chesbrough, Wim Vanhaverbeke, and Joel West, eds., New Frontiers in Open Innovation.
Chien, Colleen V. 2016. ‘Opening the Patent System: Diffusionary Levers in Patent Law,’ 89 Southern California
Coase, R.H. 1937. ‘The Nature of the Firm,’ 4 Economica 386–405.
Cohen, Wesley M., and Daniel A. Levinthal. 1989. ‘Innovation and Learning: The Two Faces of R&D,’ 99 The
Economic Journal 569–96.
Contreras, Jorge L. 2015. ‘Patent Pledges,’ 47 Arizona State Law Journal 543–608.
De Beer, Jeremy. 2015. ‘Open’ Innovation Policy Frameworks: Intellectual Property, Competition, Investment &
Other Market Governance Issues. Ottawa: Prepared for Industry Canada.
Dreyfuss, Rochelle Cooper. 2010. ‘Does IP Need IP? Accommodating Intellectual Production outside the
Competition. Washington, D.C.: Federal Trade Commission.
Fisher, William W. 2010. ‘The Implications for Law of User Innovation,’ 94 Minnesota Law Review 1417–77.
Fromer, Jeanne C. 2009. ‘Patent Disclosure,’ 94 Iowa Law Review 539–600.
Gallini, Nancy T., and Ralph A. Winter. 1985. ‘Licensing in the Theory of Innovation,’ 16 RAND Journal of
Economics 237–52.
Gans, Joshua S., and Scott Stern. 2003. ‘The Product Market and the Market for “Ideas”: Commercialization
Strategies for Technology Entrepreneurs,’ 32 Research Policy 333–50.
Ghosh, R.A. 1998. ‘Cooking Pot Markets: An Economic Model for the Trade in Free Goods and Services on
the Net,’ First Monday (March 2), accessed March 21, 2019 at http://firstmonday.org/ojs/index.php/fm/article/
view/580/501.
Gilson, Ronald J. 1999. ‘The Legal Infrastructure of High Technology Industrial Districts: Silicon Valley, Route
128, and Covenants Not to Compete,’ 74 New York University Law Review 575–628.
Gilson, Ronald J., Charles F. Sabel, and Robert E. Scott. 2009. ‘Contracting for Innovation: Vertical
Disintegration and Interfirm Collaboration,’ 109 Columbia Law Review 431–502.
Harhoff, Dieter, Joachim Henkel, and Eric von Hippel. 2003. ‘Profiting from Voluntary Information Spillovers:
How Users Benefit by Freely Revealing Their Innovations,’ 32 Research Policy 1753–69.
Heald, Paul J. 2005. ‘A Transaction Costs Theory of Patent Law,’ 66 Ohio State Law Journal 473–508.
Heller, Michael A., and Rebecca S. Eisenberg. 1998. ‘Can Patent Deter Innovation? The Anticommons in
Henkel, Joachim, Simone Schöberl, and Oliver Alexy. 2014. ‘The Emergence of Openness: How and Why Firms
Adopt Selection Revealing in Open Innovation,’ 43 Research Policy 879–90.
Howe, Jeff. 2006. Crowdsourcing: Why the Power of the Crowd is Driving the Future of Business. New York, NY:
Three Rivers Press.
Review 697–754.
Kieff, F. Scott. 2006. ‘Coordination, Property, and Intellectual Property: An Unconventional Approach to
Anticompetitive Effects and Downstream Access,’ 56 Emory Law Journal 327–437.
Kitch, Edmund W. 1977. ‘The Nature and Function of the Patent System,’ 20 Journal of Law and Economics
265–90.
Lakhani, Karim R., Hila Lifshitz-Assof, and Michael L. Tushman. 2013. ‘Open Innovation and Organizational
Boundaries: Task Decomposition, Knowledge Distribution and the Locus of Innovation,’ in Anna Grandori,
ed., Handbook of Economic Organization: Integrating Economic and Organization Theory. Cheltenham, UK
Laursen, Keld, and Ammon J. Salter. 2014. ‘The Paradox of Openness: Appropriability, External Search, and
Collaboration,’ 43 Research Policy 867–78.
Lee, Peter. 2012. ‘Transcending the Tacit Dimension: Patents, Relationships, and Organizational Integration in
Technology Transfer,’ 100 California Law Review 1503–72.
Lemley, Mark A. 2011. ‘The Myth of the Sole Inventor,’ 110 Michigan Law Review 709–60.
Lerner, Josh, and Jean Tirole. 2002. ‘Some Simple Economics of Open Source,’ 50 Journal of Industrial

Lerner, Josh, and Jean Tirole. 2005. ‘The Scope of Open Source Licensing,’ 21 Journal of Law, Economics, and
Madison, Michael J., Brett M. Frischmann, and Katherine J. Strandburg. 2010. ‘Constructing Commons in the
Cultural Environment,’ 95 Cornell Law Review 657–708.
Madison, Michael J., Katherine J. Strandburg, and Brett M. Frischmann. 2018. ‘Knowledge Commons,’ in,
Peter S. Menell and David L. Schwartz, eds., Research Handbook on the Economics of Intellectual Property,
vol. II. Cheltenham, UK and Northampton, MA, USA: Edward Elgar Publishing.
Shavell, eds., Handbook of Law and Economics, vol. II. Amsterdam: Elsevier.
Merges, Robert P. 2005. ‘A Transactional View of Property Rights,’ 20 Berkeley Technology Law Journal
1477–520.
297–306.
Ostrom, Elinor. 1990. Governing the Commons: The Evolution of Institutions for Collective Action. New York,
NY: Cambridge University Press.
Ouellette, Lisa Larrimore. 2012. ‘Do Patents Disclose Useful Information?,’ 25 Harvard Journal of Law and
Piller, Frank, and Joel West. 2014. ‘Firms, Users, and Innovation: An Interactive Model of Coupled Open
Innovation’, in Henry Chesbrough, Wim Vanhaverbeke, and Joel West, eds., New Frontiers in Open Innovation.
Polanyi, Michael. 1966. The Tacit Dimension. Chicago, IL: University of Chicago Press.
Powell, Walter W., Kenneth W. Koput, and Laurel Smith-Doerr. 1996. ‘Interorganizational Collaboration
and the Locus of Innovation: Networks of Learning in Biotechnology,’ 41 Administrative Science Quarterly
116–45.
Raymond, Eric S. 1999. The Cathedral and the Bazaar: Musings on Linux and Open Source by an Accidental
Revolutionary. Beijing, China: O’Reilly.
Reichman, J.H., and Paul F. Uhlir. 2003. ‘A Contractually Reconstructed Research Commons for Scientific Data
in a Highly Protectionist Intellectual Property Environment,’ 66 Law and Contemporary Problems 315–462.
Saxenian, AnnaLee. 1994. Regional Advantage: Culture and Competition in Silicon Valley and Route 128.
Schultz, Jason, and Jennifer Urban. 2012. ‘Protecting Open Innovation: The Defensive Patent License as a New
Approach to Patent Threats, Transaction Costs, and Tactical Disarmament,’ 26 Harvard Journal of Law and
Technology 1–67.
Schumpeter, Joseph A. 1947. Capitalism, Socialism & Democracy. New York, NY: Harper & Brothers.
Scotchmer, Suzanne. 2004. Innovation and Incentives. Cambridge, MA: MIT Press.
Shapiro, Carl. 2001. ‘Navigating the Patent Thicket: Cross Licenses, Patent Pools, and Standard Setting,’ in
Adam B. Jaffe, Josh Lerner, and Scott Stern, eds., Innovation Policy and the Economy. Cambridge, MA: MIT
Press.
Simcoe, Timothy S. 2006. ‘Open Standards and Intellectual Property Rights,’ in Henry Chesbrough, Wim
Vanhaverbeke, and Joel West, eds., Open Innovation: Researching a New Paradigm. Oxford: Oxford University
Press.
Sprigman, Christopher Jon, and Kal Raustiala. 2018. ‘When Are IP Rights Necessary? Evidence from
Innovation in IP’s Negative Space,’ in Peter S. Menell and Ben Depoorter, eds., Research Handbook on the
Economics of Intellectual Property, vol. I. Cheltenham, UK and Northampton, MA, USA: Edward Elgar
Publishing.
Strandburg, Katherine J. 2004. ‘What Does the Public Get? Experimental Use and the Patent Bargain,’ 2004
Strandburg, Katherine J. 2008. ‘Users as Innovators: Implications for Patent Doctrine,’ 79 University of
Colorado Law Review 467–540.
Strandburg, Katherine J. 2018. ‘Users, Patents, and Innovation Policy,’ forthcoming, in Rochelle Dreyfuss and
Justine Pila, eds., Oxford Handbook of Intellectual Property Law, accessed March 21, 2019 at https://papers.
ssrn.com/sol3/papers.cfm?abstract_id=2828758.
Teece, David J. 1986. ‘Profiting from Technological Innovation: Implications for Integration, Collaboration,
Licensing, and Public Policy,’ 15 Research Policy 285–305.
Torrance, Andrew W. 2017. ‘Better to Give than to Receive: An Uncommon Commons in Synthetic Biology,’ in

Katherine J. Strandburg, Brett M. Frischmann, and Michael J. Madison, eds., Governing Medical Knowledge
Commons. Cambridge: Cambridge University Press.
U.S. Congress, Office of Technology Assessment. 1995. Innovation and Commercialization of Emerging
Technologies. Washington, DC: U.S. Government Printing Office.
Von Hippel, Eric. 1994. ‘“Sticky Information” and the Locus of Problem Solving: Implications for Innovation,’
Von Hippel, Eric. 2005. Democratizing Innovation. Cambridge, MA: MIT Press.
West, Joel. 2006. ‘Does Appropriability Enable or Retard Open Innovation?,’ in Henry Chesbrough, Wim
Vanhaverbeke, and Joel West, eds., Open Innovation: Researching a New Paradigm. Oxford: Oxford University
Press.
West, Joel, and Marcel Bogers. 2014. ‘Leveraging External Sources of Innovation: A Review of Research on
Open Innovation,’ 31 Journal of Product Innovation Management 814–31.
West, Joel, and Karim Lakhani. 2008. ‘Getting Clear about Communities in Open Innovation,’ 15 Industry and
Innovation 223–61.
West, Joel, Ammon Salter, Wim Vanhaverbeke, and Henry Chesbroug. 2014. ‘Open Innovation: The Next
Decade,’ 43 Research Policy 805–11.
Williamson, Oliver. 1985. The Economic Institutions of Capitalism. New York, NY: Free Press.
Winter, Sidney G. 1987. ‘Knowledge and Competence as Strategic Assets,’ in David J. Teece, ed., The Competitive
Challenge: Strategies for Industrial Innovation and Renewal. New York, NY: Harper & Row.
Yelderman, Stephen. 2016. ‘Coordination-Focused Patent Policy,’ 96 Boston University Law Review 1565–616.

13. Prize and reward alternatives to intellectual property
Michael Abramowicz* 1
Contents
I. Introduction
II. History and Seminal Scholarship
III. Administrative Design Issues
A. Timing of Payment
B. Cash Versus Other Incentives
C. Variable Versus Fixed Payments
D. Mandatory Versus Optional
E. General Versus Targeted
F. International Coordination
IV. Valuation of Inventions and Works
A. Measurement Based on Use
B. Market Mechanisms
C. Flexible Ex Post Valuation and Hybrid Mechanisms
V. Comparison with Traditional Intellectual Property Systems and Other
Alternatives
A. Effects on Innovation
B. Effects on Redundant Innovation
C. Effects on Commercialization
D. Effects on Consumers
1. Deadweight loss
2. Distribution and labor incentives
E. Transactions Costs
F. Rent-Seeking Costs
G. Other Policy Options
1. Grants
2. Governmental purchases and tax credits
VI. Conclusion
References
I. INTRODUCTION
This chapter surveys the literature on alternatives to intellectual property, focusing espe-
cially on alternatives to patent law, but with some attention as well to copyright. It does
* Professor of Law, George Washington University.
350

Prize and reward alternatives to intellectual property 351
not consider the question whether intellectual property rights are justified, but assumes
that absent some form of governmental innovation, inventions and works of authorship
might be underproduced relative to the social optimum. The chapter thus considers how
institutions besides traditional property rights compare relative to one another and to
traditional intellectual property systems. The chapter considers only briefly the possibility
that government itself might produce inventions and works of authorship or that govern-
ment might indirectly produce these by funding research. A considerable literature already
exists on governmental peer review and research administration, and so these mechanisms
are considered only insofar as they compare to the institutions that are the primary focus
of this chapter, prize and reward systems.
A prize or reward system is one in which the government gives an award to a person
or entity that has produced an invention or work deemed socially valuable, based on
an assessment of the contribution of that invention or work. The chapter will use the
word ‘prize’ to describe systems that give an award only to a single contributor or small
group of contributors, based, for example, on who has given the best contribution in a
designated area or who is first to have solved a particular problem. A ‘reward system,’ by
contrast, will be used to refer to a system in which funds are available for a large number of
contributors, to be distributed based on assessments of the quality of their contributions.
As Burstein and Murray (2015) note, much of the academic focus has been on reward
systems, but prizes have enjoyed more recent policy prominence.
As the prize-reward distinction suggests, there are many ways in which prize and reward
systems can be structured. The form of the award may vary, consisting usually of cash but
sometimes of other governmental benefits. In some proposals, the total award is a fixed
amount, but in others, the total award may vary depending on the assessed quality of the
contributions. An award system may be an optional alternative to an intellectual property
system, or could be instituted as an exclusive alternative. A system may cover a wide range
of possible inventions or works, or it may target a particular type of invention or work,
such as pharmaceuticals or songs.
Perhaps most importantly, prize and reward systems may differ in the mechanisms used
to assess the quality of the contributions. One approach is for the government to observe
market behavior in an effort to estimate the demand for a new invention or work and
the improvement in social welfare produced by its release into the public domain. Such
observation might be conducted by an administrative agency or through judicial proceed-
ings, for example in a takings suit. Some commentators have sought to identify the specific
formulas that government should use in conducting valuation, while others have argued
that the government should have substantial freedom to take into account a wide range
of considerations in making assessments of contributions. Meanwhile, other contributors
have argued that, to restrict dangers associated with governmental discretion, a market
mechanism should be used to assess contributions.
The effectiveness of the valuation mechanism is critical to an assessment of how
alternatives to intellectual property compare to traditional intellectual property systems.
If the government miscalibrates awards, then these awards can lead to production of
inventions and works that add little to social welfare and may thus produce lower value
than alternative uses of governmental funds, including tax reductions. On the other hand,
if the government can accurately measure social welfare, then a prize or reward system
may contribute substantially to social value. If social benefits of an invention or work are

a substantial multiple of the cost of incentivizing its production, then the social benefits
of a prize or reward program may be positive. A full accounting, however, must account
for any reduction in social welfare attributable to the need to raise sufficient tax revenue
to accomplish the program.
In the real world, of course, neither a prize or reward system nor a system of intellectual
property rights will perfectly calibrate incentives and lead private parties to undertake all
investments in the creation of inventions and works that have a positive social return.
A critical virtue of intellectual property systems is that they minimize governmental
discretion in assessing the value of works. But the need to determine when to grant
intellectual property rights and to adjudicate the propriety of and infringement of such
rights dictates a considerable governmental role. Thus, the effectiveness of a patent office
and the courts are critical to assessing how existing intellectual property systems compare
to prize or reward systems. To the extent that intellectual property systems do constrain
administrative and judicial discretion with rules, the question becomes to what extent the
rules successfully align creator incentives with social welfare.
In principle, the ultimate comparison of prize systems with intellectual property
systems is an empirical one. This chapter will provide some historical context, but there
is little empirical data, particularly on the prize or reward system side of the ledger. And
even to the extent some experimentation with prize or reward systems has occurred, the
heterogeneity of such systems counsels restraint in attributing the virtues or vices of any
one such system to all of them. Of course, it is also important not to compare some hypo-
thetical ideal prize or reward system against a functioning intellectual property system.
After all, there are many competing proposals to improve patent and copyright law, and
even if there were broad scholarly agreement about the design of an ideal system, political
considerations may make suboptimal intellectual property inevitable—and would likely
mean the same for a prize or reward system.
This review thus does not seek to determine whether intellectual property systems
should be abandoned in favor of prize or reward systems or whether the latter should
be introduced as a complement to the former. Rather, it is to provide an overview of the
theoretical terrain, to identify the potential weaknesses of prize or reward systems and to
point out analogous imperfections in our existing intellectual property system. The chap-
ter will highlight some considerations, such as incentives to commercialize inventions or
works, transactions and administrative costs, and rent-seeking costs, which are sometimes
omitted from the intellectual property calculus, but must be taken into account in a full
comparison of intellectual property to prize or reward systems.
The chapter proceeds as follows. Section II will provide an overview of the history and
early scholarship on prize and reward systems, including nineteenth-century antecedents
and contemporary prizes and proposals. Section III considers administrative design
issues, other than valuation. This includes questions such as the choice between prizes
and rewards, whether awards should be general or targeted, and how international
coordination of prize or reward systems could work. Section IV then turns to valuation
mechanisms, considering traditional administrative and judicial mechanisms, as well as
market mechanisms. Finally, Section V compares prize or reward systems to the existing
intellectual property system.

II. HISTORY AND SEMINAL SCHOLARSHIP
Governmental prizes at one time represented a significant component of governmental

innovation policy and appeared to be a serious competitor to traditional intellectual
property systems. As documented by Burrell and Kelly (2014), Parliament gave numerous
invention awards from 1732 to 1840 based on its institutional assessment of the contribu-
tion of inventions. Parliament was also in the business of granting patent term extensions
on an ad hoc basis and thus used its discretion in both types of systems. Burrell and Kelly
argue that the shift from prizes to patents changed not so much because of a perception
that intellectual property systems were inherently superior to prize systems, but because
Parliament’s ‘role began to shift from that of a grand tribunal for the nation to something
more like that of a dedicated legislature’ (p. 885). Parliament’s failure, however, to vest
prize discretion in some other administrative institution, along with the failure of other
countries to create prize systems on the scale of England’s, suggests the existence of some
distrust of prize systems.
In the heyday of prizes, Parliament did not always reserve to itself the right to adjudi-
cate prizes, but instead often resorted to administrative determinations. This was true of
what is undoubtedly the most famous invention prize in history, the prize for a method
for precise determination of longitude at sea. The Board of Longitude gave awards for
two rival approaches to the problem, one involving a method of lunar distance calcula-
tion, and the other depending on accurate time-keeping (Howse, 1998). Navigators could
identify high noon local time, so if a timepiece informed them of the time at the port of
origin, they could extrapolate the extent of a ship’s progress around the globe. A modern
debate relevant to prize systems generally is whether the Board was excessively stingy
with John Harrison, who invented a timepiece allegedly capable of functioning on water,
perhaps because he was a carpenter rather than a member of the scientific establishment.
Sobel (1995) intimates that the Board failed to discharge its duty properly, though Siegel
(2009) argues that the Board’s reluctance to grant a full award to Harrison resulted from
Harrison’s inability to provide a method of replicating his timepieces. Siegel argues:
Even in the relatively simple case wherein the legislature sought to promote a solution to a single,
specific problem, and in which a single, discrete product embodied the sought-after innovation,
Parliament could not usefully specify the test that would determine whether the invention had
won the prize in a way that would eliminate the need for administrative discretion. (p. 61)
The tale can thus be seen as a cautionary one about governmental abuse of discretion
or as one highlighting the difficulty of constructing ex ante rules and the advantages of
flexibility inherent in prize systems.
Prizes had a less central role in England in the second half of the nineteenth century, but
prizes were still given in some sectors. Brunt et al. (2012) analyse annual competitions held
by the Royal Agricultural Society from 1839–1939, including 1,986 awards from 15,032
entries (p. 658). It is difficult to use econometrics to assess the incentive effects of prizes,
for the same reason that it is difficult to assess the incentive effects of intellectual property;
ordinarily, there is no control group. Brunt et al., however, exploit a system by which the
areas in which prizes were offered rotated from year to year over a 15-year period, and
they find that the availability of prizes in a particular year increased inventions in the
relevant art. Because the prize system was a complement to the patent system, the increase

in invention was manifested by an increase in patenting activity in the area covered by the
prize system in the time around the show. Similarly, Moser and Nicholas (2013) analyse
the effects of prizes at the Crystal Palace Exhibition in 1851. They show that technologies
that won prizes were more likely to receive patents than technologies that did not win a
prize. This may demonstrate that there were positive reputational benefits from receiving
prizes beyond the money offered or that judges were skilled at identifying patentable
inventions. Moser and Nicholas argue for the former interpretation.
Prizes also existed in other countries. Indeed, a number of countries had offered a
longitude prize before the prize that Harrison claimed. Nicholas (2013) documents
an extensive system of prizes in nineteenth-century Japan. Adopting an approach
similar to that of Brunt et al., Nicholas exploits variation in the size of prizes over
time to estimate the degree to which prizes affected innovation. He finds a significant
effect when measuring innovation based on patenting activity. Prizes were granted in
individual prefectures rather than nationally, and Nicholas argues that ‘within the host
prefecture the cost of the prize contests was high relative to the value of patent capital
they created’ (p. 596). Adjoining prefectures enjoyed significant knowledge spillovers
from these competitions. Nicholas’s historical account highlights a potential problem
with prize proposals that we will consider below: that a jurisdiction that sponsors a prize
may not be able to appropriate a high percentage of the benefits from invention. This
problem is, of course, particularly acute in the modern world as a result of international
information flows.
Intellectual property systems gradually began to dominate prize systems throughout
the world in the nineteenth century. Perhaps this occurred in part because patent systems
were constructed in a general way so that any inventor meeting the specified criteria could
benefit from them, while prizes were given sporadically and in an ad hoc manner. Some
commentators, however, urged the prizes be given in lieu of monopoly rights granted by
patents. Macfie (1883, p. vi), for example, argued:
In every patent there should be a condition that the State, from public moneys, or moneys sup-
plied by individuals, shall be entitled to demand that the value of the invention be estimated, and,
on this value being paid (with a liberal percentage added in consideration of ‘compulsory sale’),
the use of the invention should become free to all the Queen’s subjects (even in the Colonies, so
far as privileges granted there do not clash).
Macfie collects early commentary on prize systems, including a proposal by Sir John
Sinclair that recognized the international dimension of the problem and proposed an
agreement among the European powers and the United States to reward inventors, as well
as a proposal by Sir David Brewster for creating a board of ‘scientific and practical’ men
to administer a prize system (pp. 33–4)
By the twentieth century, prizes largely faded in the public consciousness as a method
for encouraging invention, but there were occasional proposals to revive them. For
example, Polanvyi (1944) adopted a theme similar to Macfie’s, highlighting the costs of
monopoly. He argued that ‘[i]n order that inventions may be used freely by all, we must
relieve inventors of the necessity of earning their rewards commercially and must grant
them instead the right to be rewarded from the public purse’ (p. 65). Polanvyi envisioned
not eliminating the patent system, but grafting onto it a system of compulsory licenses.
Meanwhile, the modern economic literature on prizes dates to Wright (1983), who

compares the costs and benefits of patents, prizes, and research grants. Wright formalizes
Macfie’s and Polanvyi’s concerns about monopolization, showing that prizes can improve
efficiency by eliminating the deadweight loss associated with pricing of patented products
and services above marginal cost. He also addresses a number of other issues that we will
explore below, such as the danger of excessive research into particular inventions.
The law and economics of prizes began to receive sustained academic attention in
literature beginning in the 1990s. There was also a concomitant increase in the use
of prizes as a mechanism for encouraging innovation. Prior to that, there were some
instances in which governments encouraged research through prizes. For example, the
United States Atomic Energy Act of 1946, Pub. L. No. 79-585, 60 Stat. 755 (codified as
amended at 42 U.S.C. §§ 2011–2297g), created a board to give rewards for military energy
innovations, but there are few other examples. In recent years, the Defense Advanced
Research Projects Agency (DARPA) has offered numerous prizes (Scotchmer, 2004, p. 2).
Other governmental prizes include the Department of Defense Wearable Power Prize, the
Department of Energy Grand Challenges, and the NASA Centennial Challenges (Stine,
2009). Congress required the National Science Foundation to create a program to fund
innovation inducement prizes (Science, State, Justice, Commerce, and Related Agencies
Appropriations Act, Pub. L. No. 109-108 (109th Cong. 2006)). Meanwhile, Senator
Bernie Sanders proposed replacing the patent system for pharmaceuticals with a reward
system (Medical Innovation Prize Fund Act, S. 1137, 112th Cong. (2011)). The lack of
any other apparent support in Congress highlights, however, that while prize systems have
received support as accessories to the patent system in discrete areas, there appears to
be little chance of instituting a comprehensive reward system, especially one that would
displace the patent system, in the near future.
There has also been increased private interest in prizes, particularly for philanthropic
purposes. For example, the Virgin Earth Challenge offers a $25 million prize for inventions
that might combat global warning by anthropogenic greenhouse gases (Adler, 2011). The
Ansari X-Prize has offered $10 million for the development of reusable manned aircraft,
and numerous other prizes exist as well (McKinsey & Co., 2009). Murray et al. (2012,
p. 1791) argue that the various private prizes ‘blend a myriad of complex goals, including
attention, education, awareness, credibility and demonstrating the viability of alterna-
tives.’ Burstein and Murray (2015) offer a detailed study of one such prize, the Progressive
Insurance Automotive X Prize, explaining how it achieves innovation goals that help to
overcome information asymmetries and significant uncertainty. Because prizes are still
relatively rare, they may generate more publicity than other means of privately funding
innovation, such as research grants. Adler (2011, p. 16) notes that this has sometimes led
to investments much larger in total than the value of the prize. This also may help explain
the relative popularity of prize systems compared to reward systems. A caution is that if
benefits are attributable to publicity, any successes attributable to prizes might not scale
if a prize or reward system were increased to near the size of the patent system. Burstein
and Murray (2015) argue that prize systems may not scale as well as the patent system or
other reward systems.

III. ADMINISTRATIVE DESIGN ISSUES
A. Timing of Payment
Both prize and reward systems generally make awards after the relevant research, inven-
tions, or work are complete, in contrast to research grants, which are generally given ex
ante based on a specification of the research to be conducted. Because awards are granted
ex post, inventors or creators must raise funds necessary to conduct the research or create
the work. Bakker (2013, p. 1811) argues that, before 1750, the time lag made prizes a
relatively ineffective method of inducing inventive activity. With capital markets, however,
prizes in theory should be capable of encouraging activity to be performed before awards
will be made. Venture capitalists or banks thus serve a role in assessing research plans
under a prize system, and a cost of the system is that they must be compensated for their
labor and the risk that they undertake. This may also make it difficult to offer exception-
ally large prizes; Kalil (2006, p. 7) speculates that a prize system might not be able to
induce sufficient private investment in a new particle accelerator for theoretical physics
research. Brennan et al. (2012, p. 4) argue that an assessment of who can best bear risk is
essential to evaluation of a prize system, stating that ‘prizes are likely to be better when
the needs can be specified, necessary or preferred technological solutions are not known,
and the procuring party needs to share the risk by guaranteeing a minimum return to the
first successful innovator.’
The payment of prizes and rewards after research is complete still leaves a number of
decisions concerning the payout approach. The payout can occur as soon as research is
complete, though Abramowicz (2003) argues that payments should be delayed to provide
better information to decisionmakers about the contribution of inventions. Especially
if the intellectual property system is retained and a prize or reward system is an option,
this can reduce the danger of an adverse selection problem in which only those inventors
with ‘lemons’ agree to forego their intellectual property rights in exchange for a prize
or reward. Love and Hubbard (2007, pp. 1539–40) suggest that payouts be made over
time, with subsequent payments changing based on observation of markets. Similarly,
Williams (2012, p. 772) urges that some payment be made based on adherence to technical
requirements, but additional payments should be based on whether an invention achieves
sufficient market acceptance.
B. Cash Versus Other Incentives
Virtually all proposals assume that the benefits to be granted by the government will
come in the form of cash. Galle (2014, p. 881) notes, however, that some cash prizes
may ‘reduce positive contributions from inframarginal producers by crowding out their
internal motivation’ and thus noncash prizes may sometimes be superior. Prizes can also
provide financial benefits without directly providing cash. Ridley et al. (2006) suggest that
a prize might be the right to obtain expedited review of another drug. This proposal was
adopted into law, with respect to specified diseases, by the Food and Drug Administration
Amendments Act of 2007, § 1102, codified at 21 U.S.C. § 360n, which created a system
of priority review vouchers. Similarly, Fisher and Syed (2012, p. 16) note the possibility
that pharmaceutical companies might be rewarded for creating a medication for poor

c onsumers by having patent life extended on some other drug. Of course, the government
would need to exercise discretion in determining how much patent life to grant. Wright
(1983) notes that the fact that the government does not generally tailor patent terms to
inventions suggests a concern that government may not exercise its discretion well, a
concern that might affect both cash and other prizes.
Another alternative to cash prizes is for the government to purchase innovative
products and services for itself or to help consumers with such purchases. Lichtman
(1997) argues that the government might be able to reduce deadweight loss more cheaply
by giving coupons for pharmaceuticals to the poorest consumers than by buying out the
patent. His model assumes linear demand and that the government knows the most each
consumer is willing to pay for a product (pp. 131, 135). If the government gives coupons
to just one-quarter of customers, in magnitudes just large enough to bring them up to
the top of the bottom quartile, then the pharmaceutical patent owner’s incentive will be
to lower the price so that it can gain sales from these customers, as well as from the next
quartile of consumers (p. 135). Of course, Lichtman recognizes that the government will
not have perfect information, but he argues that if the government can spend just 1 in
8 dollars correctly, this approach may be more effective dollar-for-dollar than a patent
buyout (p. 136). Lichtman also notes that the government may have tax data and other
data enabling the government to make fine-grained assessments of consumer valuations.
If, however, the government gives coupons to some consumers who would not need
them, then that would not merely undermine the government’s plan by leading to unneces-
sary spending, but would also reduce the pharmaceutical patent owner’s incentive to lower
prices to capture the consumers with coupons. Abramowicz (2003) offers a simulation
model varying the quality of the government’s estimates of demand. With sufficiently
good governmental information, coupons will be cheaper per dollar of deadweight loss
removed than patent buyouts. But below some point, patent buyouts are likely to be more
expensive. This simulation still assumes linear demand, however. Fisher and Syed (2012,
p. 21) argue that this assumption is unrealistic in the context of drugs in the developing
world, because the vast majority of consumers are able to pay very little in comparison
to consumers in the developed world. One could argue that their poverty gives these
consumers low valuations of the drugs, however, and that the deadweight loss from their
exclusion from the market is low. Fisher and Syed’s argument depends on the proposition
that willingness to pay is not a valid basis for assessing social welfare and that alternative
means of achieving redistribution are not viable.
C. Variable Versus Fixed Payments
Concerns about governmental exercise of discretion are also central to the decision of
whether to grant a fixed prize or a variable reward. Clancy and Moschini (2013) offer
a model showing that ideally the government would offer rewards equal to the social
innovation value. With a guaranteed prize at a certain level, the system cannot take
advantage of private researchers’ information about the expected value of inventions
that they might create; Wright (1983, p. 704) notes that the absence of such information
creates an informational advantage of patents over prizes. A guaranteed prize may also
be problematic if the winner of a contest may opt for some other form of compensation,
such as the patent or copyright system, instead of the contest. Ding and Wolfstetter (2011)

note that if a prize is set too low, only those with low contributions that would not fare
well on the market will submit their work for the prize.
Another approach, initially suggested by Abramowicz (2003), is to have a fixed ‘pot’ to
be distributed among an unknown number of contributions in proportion to assessments
of the value of the contributions. Hollis and Pogge (2008) elaborate how this might be
used in the context of pharmaceuticals. Love and Hubbard (2007, p. 1536) argue that a
benefit of this approach is that it provides a government assurance of how much it will
have to pay. A fixed pot reduces the risk, cited by Wright (1983, p. 703), that with a variable
prize, the government may lowball its evaluation in the hope of saving money. As Wright
recognizes, however, if a reward system is repeated over time, the government will have at
least some reputational incentives not to renege on its commitments. Another disadvan-
tage of both the fixed pot and the fixed single prize relative to variable prizes is that these
mechanisms do not tend to move the system toward the point where the marginal benefit
of prizes equals the marginal cost. If the government can be trusted to value inventions
based on social welfare, then the system should equilibrate at this point without further
governmental invention to increase the size of the prize pool.
D. Mandatory Versus Optional
Whether payments are fixed or variable, a critical question is whether participation in

the system is mandatory, or optional, in which case participants may choose whether to
seek traditional intellectual property rights instead. Shavell and van Ypersele (2001) note
that adverse selection is a potential drawback of an optional system. They assume that an
inventor knows the demand for an invention created, but the government does not, so the
only inventors to choose a prize will be those who expect to be overcompensated relative
to the patent system. To the extent that the patent system’s valuations are correlated with
social value, this creates a lemons problem. To combat this problem, the government can
offer a reward corresponding to the lowest possible demand for the inventor’s product,
based on the government’s calculations. Alternatively, the government might offer higher
rewards, thus inducing greater participation (and greater reduction of deadweight loss)
but at some cost of unnecessary payments.
Love and Hubbard (2007) defend a mandatory system, particularly in the pharma-
ceuticals context. A mandatory system eliminates the transactions costs associated with
patents and guarantees to researchers and entrepreneurs that they need not clear rights
to existing technology. A disadvantage of a mandatory system is that it makes it more
difficult to experiment and transition, though a mandatory system might be introduced
for some relatively small class of inventions. An additional disadvantage is that if the
government pays too little, there will be no alternative system for encouraging innovation.
With an optional system, low governmental payments may reduce social welfare, but
only by reducing the benefits associated with the prize system; the traditional intellectual
property system can continue to function.
E. General Versus Targeted
In principle, proposals that target pharmaceuticals could be broadened to cover any form
of intellectual property. Indeed, commentators often argue that intellectual property

works relatively well in the context of pharmaceuticals in contrast to other areas of patent
law, because patent scope can be relatively clearly defined, so the focus on pharmaceuticals
may seem inappropriate. Kremer and Williams (2010), however, argue that an alternative
to the intellectual property system might be tried in specific sectors before it is generalized,
and Fisher and Syed (2012) argue that there are some factors that make pharmaceuticals
a particularly appropriate area for experimentation. They note that ‘government agencies
regularly collect and assess data concerning the incidence and impact of diseases and thus
are well positioned to ascertain the welfare gains that could be reaped by developing and
distributing vaccines or treatments for each ailment’ (p. 3).
The advantage of a more general system is that it may be able to stimulate precisely
those innovations that are underfunded by the patent system. Abramowicz (2003) notes
that with a fixed pot, inventors are most likely to opt in if they believe that the inven-
tions that they will produce will have high measured social welfare relative to the private
gains that could be appropriated through a patent. This could reflect a form of adverse
selection, but only if there are reasons to believe that the social welfare calculations will
be systematically inaccurate. Fisher and Syed likewise argue that targeting a fund toward
specific diseases, rather than to pharmaceuticals in general, might be a mistake, because
a more general prize fund will tend to stimulate work on inventions that are expected to
have large impacts on welfare that cannot easily be monetized by the patent system.
F. International Coordination
International harmonization has reduced the ability of countries to free-ride on the

intellectual property system, though at the cost of greatly reduced access to inventions,
particularly pharmaceutical inventions, in the developed world. If individual countries
gave prizes or rewards to inventors or creators of works in an uncoordinated way, then
there would be incentives to free-ride. Duffy (2004a, p. 54) notes that each government
‘has incentives to behave strategically in contributing toward the reward.’ A government
might not give prizes or rewards but still reap the benefits of innovations generated from
prizes or rewards. A partial response for this is to grant prizes or rewards solely based on
contribution to the jurisdiction providing the prizes or rewards. Calandrillo (1998, p. 354)
suggests offering awards based on use and utility as observed in the jurisdiction providing
awards. If, for example, the United States alone offered prizes, innovators would still be
able to exploit their intellectual property in other countries. This would have the benefit
of reducing free-riding, but of course would mean continued difficulty in the developed
world accessing patented inventions.
If, however, countries can choose whether to offer intellectual property or rewards, then
there may be a strategic incentive to offer rewards but then to undervalue the rewards. In
principle, a system of harmonization could seek to ensure that countries (or perhaps, only
developed countries) that offer rewards offer them at a sufficiently high level and also that
countries do not discriminate against foreign inventors and creators in paying rewards.
Meanwhile, if some countries offered rewards while others offered intellectual property,
this could mean that inventors would face increased transactions costs, reducing some of
the benefits that have been achieved from international cooperation among patent offices.
These issues would be reduced if inventors and creators could opt into a prize or reward
system but still retain their intellectual property rights.

In theory, it might be possible to create a system of international coordination and

harmonization for prize and reward systems. It also might be possible that the developed
countries would agree to underwrite an international prize fund to benefit the developing
world, as suggested by Stiglitz (2006). But any such initiatives could require a long time
and a great deal of diplomatic effort. Under current law, it is not clear whether it is even
permissible for a country to abandon intellectual property for a reward system. Fisher and
Syed (2012, p. 34) argue that a mandatory system would violate the Agreement on Trade-
Related Aspects of Intellectual Property Rights (TRIPS), article 27 of which requires that
‘patents shall be available for any inventions . . . in all fields of technology.’ Article 30,
however, permits ‘limited exceptions [that] do not unreasonably conflict with a normal
exploitation of the patent and do not unreasonably prejudice the legitimate interests of
the patent owner.’ Perhaps the alternative revenue provided by a prize or reward system
would lead to a conclusion that such a system would not prejudice rights holders’ interests.
IV. VALUATION OF INVENTIONS AND WORKS
By far the most important practical problem in the design of a reward system is the
challenge of valuing contributions. A prize system avoids this problem to some extent
by setting fixed prizes in advance, though that means that the government must at least
be able to determine how much to offer for particular types of inventions. As Roin
(2014, p. 1035) notes, ‘Social value is notoriously difficult to measure objectively in most
circumstances, and measuring the social value of innovations—which are unique goods
by definition—may be particularly difficult.’ While we will explicitly compare prize and
reward systems to traditional intellectual property in Section V, the question whether a
particular prize or reward system would track social value better than traditional intel-
lectual property looms large in the calculation.
As a preliminary matter, one must have a handle on what counts as the social value
that a prize or reward system should be awarding. Several points are worth making. First,
the baseline target is the gross social surplus from invention. One should not subtract the
inventor’s direct costs or the cost of risk facing the inventor. The assumption underlying
both traditional intellectual property and a prize or reward system is that private incentives
will align with social ones if creators expect to receive the gross social surplus, because
they will invest until the marginal increase in expected gross social surplus increases the
marginal cost of investment. To be sure, what ultimately matters is the net surplus, but we
will assume that if the inventor receives an award equivalent to gross surplus, the inventor
will have incentives to optimize net surplus. Consistent with this observation, existing
intellectual property systems do not concern themselves with how much the inventor has
invested.
Second, a payment to a creator should ideally net out the value of any independent
rents that the creator receives for the creation. Suppose, for example, that by inventing a
new flying car, an inventor would receive substantial first-mover advantages, aided by the
trademark system. If a reward system is used as an alternative to a patent system, then
the ideal reward would be equal to the social benefits of the flying car minus the rents the
inventor receives naturally or with the help of trademark law. A reward system that paid
the full social value of an invention and also allowed inventors to appropriate value for it

in other ways would be overcompensating inventors, leading to investments past the point
at which the marginal social benefits justified the cost.
Third, and more subtly, the payment should also net out gross social surplus that would
have been achieved absent the invention without the incentive of the reward system but
will not be achieved as a result of the invention. This can be relevant when an inventor’s
diligence leads to a discovery shortly before it would have been made in the absence of a
patent or reward system. Because invention is cumulative, sometimes Invention A makes
it possible to invent Invention B. If Invention B would have been invented soon even if
no patent or reward incentive existed, then the only increase in social welfare that should
be attributable to Invention B should be any acceleration in invention that occurred
as a result of the reward system. The analogous point in the patent context is that the
nonobvious doctrine should depend on whether an invention was induced by the patent
system (see Graham v. John Deere Co., 383 U.S. 1, 11 (1966); Abramowicz and Duffy,
2011).
A. Measurement Based on Use
Perhaps the most obvious approach to assessing the social value of an invention is to
measure use of the product by consumers. This might be accomplished either by forcing
some consumers to pay for the product or to consider purchases by consumers of the
product once it is placed in the public domain. Guell and Fischbaum (1995, p. 225), who
propose patent buyouts for pharmaceutical products, suggest that the pharmaceutical be
sold initially in a test market. Shavell and van Ypersele (2001) suggest that the government
wait until the product is placed in the public domain and the government observes the
goods sold.
This approach runs into at least three significant difficulties. First, there is not an
identity between a patented invention and a consumer product (Abramowicz, 2003,
p. 143). (There often is an identity, however, between a copyrighted work and a consumer
product, so this approach may be easier to implement in the copyright context.) Often,
many innovations are inputs into a single product, and an innovation may be an input
into many different products (Hopenhayn et al., 2006, p. 1047). A further complication,
identified by Chari et al. (2012), is that producers may have an incentive to manipulate
market signals about the value of their inventions, for example by bribing other producers.
Chari et al. conclude that the relative strength of patent and reward systems depends on
how costly it is for producers to manipulate market signals.
Second, measurement can be difficult even when there is an identity between product
and contribution. One problem, identified by Coase (1946) and applied to the patent
reward context by Duffy (2004a), is that buyouts can cause misallocations of resources
even where a good is used. For example, government transportation may lead consumers
to live inefficiently far from a city. The fact that consumers take advantage of the transport
does not mean that they value it at the cost of the production. The relevant question is how
much consumers ‘would have demanded the good if government policy were different and
they had expected to pay the full cost of the product’ (p. 43). Duffy points out that we should
like to know the most that consumers would pay for a product, but ‘the government will
have no data to answer that question because consumers are not paying full costs and
never expected to do so’ (Duffy, 2004a, p. 43). Meanwhile, manipulation by producers

can cause distortions. Roin (2014, p. 1037 n. 174, pp. 1059–60) observes that producers
may sell goods below marginal cost to inflate apparent demand, not only thwarting value
measurement but also leading to inefficient overuse.
Third, even measurement of use may face administrative complications. For a
pharmaceutical, whose use is probably considerably easier to measure than many other
inventions, we would need to be able to determine total manufacturing by generic firms,
perhaps not an insurmountable problem but one requiring some degree of coordination.
Sometimes, once use is known, it may be possible to estimate consumer value, subject to
the objections above, but often, the government may take shortcuts, such as assuming
benefits based on clinical trials. Hollis and Pogge (2008, pp. 30–31) point out that this
approach may be flawed, because clinical trials may feature the consumers most likely to
benefit from an invention.
B. Market Mechanisms
An alternative approach, or perhaps sometimes a complementary one, is to use a market

mechanism to estimate the value of intellectual property forfeited by its owner in exchange
for a buyout price. Of course, such a system can function only as an adjunct to the exist-
ing intellectual property system. The mechanism in the literature that has received the
most attention is offered by Kremer (1998), who suggests allowing an optional system in
which an inventor can submit an invention for an auction to the public. With some small
probability, the auction is consummated; the rest of the time, the invention is placed in
the public domain (p. 1147). The virtue of this system is that the government can harness
private information. The government pays the patent owner an amount based on the bids
when the auction is not consummated.
The system, however, invites a number of objections. A patentee might make side
payments to bidders. Collusion can even occur implicitly; a company might bid often on
another company’s intellectual property in the hope of reciprocation. Kremer suggests
combating this by paying some multiple of the third-highest bid (p. 1158). The effec-
tiveness of this solution depends on the number of plausible bidders. More generally,
an auction might not be competitive, especially for exceptionally valuable intellectual
property. It may not be worthwhile to undertake the expense of researching the value of
intellectual property if there is only a small probability that the sale will be consummated.
Clancy and Moschini (2013, p. 229) note also that potential bidders must have sufficient
information to make meaningful evaluations. In addition, truly random assignment will
lead to some patents being kept in private hands. ‘The government must be willing to
allow the cure for AIDS to remain in private hands if random chance would have it’
(Duffy, 2004a, p. 48).
A more complex problem that Kremer addresses is that different patents may be
substitutes or complements. If two patents are complements, rewards may be too high if
they are auctioned separately; given the high probability each of two perfect complements
will be assigned to the public domain, the other will command nearly the full value of
both together. Meanwhile, the value of one patent will be greatly lowered if a substitute
patent will likely be placed in the public domain. Kremer responds by suggesting joint
randomization, that is, that all close complements and substitutes be auctioned simultane-
ously, with all either randomized to the public domain or to private ownership, though

not necessarily by the same owners. This may, however, not always be administratively
feasible, if different inventions are produced at different times.
An alternative market mechanism, still subject to concerns about substitutes and com-
plements but avoiding some of the other difficulties, is suggested in a footnote by Duffy
(2004a, p. 47 n. 37). Under this mechanism, the government announces an intention to
take some intellectual property owned by a publicly traded firm, but it does not announce
the precise time at which this taking will occur. At some moment, trading is suspended
and the government announces that it has taken the property, but that the takings amount
will be paid to the shareholders of record at the moment of the taking. Trading is then
resumed. Because the rights to the takings payment are no longer attached to the stock,
the stock price should fall. The total takings payment would be equal to the decline
in net capitalization. An advantage of this mechanism is that it relies on a firm’s own
shareholders to value its intellectual property. It seems likely to work effectively, however,
only with relatively high value intellectual property, and there is some danger of noise
associated with contemporaneous market events. A variation might be for the government
to sell derivatives whose value is equal to the stock price decline observed over a set time
interval and then to reimburse shareholders based on the expectation of the stock price
decline. Of course, the government must guard against stock manipulation; it might help
to randomize the exact time to be used for measurement of stock value.
C. Flexible Ex Post Valuation and Hybrid Mechanisms
A final approach is to allow for a relatively unconstrained assessment of social value to

be made ex post. This ex post evaluation could be made judicially, for example in takings
litigation as suggested by Guell and Fischbaum (1995), or through some form of adminis-
trative process. The case for flexible ex post valuation is that even if the government makes
large ex post errors, these errors will have effects on ex ante investments only to the extent
that they are systematic and can thus be anticipated (Abramowicz 2003, pp. 218–24).
Some systematic errors might exist, however. For example, Khan and Sokoloff (2009,
p. 25) demonstrated in an empirical study of prize awards that they often depended on
factors such as elite education and geographical proximity. Perhaps decisionmakers also
would favor consumer products relative to products that are inputs into other products. It
is also possible that decisionmakers might adjust based on distributional considerations,
favoring the poor. Arguably, such adjustments would help ameliorate inequality, though
whether a reward system is an efficient mechanism for redistribution depends on the
suboptimality of the existing distribution of wealth, as well as on the relative efficiency
and plausibility of this mechanism in comparison to other vehicles for redistribution,
such as taxation.
Duffy (2004a, p. 50) points out that if this worked ‘it could also be applied to other
fields of natural monopoly or even to government procurement.’ That such an approach
has not been applied even in those contexts may suggest that policymakers care about ex
post fairness rather than just ex ante efficiency, or that policymakers believe that such a
mechanism would not be effective. It is also possible that ex post valuation is less needed
with natural monopoly or government procurement, because alternative approaches may
be more viable than the patent system. Duffy also expresses a concern that flexible ex
post valuation systems might invite rent-seeking (pp. 50–51), citing the pioneer preference

program of the Federal Communications Commission. This concern is reduced if assess-

ments are to be made judicially rather than administratively.
The most significant advantage of ex post valuation relative to other valuation mecha-
nisms is that decisionmakers can in principle take into account the full range of potentially
relevant considerations. Kapczynski and Syed (2013, p. 1955) argue that a principal
justification for reward systems is that existing intellectual property systems, such as
patent law, may distort inventor incentives toward contributions where excludability is
high and away from contributions where intellectual property owners will not be able to
police violations. For example, the patent system may undervalue inventions whose value
will be realized only with a long time delay; while the patent system appropriately causes
potential inventors to discount anticipated revenues to the present, they will ignore antici-
pated use beyond the patent term. Market mechanisms also will not work as a mechanism
for encouraging subpatentable inventions, that is, small contributions that would not be
worth the social cost of a patent but are nonetheless worth encouraging.
A market mechanism that simply attempts to replicate the private value that would
be achieved by the patent system cannot overcome the limits of such a system. It is pos-
sible, however, to envision a hybrid system. For example, the government might hold a
Kremer-like auction to provide a baseline for private valuation, but then use flexible ex
post decisionmaking to determine the ultimate amount to be paid. This would allow the
decisionmakers to focus specifically on the factors that would cause deviation between
private and social value under a traditional intellectual property system. Similarly, if
the Duffy taking system were employed, the government might be allowed to deviate
up or down from the stock market valuation of a patent based on such considerations.
Of course, an auction mechanism could also be combined with estimates based on sales
(Kremer, 1998, pp. 1160–61). A flexible ex post valuation could take into account a market
valuation as well as a valuation based on sales.
V. COMPARISON WITH TRADITIONAL INTELLECTUAL

PROPERTY SYSTEMS AND OTHER ALTERNATIVES
The above sketch of the history of prize and reward systems and description of design
choices hint at the comparison between those systems and existing intellectual property
institutions. The case for each approach focuses on the flaws of the other. A patent
or reward system will seem attractive to those who look at the patent system and see
deadweight loss, inefficient races, and high transactions costs. The patent system will seem
preferable for those who believe that government would be ineffective in determining the
appropriate size of awards. John Stuart Mill (1858) stated this case most famously and
perhaps most persuasively:
[I]n general, an exclusive privilege, of temporary duration, is preferable, because it leaves nothing
to anyone’s discretion; because the reward conferred by it depends upon the invention’s being
found useful, and the greater the usefulness, the greater the reward; and because it is paid by
the very persons to whom the service is rendered, the consumers of the commodity. So decisive,
indeed, are these considerations, that if the system of patents were abandoned for that of rewards
by the State, the best shape which these could assume would be that of a small temporary tax,
imposed for the inventor’s benefit, on all persons making use of the invention.

The advocate of prizes might suggest the reverse deconstruction, recommending that the
state alter the rewards of the market to account for its imprecision in rewarding social
value, until ultimately there would be no remaining need to grant an exclusive property
right and much savings from eliminating the patent apparatus. We will not adopt either of
these polar opposite positions, but instead will consider the comparison along a number
of dimensions, with attention to the many ways we have identified that prize and reward
systems can be structured.
A. Effects on Innovation
Mill’s quotation embraces the position that patent law is well calibrated to reward social
value, because private returns will depend on success in the market. The views of consum-
ers spending their own money should be preferred to the views of bureaucrats spending
the money of others, for they may be more influenced by political considerations than
economic ones. Spulber (2014), for example, highlights the Hayekian role that prices serve
in a patent system and the difficulty that government will have in replicating the invisible
hand. Partisans of both positions generally agree that the increase in consumer surplus
created by an invention should count as an increase in social welfare, and so a system that
provides inventors with some proportion of this increase is generally effective. On the
other hand, one may argue that ‘the social worth of a good depends (at least in part) on
values that are distinct from its utility to consumers’ (Roin, 2014, p. 1029). This motivates
especially proposals regarding pharmaceuticals.
Advocates of prizes, however, emphasize the imperfection in the correlation between
social value created by an invention and the amount that can be appropriated by inventors.
Appropriability, Kapczynski and Syed (2013) argue, depends largely on excludability.
Many innovations—such as hospital checklists and the provision of information that
drugs do not work—are not easily excludable, and so these will tend to be underproduced
by the patent system (p. 1908). Meanwhile, it may be difficult to monitor infringement
with some inventions (p. 1917). Some inventions, such as vaccines, may produce high
positive externalities that the patent system does not value (Fisher and Syed, 2012, p. 4).
Others may produce a high percentage of their social value after the duration of any
exclusive right, and will thus be underprovided. Finally, relative to social value, there will
be relatively little incentive to produce discoveries and inventions that have high spillovers.
Basic research exemplifies this. Patent law rejects patents on abstract ideas in part because
of the difficulties of defining property rights on such ideas, but a prize or reward system
can easily incentivize research that cannot be easily monetized through the patent system.
Thus, a critical inquiry is to examine what proportion of useful research—not necessarily
research being conducted, but research that might be incentivized either by patents or by
prizes or rewards in a hypothetical optimal system—is not easily excludable. The higher
this proportion, the stronger the case for a prize or reward system, though the possibility
that other systems might be used to encourage such research (such as research grants, to
be considered below) must also be considered.
The above are concerns about false negatives—that the patent system does not incentiv-
ize progress on much important research. But one can also argue for a reward system
on the ground that patent systems have too many false positives. An examination of
the literature on the effectiveness of patent offices in weeding out low-quality patents is

well beyond this chapter’s scope. One who perceives that many patents provide outsized
rewards for minimal contributions may believe that a reward system may be more effective
in preventing such distortions. A significant challenge in administering a patent system
is rendering determinations of whether patents are nonobvious. Part of the problem may
be that nonobviousness in patent law is a binary determination; in a reward system, by
contrast, small rewards could be given for small degrees of nonobviousness and large
rewards for greater degrees of innovation. In principle, patent law can adjust scope based
on the degree of innovativeness of an invention, but this may not be easy to accomplish
administratively, and patent offices may not have strong incentives to push back against
applicants who have made small innovations but claim broad scope. Even if these are
problems of the patent system, however, the question becomes whether the best course is
to abandon the patent system or to work to improve the patent system from within, for
example by increasing the rigor of the nonobviousness standard, reforming patent scope,
or changing patent litigation procedure.
In principle, a reward system can consider all forms of social welfare produced by an
invention, including relatively nonexcludable forms. (A prize system, meanwhile, can also
be used to reward any sort of creation, but the tendency of prize systems to specify one
or more areas of interest in advance means that they will necessarily exclude considera-
tion of many important discoveries outside their scope.) The heart of the objection to
a reward system is that such a system may not effectively capture the benefits that the
patent system misses and moreover may misestimate the benefits that the patent system
captures. Calandrillo (1998, p. 340 n. 155) notes that, historically, prize systems have been
rejected for largely this reason. Skepticism among economists about prizes stems largely
from distrust of bureaucracies. Scherer (1970, p. 398) argues, ‘[A]ny bureaucratic council
entrusted with the job is bound to make mistakes and perpetuate inequities.’ Similarly,
Tirole (1988, p. 401) notes that ‘[i]nformation about demand is crucial for determining the
size of the award, which, in turn, influences the research incentives.’ It may be difficult to
estimate demand when a product is offered for free.
Virtually no empirical progress has been made in addressing these questions. Part of
the problem, of course, is that we do not have reward systems. In principle, however,
governmental officials or laboratory subjects might be tasked with estimating social value
for experimental purpose. But such experiments might not be useful. It would be difficult
to assess whether estimates of social value were correct, though it might be possible to
determine whether subjects could accurately estimate the private value created by certain
discoveries. Perhaps one could determine whether subjects were relatively consistent
in their evaluations. But as long as rewards are given ex post, what matters are ex ante
forecasts of evaluations, and even highly inconsistent decisionmakers might come close to
approximating social value in expected value terms. Even if one could measure expecta-
tions of what average decisionmakers would decide, it may be impossible to know whether
these expectations represent an improvement over expectations of what the market would
reward.
Market mechanisms, like those described in Section IV.B, largely avoid the debate on
whether patents or rewards are better calibrated to social welfare. Such a market mecha-
nism seeks to buy out intellectual property at its private value, so it does not entail the risk
that government officials will systematically misestimate invention value, nor does such a
mechanism offer the potential of government officials steering inventive activity toward

what is socially valuable. Thus, those mechanisms must be assessed primarily on their
effectiveness at capturing private value and on benefits such as reduction in deadweight
loss. Such a mechanism could be used to provide a baseline compensation level, with
governmental officials permitted to deviate from that level only within some range. One
might distrust the government to generate social value estimates of innovations being
placed in the public domain but believe that governmental adjustments to the baseline of
private value may improve the correlation between social value and private value.
A final point is that such correlations are not all that matters. A system for promoting
intellectual property might reward contributions in exact proportion to their social value
yet still fail if the size of the total contributions is too low or too high. An additional
consideration is thus whether total incentives to innovate will be higher with rewards
or with traditional intellectual property. Assuming private appropriation is likely to be
less than social welfare on average with either system, the system that produces greater
total rewards may be preferable. Reward system advocates highlight that rewards could
be augmented to capture social value better. Kremer (1998, p. 1411) notes that patent
buyouts may be at a multiple of private value and suggests that a markup between 2.5
and 3.33 may be socially desirable, and Shavell and van Ypersele (2001) point out that if
government knows the demand curve, government can play social surplus and achieve a
first-best outcome. Of course, what the government could do is not necessarily what the
government will do. Kalil (2006, p. 11) notes that uncertainty of whether the government
will pay reduces the effectiveness of a prize or reward. If government is thought likely
to cheat or renege on open-ended commitments, then the optimal policy may be for the
government to commit to a fixed pot (Abramowicz, 2003). This eliminates the potential
benefit of increasing total funds offered for innovation but also reduces the danger of
underproviding for innovation, and it may be useful as a transitional mechanism, perhaps
as a complement to the patent system.
B. Effects on Redundant Innovation
Both traditional systems of intellectual property and prize or reward systems may lead to
inefficient redundant research. In a patent race, multiple competitors may attempt the same
approaches, though sometimes they will attempt different strategies. Sometimes redundant
efforts can be welfare-improving, especially where an invention has very high social value,
but it is not a free lunch. Although patent races can lead to earlier invention, the anticipa-
tion of competition reduces potential participants’ estimates of the expected payout from
a race. In equilibrium, this can lead to later racing (Duffy, 2004b). Meanwhile, attempts to
invent around an invention can also lead to inefficient duplication. Redundancy can also
occur with a prize or reward system. If there is a single prize, the dynamics of racing will
be quite similar to those in a patent system. With rewards, redundancy may still occur,
whether competitors believe only the first to complete an achievement will receive a reward
or whether competitors anticipate that the reward will be shared among all participants.
Thus, patents, prizes, and rewards may all be inferior in this particular respect to a program
of centralized government research, which can allocate funds to various research programs
in a way that attempts to minimize or optimize redundancy.
As this analysis suggests, commentators generally agree that prizes and rewards do
not necessarily solve the common pool problem. Calandrillo (1998, p. 353) notes that a

‘government-run reward system will not solve the race to be the first party to generate a
given piece of information in order to obtain the reward for it,’ and Wright (1983) observes
that an inefficiently high number of firms may respond to a prize incentive. In principle,
however, operators of a reward system might seek to adjust rewards or prizes to combat
redundancy, just as rewards in principle could be adjusted for any other purpose (Roin,
2014, p. 1033). The problem is that it is not obvious exactly what adjustments should be
made.
One possibility is to give considerably lower rewards where interest in a particular
inventive path was stimulated by an exogenous shock, such as the emergence of a com-
plementary invention. If Invention A enables Invention B, then the payment to the first
to invent B need not be the full social welfare value of B over A. As noted above, social
value is better measured against a baseline of what would have been invented absent an
incentive, and an extension to this principle is that even if some incentive is necessary to
encourage innovation, sometimes that incentive may be much lower than the full social
value of the invention. Those operating the reward system might reduce the ex post award
to the level that they believe would have produced optimal ex ante entry into the race. Of
course, this introduces the risk of hindsight bias.
Another possibility is for the reward judges to provide strong incentives for very rapid
release of information. If Inventor A performs Step X but does not report it, so Inventor
B repeats Step X unnecessarily, then the reward judge might allocate the proportion of the
reward attributable to Step X between Inventor A and Inventor B. If Inventor A reports
Step X and is the first to complete Steps Y and Z necessary to produce the invention,
then the case for sharing the reward with a later inventor who replicates A’s steps is greatly
reduced. Of course, a risk for Inventor A is that Inventor B will replicate and invent first.
In this case, however, there is a strong case for giving Inventor A at least the credit for the
portion of the reward attributable to Step X and perhaps more, if there was no reason
to think that Inventor B would make a stronger contribution. A broader observation is
that reward systems can provide incentives for taking intermediate steps on the way to a
discovery, and more granular rewards can reduce redundancy. Reward systems may also
choose to reward negative results. This presents a danger that inventors will choose paths
that they believe are unlikely to be successful, but also encourages information sharing.
With this determination as with others, it is difficult to identify principles to determine
just how much credit should be given. A reward system will generally be more attractive
if such micro-adjustments successfully adjust for distortions attributable to inefficient
racing, and less attractive if the government is poorly situated to make the adjustments
and may overcompensate.
C. Effects on Commercialization
Although the goal of the patent system is generally viewed as being encouraging research
and invention, some commentators have emphasized that intellectual property may
serve an important purpose in encouraging commercialization of invention. The patent
prospect theory is widely associated with Kitch (1977, p. 277), who argues, ‘Only in the
case of a patented product is a firm able to make the expenditures necessary to bring the
advantages of the product to the attention of the customer without fear of competitive
appropriation if the product proves successful.’ Kieff (2001) extends this insight to the

literature on prizes and rewards, noting that if payments are given for inventive effort
alone, but consumer demand is highly uncertain, then second-mover advantages may
dominate first-mover advantages. Kieff emphasizes that patents serve a role in coordinat-
ing inventors and market players.
A recent literature has assessed how intellectual property rights might be used to
further the commercialization function (see Abramowicz and Duffy, 2008; Sichelman,
2010). Hrdy (2015) emphasizes that in fact states frequently give awards to encourage
commercialization, and these awards can be seen as analogous to a reward system. It is
possible to imagine channeling such awards into a quasi-judicial process. Entrepreneurs
might be rewarded for bringing products to market where consumer demand is highly
uncertain and second movers can be expected to enter the market. The question becomes
whether reward judges can effectively determine the magnitude of such rewards and
whether inaccurate rewards or insufficient attention to the commercialization function
might lead to inferior performance relative to a patent system.
D. Effects on Consumers
1. Deadweight loss

The most common justification for a prize or reward system alternative to the patent
system is that a patent or reward system can reduce or eliminate the deadweight loss
attributable to patents. The observation that monopolists will set output at a level that
is inefficiently low from a social perspective is elementary. Wright (1983) was the first
to develop this advantage of a prize or reward system in a rigorous analytical way. One
question is how important deadweight loss is relative to other systemic imperfections.
Arguably, static inefficiency is not as important as ensuring that the patent system gener-
ates as much economic growth as possible, though the dynamic benefits of intellectual
property and prizes or rewards must be discounted to present value. Another question is
how serious the problem of deadweight loss is. Leslie (2011, p. 813) notes that patentees
may reduce deadweight loss by engaging in price discrimination, for example through
tying arrangements if permitted by antitrust laws. Love and Hubbard (2007, pp. 1548–9)
counter that price discrimination is difficult, arguing that many pharmaceuticals are sold
only to the world’s richest group of consumers, because pharmaceutical companies worry
about arbitrage. Roin (2014, pp. 1024–5) agrees that the difficulty of stopping arbitrage,
combined with the challenge of identifying who is unwilling to pay the monopoly price,
may make price discrimination difficult. But prescription drug insurance produces a
two-part pricing that may make the two-part pricing achieved by a prize or reward system
unnecessary (pp. 1048–9).
The deadweight losses associated with patents must in any event be compared against
the deadweight losses associated with taxation. Wright’s original model featured the
simplifying assumption that ‘lump sum taxation is possible (or the marginal welfare
cost of general revenue is negligible)’ (Wright, 1983, p. 693). But the literature highlights
that taxation causes economic distortions (Calandrillo, 1998, p. 337; Abramowicz, 2003,
200-201). Roin (2014, pp. 1026–7) notes the possibility that the loss from taxation might
be even higher than the loss from patents. On the other hand, Calandrillo (1998, p. 314)
suggests that taxes are less distortionary, because they apply to all goods or income
rather than to just some subset of them. Hemel and Ouellette (2013, p. 315) counter

that ‘numerous deductions and exclusions . . . upset this equivalence and increase the
distortionary effect of income taxation.’
Whatever the actual distortionary cost of taxation, commentators widely agree that
there is political resistance to taxation. Hemel and Ouellette (p. 312) note that patents are
unique in that they do not ‘necessarily impose a budgetary cost on the government (apart
from the costs of administering the patent system).’ If taxpayers prefer, or are more easily
fooled by, the implicit taxation associated with intellectual property than the explicit taxa-
tion necessary for prizes or rewards, then prizes or rewards might be politically infeasible.
This is not necessarily an argument on the merits against prize or reward systems, but a
corollary may be that prize or reward systems are likely to be underfunded because of
their tax cost.
2. Distribution and labor incentives

A prize or reward system might benefit some consumers at the expense of others. Most
obviously, if the social value is judged independently of consumers’ willingness to pay, for
example because health benefits to the poor and rich are counted alike rather than based
on willingness to pay, then a prize or reward system would tend to encourage innovation
that benefits all people, rather than simply innovation that benefits the wealthy. Even if
the reward is based on willingness to pay, there may be a progressive redistributive effect.
If marginal cost is sufficiently low, the poor can consume a product placed in the public
domain, regardless of whether their utility is taken into account in the reward calculation.
Hemel and Ouellette (2013, p. 308) argue that any redistribution’s normative desirability
may depend on the type of good; that the patent system is subsidized by users may be
more justifiable for luxury goods (say, improved boat hull designs) than for life-saving
pharmaceuticals. Of course, a normative analysis of any distributional consequences
depends on one’s views about the existing distribution. In principle, one might argue that
distribution in favor of the poor would reduce incentives to work and thus harm effi-
ciency, though the general assumption in the reward literature has been that progressive
redistributions would be desirable.
Not all of a reward system’s distributional effects can be classified entirely in terms of
effects on the wealthy or the poor. The copyright system can be seen as a tax on readers,
and the patent system as a tax on users of inventions. Reward or prize alternatives con-
tinue to provide benefits to those same groups, but the costs are imposed more broadly.
Whether it is socially desirable, independent of efficiency effects, to redistribution from
nonreaders to readers or from Luddites to those who enjoy the fruits of invention, is ulti-
mately a normative question. Duffy (2004a, p. 44) cites Coase (1946, p. 176) in observing
that rewards ‘would redistribute income in favor of those who consume declining average
cost goods.’ Of course, if consumers are relatively homogeneous in their preference for
declining average cost goods, this may be of little concern.
E. Transactions Costs
The costs of administering either a traditional intellectual property system or a system

of rewards of prizes reduce social welfare. Calandrillo (1998, p. 341) argues that ‘the
reward regime would save on the legal, private, and social enforcement costs involved
in protecting property rights from theft, infringement, or copying by others.’ These

costs are especially significant in the context of the patent system, encompassing both
governmental and private costs associated with the prosecution and litigation of patents.
Calandrillo argues that transaction costs for both copyright and patent are high in part
because of complex doctrine (pp. 331–4). Of course, a reward system must evaluate
every invention or work and this entails some administrative costs as well. Though such
evaluations might not track traditional doctrinal argument, they might involve complex
economic analysis. Economists may not come cheaper than lawyers, who no doubt would
find a way to become involved. Burrell and Kelly (2014, p. 887) argue that, historically, the
administrative costs of prizes may be higher than those of patents.
A full accounting must take into account not only the costs involved in dealings with
government agencies and courts, but also the costs associated with private negotiation and
evaluation. The patent system encourages private ordering, but private ordering depends
on negotiation by private parties. Those parties may be able to find means of reducing
their costs, for example by joining patent pools. A reward system may reduce the need
for private negotiation, because intellectual property placed in the public domain can be
used without charge. This may be especially important in the context of ‘patent thickets’
(Heller and Eisenberg, 1998). If a sufficient number of inventions are placed into the
public domain, especially in a mandatory system, enough of the thicket may be cleared
to facilitate follow-on invention.
Market-based reward systems must be evaluated separately, because they combine
elements of the other approaches. Because a market-based reward system requires that
traditional intellectual property institutions still exist and that patents still be obtained,
the transactions costs associated with the patent office remain. Litigation, however, may
be greatly reduced with respect to the patents randomized to the public domain. But the
market mechanism itself involves transactions costs, as auction theory predicts that the
private parties participating in a Kremer auction will lower their bids to account for their
research costs. Such transactions costs may be reduced with the Duffy market valuation
approach, because each individual shareholder has only a limited incentive to engage in
research. But the downside of that may be less accurate market valuation. As long as
shareholders value patents correctly on average, however, this should not adversely affect
ex ante investments in research, except to the extent that it makes them somewhat riskier.
F. Rent-Seeking Costs
Perhaps one of the great achievements of the patent system is that it has greatly reduced
rent-seeking by inventors seeking special treatment for their individual inventions. In
principle, Congress can grant patent term extensions on an invention-by-invention basis,
but because the patent system exists and imposes uniform rules governing patent term, this
rarely occurs. A concern with the creation of a prize or reward system is that owners of
rights to inventions might lobby for those inventions to receive high rewards. Hemel and
Ouellette (2013, p. 327) worry that rewards might lead to ‘politicization, rent-seeking, and
mismanagement,’ pointing out that ‘the social rate of return on R&D funded through gov-
ernment grants has been estimated to be lower than on private R&D.’ Duffy (2004a) also
points out that consumers might lobby for the government to choose products that they
use to be placed in the public domain. The investments in such lobbying would be social
losses, and they might distort decisionmaking. In a judicialized reward process, however, it

may be possible to maintain decisionmaker independence. In that case, rent-seeking could

be accomplished only by persuading the reward judges of the worth of one’s invention.
Inventors, however, may have incentives to mislead judges about the value of their contri-
butions, for example by suppressing negative studies about their inventions’ effectiveness.
G. Other Policy Options
1. Grants
Most of this chapter has compared traditional intellectual property systems to a range of
possible prize or reward systems. But these are, of course, not the only means by which
the government can encourage innovation. An approach closely related to that of rewards
is grants. Grant decisions are made by centralized agencies ex ante based on research
proposals. Adler (2011) argues that prize or reward systems may be preferable, because
the centralized grant decisions limit ‘the range of promising ventures that may receive
funding.’ Hanson (1998) argues that grants have become more important than prizes
because they tend to concentrate power in the hands of scientific rather than political
leaders. While legislatures would like to maintain such power in their own hands, the
power to administer prizes is more likely executive than legislative, and legislatures prefer
to grant power to scientific leaders than to the executive. The general normative case for
grants, particularly for basic research, is that ‘a large share of the most valuable uses of
basic research will take highly abstract, intangible forms, rendering the output of such
research highly nonexcludable’ (Kapczynski and Syed, 2013, p. 1951).
2. Governmental purchases and tax credits

Above, we noted that the government might create something akin to a prize system by
giving coupons to low-income consumers. There may also be other ways by which the gov-
ernment can help consumers obtain patented products in a way that reduces deadweight
loss, yet possibly without the need to evaluate the contributions of individual inventions.
For example, the US government subsidizes health insurance through its tax code. Roin
(2014, p. 1051) argues that this may be administratively much simpler than giving coupons
to individual consumers. Health insurance creates a two-part pricing scheme for pharma-
ceuticals. Consumers pay a fixed amount for health insurance and then a copayment for
individual prescriptions. This copayment may be relatively close to marginal cost. Roin
also argues that the government might combine subsidies for health insurance with price
controls to ensure that the consumer price ends up relatively close to marginal cost. Many
national health systems help to drive prices for consumers to levels near marginal cost
without creating a formal prize system. The effect, he argues, is to create the benefits of
a patent buyout scheme while retaining the system of intellectual property and arguably
reducing the informational demands on the government. It may not be easy to apply the
same logic outside the health insurance context, though in principle a combination of pri-
vate ordering and governmental subsidies could reduce consumer costs for other goods.
Consumers often buy goods (such as cable television subscriptions) with a high fixed and
low marginal price component, and government policy could further encourage this.
We have, of course, only scratched the surface of the ways in which the government
can create innovation through its own purchases. Carroll (2009, p. 1369–70) offers an
expanded taxonomy of means by which the government can encourage innovation, noting

the possibility of ‘direct provision of creators or innovators employed by the government.’

Kalil (2006, p. 7) argues that a benefit of prize and reward systems is that they avoid the
complexities associated with procurement regulations and can thus attract teams who
otherwise would not be positioned to do business with the government. Brennan et al.
(2012, p. 27) offer some guidance on when procurement may be appropriate. While grants
will be useful when the government wishes to reduce the risks of inventors but does not
know the best approach to take, procurement will work better for a known objective and
solution. Prizes may reduce risk somewhat and may be especially appropriate where a
specific objective is known but the means to obtain that objective is unknown.
Hemel and Ouellette (2013) argue that tax credits also have potential advantages as
means of funding innovation. Tax credits require less direct governmental involvement
and supervision than any of the other methods of encouraging innovation. Tax credits,
they argue, ‘may be optimal where potential innovators have private information about
project prospects and limited access to outside capital’ (p. 303). Tax credits can be
structured as deductions or as refundable, which provides stronger incentives for startup
companies. Tax credits can either be administered at a national level or at a state level.
Hrdy (2013) argues that in part because a variety of doctrines limit the ability of states to
create patent policies, states have focused largely on providing tax credits as a mechanism
for encouraging innovation. As Hrdy’s argument suggests, a significant drawback of tax
credits relative to rewards is that they provide the government relatively little ability to
adjust for ways in which the patent system may not align private and social incentives.
VI. CONCLUSION
The arguments for and against prize and reward systems are many and all depend on
empirical considerations that, without experimentation, are difficult to assess with
confidence. On one hand, there are strong theoretical reasons that an ideal prize or reward
system could dominate an ideal patent system because of the reduction of deadweight
loss and because rewards can be adjusted to encourage or discourage any desired or
disfavored activity. But no system is ideal. As Kremer (1998, pp. 1162–3) points out, new
mechanisms will have ‘risks and shortcomings,’ but ‘existing mechanisms of encouraging
innovation have serious flaws’ as well. Thus, a full comparison depends on considerations
such as which system can better encourage commercialization, which system entails larger
transactions costs, and which system can best reduce inefficient rent-seeking. Perhaps
limited governmental experimentation with prize or reward systems can reduce some
of the uncertainty about these costs, but individual governments may have only limited
incentives to experiment, particularly given their obligations to maintain intellectual
property rights under existing trade laws. Perhaps the increased scholarly interest in
prize and reward systems anticipates increased governmental interest in these systems.
Levmore (2013) argues that the increasing importance of ideas in the modern economy
and the difficulty of creating property rights in ideas and the drawbacks in enforcing such
property rights mean that prizes and rewards will take on increasing importance in the
future. Pharmaceuticals may be the most likely target for experimentation (pp. 156–7),
given concerns about deadweight loss, scholarly and policy interest in that area, and the
already large spending and research by governments on health.

REFERENCES
Abramowicz, Michael, and John F. Duffy. 2008. ‘Intellectual Property for Market Experimentation,’ 83 New
York University Law Review 337–410.
Abramowicz, Michael, and John F. Duffy. 2011. ‘The Inducement Standard of Patentability,’ 120 Yale Law
Journal 1590–680.
Adler, Jonathan H. 2011. ‘Eyes on a Climate Prize: Rewarding Energy Innovation to Achieve Climate
Stabilization,’ 35 Harvard Environmental Law Review 1–45.
Bakker, Gerben. 2013. ‘Money for Nothing: How Firms Have Financed R&D-Projects since the Industrial
Revolution,’ 42 Research Policy 1793–814.
Brennan, Timothy J., Molly K. Macauley, and Kate S. Whitefoot. 2012. ‘Prizes or Patents for Technology
Procurement: An Assessment and Analytical Framework,’ Resources for the Future, Discussion Paper No.
11-21-REV, http://www.rff.org/research/publications/prizes-patents-and-technology-procurement-proposed-
analytical-framework (last accessed Mar. 11, 2019).
Brunt, Liam, Josh Lerner, and Tom Nicholas. 2012. ‘Inducement Prizes and Innovation,’ 60 The Journal of
Industrial Economics 657–96.
Burrell, Robert, and Catherine Kelly. 2014. ‘Public Rewards and Innovation Policy: Lessons from the Eighteenth
and Early Nineteenth Centuries,’ 77 The Modern Law Review 858–87.
Burstein, Michael J., and Fiona E. Murray. 2015. ‘Governing Innovation Prizes,’ unpublished manuscript, http://
static.squarespace.com/static/528679bee4b0fb796a381814/t/5357f587e4b04cbb643a8790/1398273415651/Bur
stein-Murray%20Governing%20Innovation%20Prizes.pdf (accessed Sept. 24, 2015).
Calandrillo, Steve P. 1998. ‘An Economic Analysis of Property Rights in Information: Justifications and
Problems of Exclusive Rights, Incentives to Generate Information, and the Alternative of a Government-Run
Reward System,’ 9 Fordham Intellectual Property, Media & Entertainment Law Journal 301–60.
Carroll, Michael W. 2009. ‘One Size Does Not Fit All: A Framework for Tailoring Intellectual Property Rights,’
70 Ohio State Law Journal 1361–434.
Chari, V.V., Mikhail Golosov, and Aleh Tsyvinski. 2012. ‘Prizes and Patents: Using Market Signals to Provide
Incentives for Innovations,’ 147 Journal of Economic Theory 781–801.
Clancy, Matthew S., and GianCarlo Moschini. 2013. ‘Incentives for Innovation: Patents, Prizes, and Research
Contracts,’ 35 Applied Economic Perspectives and Policy 206–41.
Coase, R.H. 1946. ‘The Marginal Cost Controversy,’ 13 Economica 169–82.
Ding, Wei, and Elmar G. Wolfstetter. 2011. ‘Prizes and Lemons: Procurement of Innovation under Imperfect
Commitment,’ 42 RAND Journal of Economics 664–80.
Duffy, John F. 2004a. ‘The Marginal Cost Controversy in Intellectual Property,’ 71 University of Chicago Law
Review 37–56.
Duffy, John F. 2004b. ‘Rethinking the Prospect Theory of Patents,’ 71 University of Chicago Law Review 439–510.
Fisher, William, and Talha Syed. 2012. ‘Prizes,’ in Infection: The Health Crisis in the Developing World and What
We Should Do About It, unpublished manuscript, http://cyber.law.harvard.edu/people/tfisher/Infection.htm
(last accessed Mar. 11, 2019).
Galle, Brian. 2014. ‘Tax, Command . . . or Nudge?: Evaluating the New Regulation,’ 92 Texas Law Review 837–94.
Guell, Robert C., and Marvin Fischbaum. 1995. ‘Toward Allocative Efficiency in the Prescription Drug
Industry,’ 73 The Milbank Quarterly 213–30.
Hanson, Robin. 1998. ‘Patterns of Patronage: Why Grants Won Over Prizes in Science,’ Working paper, http://
www.atriumtech.com/muic2008/icsc302/papers/WhyGrantsOverPrizes.pdf (accessed Sept. 24, 2015).
Hemel, Daniel J., and Lisa Larrimore Ouellette. 2013. ‘Beyond the Patents-Prizes Debate,’ 92 Texas Law Review
303–82.
Hollis, Aidan, and Thomas Pogge. 2008. The Health Impact Fund: Making New Medicines Accessible for
All, Incentives for Global Health, http://healthimpactfund.org/wp-content/uploads/2012/11/hif_book.pdf
(accessed Sept. 24, 2015).
Hopenhayn, Hugo, Gerard Llobet, and Matthew Mitchell. 2006. ‘Rewarding Sequential Innovators: Prizes,
Patents, and Buyouts,’ 114 Journal of Political Economy 1041–68.
Howse, Derek. 1998. ‘Britain’s Board of Longitude: The Finances, 1714–1828,’ 84 The Mariner’s Mirror 400–17.
Hrdy, Camilla A. 2013. ‘State Patents as a Solution to Underinvestment in Innovation,’ 62 University of Kansas
Hrdy, Camilla A. 2015. ‘Commercialization Awards,’ 2015 Wisconsin Law Review 13–86.
Kalil, Thomas. 2006. ‘Prizes for Technological Innovation,’ Brookings Institution, Discussion Paper 2006-08,
http://www.brookings.edu/research/papers/2006/12/healthcare-kalil (accessed Sept. 24, 2015).

Kapczynski, Amy, and Talha Syed. 2013. ‘The Continuum of Excludability and the Limits of Patents,’ 122 Yale
Khan, B. Zorina, and Kenneth L. Sokoloff. 2009. ‘A Tale of Two Countries: Innovation and Incentives among
Great Inventors in Britain and the United States, 1750–1930,’ in Roger E.A. Farmer, ed., Macroeconomics in
the Small and the Large. Cheltenham, UK and Northampton, MA, USA: Edward Elgar Publishing.
Review 697–754.
Kitch, Edmund W. 1977. ‘The Nature and Function of the Patent System,’ 20 Journal of Law & Economics
265–90.
Kremer, Michael. 1998. ‘Patent Buyouts: A Mechanism for Encouraging Innovation,’ 113 The Quarterly Journal
Kremer, Michael, and Heidi Williams. 2010. ‘Incentivizing Innovation: Adding to the Tool Kit,’ in Josh Lerner
and Scott Stern, eds., Innovation Policy and the Economy, Volume 10. Chicago, IL: University of Chicago Press.
Leslie, Christopher R. 2011. ‘Patent Tying, Price Discrimination, and Innovation,’ 77 Antitrust Law Journal
811–54.
Levmore, Saul. 2013. ‘The Impending iPrize Revolution in Intellectual Property Law,’ 93 Boston University Law
Review 139–62.
Lichtman, Douglas Gary. 1997. ‘Pricing Prozac: Why the Government Should Subsidize the Purchase of
Patented Pharmaceuticals,’ 11 Harvard Journal of Law & Technology 123–39.
Love, James, and Tim Hubbard. 2007. ‘The Big Idea: Prizes to Stimulate R&D for New Medicines,’ 82 Chicago-
Kent Law Review 1519–54.
Macfie, R.A. 1883. Copyright and Patents for Inventions. Edinburgh: T. & T. Clark.
McKinsey & Company. 2009. ‘And the Winner Is . . .’: Capturing the Promise of Philanthropic Prizes, https://www.
mckinsey.com/~/media/mckinsey/industries/social%20sector/our%20insights/and%20the%20winner%20is%20
philanthropists%20and%20governments%20make%20prizes%20count/and-the-winner-is-philanthropists-and
-governments-make-prizes-count.ashx (last accessed Mar. 11, 2019).
Mill, John Stuart. 1858. Principles of Political Economy. Stephen Nathanson, ed., 2004. Indianapolis, IN:
Hackett Publishing.
Moser, Petra, and Tom Nicholas. 2013. ‘Prizes, Publicity and Patents: Non-Monetary Awards as a Mechanism
to Encourage Innovation,’ 61 The Journal of Industrial Economics, 763–88.
Murray, Fiona, Scott Stern, Georgina Campbell, and Alan MacCormack. 2012. ‘Grand Innovation Prizes: A
Theoretical, Normative, and Empirical Evaluation,’ 41 Research Policy 1779–92.
Nicholas, Tom. 2013. ‘Hybrid Innovation in Meiji, Japan,’ 54 International Economic Review 575–600.
Polanvyi, Michael. 1944. ‘Patent Reform,’ 11 The Review of Economic Studies 61–76.
Ridley, David B., Henry G. Grabowski, and Jeffrey L. Moe. 2006. ‘Developing Drugs for Developing Countries,’
25 Health Affairs 313–24.
Roin, Benjamin N. 2014. ‘Intellectual Property versus Prizes: Reframing the Debate,’ 81 University of Chicago
Scherer, F.M. 1970. Industrial Market Structure and Economic Performance. Chicago, IL: Rand McNally.
Scotchmer, Suzanne. 2004. Innovation and Incentives. Cambridge, MA: MIT Press.
Shavell, Steven, and Tanguy van Ypersele. 2001. ‘Rewards Versus Intellectual Property Rights,’ 44 Journal of
Siegel, Jonathan R. 2009. ‘Law and Longitude,’ 84 Tulane Law Review 1–66.
Sobel, Dava. 1995. Longitude. New York, NY: Penguin Books.
Spulber, Daniel F. 2014. ‘Prices versus Prizes: Patents, Public Policy, and the Market for Inventions,’ Northwestern
Law & Economics, Research Paper No. 14-15, http://papers.ssrn.com/sol3/papers.cfm?abstract_id=2488095
(accessed Sept. 24, 2015).
Stiglitz, Joseph E. 2006. ‘Scrooge and Intellectual Property Rights: A Medical Prize Fund Could Improve the
Financing of Drug Innovations,’ 333 BMJ: British Medical Journal 1279–80.
Stine, Deborah D. 2009. ‘Federally Funded Innovation Inducement Prizes,’ Congressional Research Service
Report, R40677, June 29, 2009.
Tirole, Jean. 1988. The Theory of Industrial Organization. Cambridge, MA: MIT Press.
Williams, Heidi. 2012. ‘Innovation Inducement Prizes: Connecting Research to Policy,’ 31 Journal of Policy
Analysis and Management 752–76.
Wright, Brian D. 1983. ‘The Economics of Invention Incentives: Patents, Prizes, and Research Contracts,’ 73
The American Economic Review 691–707.

PART III
IP COSTS

14. Tailoring intellectual property rights to reduce
uniformity cost
Michael W. Carroll* 12
Contents
I. Introduction
II. Intellectual Property’s Uniformity Presumption
III. The Problem of Uniformity Cost
IV. The Argument for One-Size-Fits-All
V. Minimizing Uniformity Cost
A. Tailoring Through Private Ordering
B. Standards Instead of Rules
i. Standards in patent law
ii. Standards in copyright law
C. Tailoring Through Real Option Pricing
VI. Tailoring Exclusive Rights
VII. A Framework for Tailoring
A. Alternative Incentives to Create or Innovate and Distribute
B. Alternative Appropriability Mechanisms
C. Overlapping Rules and Rights
D. Demand-Side Features—Positive Spillovers
VIII. Industrial Innovation Decision Structures
IX. Conclusion
References
I. INTRODUCTION
This chapter focuses on solutions to a second-order problem that arises with the creation
of intellectual property (IP) rights—the problem of uniformity cost. The incentives cre-
ated by one-size-fits-all patents and copyrights often are misaligned with those necessary
to attract the optimal level of investment of capital and creative labor. Uniformity cost
is the social cost attributable to this uniform approach to institutional design. Licensing
and other forms of private ordering minimize some uniformity cost. Current intellectual
property law also deploys a range of strategies to minimize residual uniformity cost: (1)
defining rights through adaptable standards; (2) some use of real options instead of direct
* Professor of Law and Director of the Program on Information Justice and Intellectual
Property, American University Washington College of Law.
377

entitlement grants; and (3) selective use of tailored exclusive rights. After brief discussion
of these first two strategies, the remainder of this chapter focuses on the last.
The economic reasoning supporting tailoring of exclusive rights is a logical extension
of the basic economic justification for intellectual property. In a competitive economy,
we should expect underinvestment in creative and inventive endeavors without some form
of government assistance because of declining marginal cost. Once an author, inventor,
or financier has paid for the creation, commercialization, and distribution of a valuable
creative or innovative work, competitors can reproduce and distribute that work at
prices too low for those who invested in the creation to recoup their investments (Menell
and Scotchmer, 2007). Some scholars qualify their views or disagree that this baseline
appropriability problem is present in markets for works of authorship (Bracha and Syed,
2014; Lunney, 2009; Yoo, 2008).
The standard economic solution to this appropriability problem is government action.
For some, creating exclusive rights to stimulate market exchange is the obvious response.
But, this move from problem to property is socially expensive (Macaulay, 1900) and merits
a more robust justification because policymakers may select from, or combine, six types
of policy intervention to solve the appropriability problem:
1. direct provision of creators or innovators employed by the government;

2. direct compensation ex ante to the innovator for producing the information (leaving
the costs of reproduction and distribution to be borne by participants in a competi-
tive market), such as through a grant to a promising innovator;
3. direct compensation to the innovator ex post through a reward or prize system for
innovations already created, such as prizes awarded by a number of federal govern-
ment science agencies;
4. indirect compensation to the innovator through tax policy, by, for example, giving tax
credits for investments in research and development;
5. protection of innovators from competition through the grant of exclusive production
and distribution rights (thereby encouraging monopoly pricing), by creating patent
and copyright law; or
6. other forms of protection from competition or misappropriation by increasing
excludability of valuable information, by, for example, prohibiting circumvention of
technological protection measures (Fisher, 2004; Frischmann, 2000; Lessig, 2001).
These strategies can be, and usually are, combined in a number of ways. Discussion of
how to determine the optimal mix of strategies has begun (Hemel and Ouellette, 2013),
and it is now clear that the intellectual property strategy must compete with these others
for its place in the policy mix (Burstein, 2012). The early law and economics literature
equated a property-based solution with capitalism (Menell and Scotchmer, 2007). Arrow
(1962), for example, declared that ‘[i]n an ideal socialist economy, the reward for invention
would be completely separated from any charge to the users of the information’ whereas
‘[i]n a free enterprise economy, inventive activity is supported by using the invention to
create property rights’ (emphasis added). By framing the choice as between socialism and
‘free enterprise,’ readers were instructed that a reward system for invention and a free
enterprise economy are incompatible options.
Today, the case for intellectual property rights cannot disregard the direct-

Tailoring intellectual property rights to reduce uniformity cost 379
compensation alternatives so summarily. For example, traditional economic analysis

offered only equivocal support for the efficiency of the patent system (Machlup, 1958).
More radical economists have reformulated an older critique (Janis, 2001; Machlup et al.,
1950) and have concluded that Machlup was too timid and that the absence of historical
data sufficient to support the case leads to the conclusion that ‘intellectual property is an
unnecessary evil’ (Boldrin and Levine, 2008). Even if government intervention is justi-
fied, direct compensation to innovators appears to be preferable to intellectual property
rights at first glance because this strategy avoids the social costs of underutilization
(Kesan and Gallo, 2009; Polanvyi, 1944). In the modern context, an emerging literature
supports this view, advocating for a range of institutional forms of direct compensation
(Abramowicz, 2003; Kremer, 1998; Lichtman, 1997; Shavell and van Ypersele, 2001).
Stiglitz (2006b) has endorsed the use of prizes in the drug discovery context, and Shavell
(2004) further observes that ‘in many plausible situations, the reward system would be
superior to the property rights system.’ Shavell and Ypersele (2001) show that giving the
innovator the option to choose either a reward calculated from ex post data, such as sales
figures, or exclusive rights would be preferable to the current patent system. However,
in rebuttal, Roin (2014) argues that even if the government’s information is as good as
private innovators,’ a range of reasons supports supplementing intellectual property
rights with various government interventions rather than replacing intellectual property
rights with prizes.
II. INTELLECTUAL PROPERTY’S UNIFORMITY

PRESUMPTION
Notwithstanding the advantages of competing strategies, the case for intellectual property
rights retains its vitality in light of three practical considerations necessary to assessing
the feasibility and desirability of the various strategies for solving the appropriability
problem: (1) the government’s ability, relative to private actors, to value certain types or
classes of creation or innovation (Shavell and van Ypersele, 2001); (2) the comparative
administrative cost of a strategy; and (3) considerations of political economy associated
with a particular strategy. By testing the case for intellectual property in relation to these
practical metrics, the empirical and practical premises of the case are laid bare.
The economics of innovation make direct investments of public funds challenging.
Investments in innovation and new cultural works often are uncertain and risky because
the likelihood of any success is often difficult to predict and the distribution of returns
on investment is highly skewed (Scherer, 2001b). While, in theory, the government can
manage this uncertainty through post hoc rewards or prizes, the government will have
difficulty in many circumstances calibrating the reward to the social value contributed by
the creator or inventor.
These characteristics give rise to the information-asymmetry justification for intellectual
property rights. The government’s relative ignorance about the incentives required to lure
particular creators or innovators into the information production and distribution game
and to keep them in it justifies the social costs imposed by intellectual property rights.
Inventors or creators will have, on average, marginally better information about their
potential success in the markets for their respective intellectual outputs, and government

strategies for socializing this private information through some form of mechanism design
or otherwise are not likely to fully succeed. For this reason, increasing an innovator’s
ability to exclude (or at least deter) competitors through exclusive rights is superior to
a reward system because the intellectual property system is driven by the marginally
superior private information that innovators enjoy (Kitch, 2000; Lunney, 2003; Shavell
and van Ypersele, 2001).
The cognate risk-spreading justification focuses on the costs of failure. If the govern-
ment opts for direct procurement of innovation through employment or grants, the
government assumes the risk of incurring these costs. Intellectual property rights spread
this risk among potential rightsholders, who either may fail to produce information that
qualifies for protection or who may succeed in acquiring rights only to have the market
deem these rights worth less than the cost of acquisition. Spreading the risk of failure
among investors with relatively superior private information is likely to allocate resources
marginally more efficiently than concentrating this risk in the government.
Some argue that the risk-spreading justification is not a persuasive comparative jus-
tification for the intellectual property strategy because the reward or prize strategy also
enables the government to spread the risk of failure among potential innovators while
avoiding the social costs of intellectual property rights. However, prizes or rewards must
be designed to produce the desired expected value in the mind of innovators and creators,
and there are three types of risk that the government must manage: (1) identifying the
kinds of inventions and creative works eligible for reward; (2) identifying the stage of
development at which to grant rewards; and (3) quantifying the reward. The prize or
reward strategy concentrates the risk of error in any of these three decision points in the
government.
When designing intellectual property rights, the government still risks error at each of
these same decision points, but the magnitude of risk is reduced because markets enabled
by intellectual property rights have flexible features that correct to some degree for misal-
location of rights. These markets potentially spread decisions about which risks should
be undertaken and who should bear them. Markets also spread discipline for those who
waste assets in pursuit of creative or innovative goals. Finally, a less tangible risk that
intellectual property rights spreads is the risk of counterparties cheating. For a prize or
reward strategy to succeed, potential innovators must trust that the government will pay
when a reward has been earned. In the markets enabled by intellectual property rights,
the potential sources of revenue are spread among consumers, and the risks that they will
take valuable information without paying also is spread.
One significant limit on the efficiency of markets enabled by intellectual property rights
should be recognized. The gap between willingness-to-pay and ability-to-pay poses a
significant obstacle for the standard argument favoring decentralization through private
investment and market exchange and is a drawback that receives insufficient attention
in the literature (Frischmann, 2001). The information and product markets supported
by intellectual property rights operate on the basis of users’ ability to pay rather than
willingness to pay to reflect the social value of innovation. As a result, the innovations
or innovators selected for reward by market exchange will skew toward the interests of
those with an ability to pay. Using prices to allocate access to goods and services does not
accurately reflect how relatively important this access is to different individuals because
of the declining value of the marginal dollar (Shaver, 2014).

Reliance on patents to facilitate financing of drug discovery is the case that draws
the most attention to the selection effects of ability to pay. The patent system directs
significant resources to the discovery and development of so-called ‘me-too’ and ‘lifestyle’
drugs for which wealthy consumers in industrialized economies are able to pay a hefty
premium (Love and Hubbard, 2007). These resources are not directed toward discovery
and development of cures for tropical diseases because the likely beneficiaries lack
the ability to pay such premiums, even though their willingness to pay for a drug that
would keep a child alive almost certainly is greater than what an aging consumer in the
industrialized world would be willing to pay to enhance his or her sexual performance
(Fisher and Syed, 2006). The likely distortionary effects on resource allocation that follow
from the gap between willingness-to-pay and ability-to-pay for innovation should serve
as a signal for policymakers when choosing the right mix of intellectual property and
direct-compensation approaches. For fields of creativity or innovation in which the gap is
significant, the direct-compensation strategy, coupled perhaps with a tax strategy, is likely
to perform better (Saulo et al., 2008).
In sum, as against a centralized government compensation scheme or prize fund, the
intellectual property strategy offers the benefits of decentralization (Lunney, 2003; Wu,
2006). This strategy harnesses the sometimes superior private information that creators
and innovators have about the value of their cultural or technological contributions
and spreads the risk that they and their financial backers may be mistaken about the
practical feasibility of a creative or innovative idea and about its market valuation if
realized (Knight, 1921). It also spreads the risk of mismeasuring the timing or amount
of reward necessary to induce desired innovations and creative works. The social cost
of this strategy is that resources are channeled to innovations likely to serve those with
an ability to pay, and these innovations are likely to be underutilized by those who are
priced out of the market by monopoly prices or by those who are denied licenses because
the transaction costs are too high or the rightsholder refuses to license. The market also
is likely to skew toward production of consumption information goods rather than
productive information goods because of the problem of valuing positive spillovers
(Frischmann and Lemley, 2007). After balancing these costs and benefits, policymakers
must then assess whether whatever net benefits this strategy yields are comparatively
worth the administrative cost.
The literature is relatively silent about how the above justification for the intellectual
property strategy informs the institutional design of intellectual property rights. The
consensus supports the proposition that the subject matter, scope, and duration of intel-
lectual property rights should be no more robust than necessary to induce the desired
level of investment in cultural and technological innovation (Fisher, 1998; Lunney, 2003;
Wasserman, 2013). Some scholars disagree about the degree to which the scope of intel-
lectual property rights can explain market behavior (Duffy, 2005; Lemley, 2005). And, most
commentators recognize that the magnitude of the appropriability problems that these
rights are designed to remedy varies considerably across and even within industries (Burk
and Lemley, 2009; Cotropia and Gibson, 2010). From these two propositions, one would
expect an institutional design that relies on context. Instead, two aspects of current law are
generally assumed rather than justified: (1) patents and copyright appropriately offer differ-
ent levels of protection (Dinwoodie, 2010; Karjala, 2003; Long, 2004); and (2) within each
branch of law, exclusive rights are generally uniform across all protected subject matter.

III. THE PROBLEM OF UNIFORMITY COST
From a theoretical perspective, this policy of one-size-fits-all patents and copyrights

necessarily is inefficient (Carroll, 2007). In particular, this policy imposes a uniformity
cost on society, defined as the social costs attributable to the misalignment between the
general level of exclusion provided by uniform rights and the specific level of exclusion
that would be optimal with respect to any given author or inventor (Carroll, 2007; Lunney,
2003). Early literature recognized uniformity cost with respect to certain dimensions of
intellectual property rights—in particular the costs imposed by a uniform patent term
(Cornelli and Schankerman, 1999), discussed more fully in Chapter 13 of volume II
(Love, 2018). But, the generality of the problem and modes of analysing solutions beyond
formal modeling are more recent developments.
Lunney (2003) models uniformity cost by assessing the trade-offs between strictly
uniform rights, rights tailored to individual innovations, and certain intermediate options.
Adams (2002) adds that an economy’s technological capacity must also be considered
when assessing these options. At bottom, Lunney shows that ‘[e]ven where an innovative
product represents the most valuable use of available resources . . . an optimal uniform
scheme of protection will provide protection that will leave some desirable innovative
products unprofitable’ (Lunney, 2003; Abramson, 2001). Carroll’s (2007) model applies to
both patents and copyrights to similar effect. Bracha and Syed (2014) and Cotropia and
Gibson (2010) refine the description of both static and dynamic uniformity cost.
Some empirical data provide industry-specific examples of uniformity cost. Mansfield
(1986) interviewed research and development managers from 100 randomly selected
firms to ask what percentage of each firm’s inventions would have been developed and
brought to market in the absence of patent protection. Although any counterfactual
query introduces certain biases and uncertainties, especially when posed to interested
parties, Mansfield’s data indicate that: (1) a significant percentage of inventions would
have been developed and brought to market without the prospect of patent protection;
(2) this effect varies significantly by industry; and (3) the availability of protection
resulted in 80 percent of patentable inventions being patented in industries with high
patent-dependencies (pharmaceuticals, chemicals, petroleum, machinery, and fabricated
metal products) and 60 percent of inventions being patented in less patent-dependent
industries (primary metals, electrical equipment, instruments, office equipment, motor
vehicles, rubber, and textiles) (Mansfield, 1986). Similar results have been found in
the semiconductor manufacturing industry (Hall and Ziedonis, 2001; Ziedonis, 2004).
Similarly, W. Cohen et al. (2000) and Levin et al. (1987) found respondent R&D managers
in their respective surveys to be skeptical about the contribution of patents to the firm’s
ability to appropriate returns on investment.
The scope of the uniformity-cost problem is exacerbated by the fact that the distribu-
tion of rewards from both cultural and technological innovation is highly skewed (Scherer,
2001b; Scherer et al., 2000). For example, uncertainty about demand or about feasibility
leads recording companies, motion picture studios, pharmaceutical companies, and
biotechnology research firms to invest millions of dollars that will never be recouped in
innovation (Grabowski, 2002; Mansfield et al., 1977). In these industries, profits from
chart-topping songs, blockbuster movies, and blockbuster drugs must be sufficient to
cover the losses incurred on other investments (DeVany and Walls, 2004). Consequently,

industries such as these demand robust intellectual property rights to maximize the profit-
ability of successful innovations. When these rights apply uniformly, the social costs are
magnified.
IV. THE ARGUMENT FOR ONE-SIZE-FITS-ALL
The literature lacks a clear defense of uniform patents and copyrights that responds to
the uniformity-cost problem. This section attempts to extrapolate the best economic
argument for adopting one-size-fits-all patents and copyrights from the basic justification
for intellectual property, at least as a matter of initial entitlement design. Acknowledging
that uniform rights are a second best approach to institutional design, the argument for
uniformity, at bottom, is that resort to the second best in both cases is grounded in the
same information-asymmetry and risk-spreading justifications (DeBrock, 1985).
For the government to tailor rights according to subject matter (e.g., software) or the
status of the innovator (e.g., a university researcher), the government would require some
basis for distinguishing among classes of innovation or innovators. The information-
asymmetry justification for intellectual property rights also justifies adopting uniform
rights targeted at the average level of exclusion needed to stimulate the desired level of
investment in innovation throughout the economy. The government initially lacks the
information necessary to make principled distinctions among innovators or classes of
innovation.
Moreover, innovation is inherently dynamic (Merges and Nelson, 1990). Even if the
government were able to gather information sufficient to make a principled distinction
between the rights accorded to innovation category A and innovation category B, the
boundaries of those categories are likely to change and are subject to legal arbitrage (Burk
and Lemley, 2009). Technology-specific or subject matter-specific laws will become out-
dated quickly and therefore fail to provide the efficiency gains sought by tailoring rights.
The uniform rights proponent, acknowledging difficulties in measuring the total social
costs of intellectual property rights, is likely to rely on an intuitive sense that evidence of
robust intellectual property licensing activity combined with legal entrepreneurship to
reduce transaction costs through innovative licensing structures means that the magnitude
of uniformity cost is not significant enough to justify very much tailoring. The proponent
is likely to argue that broadly defined uniform rights facilitate licensing and enforcement,
and so the policy choice really is one between tailoring through public or private ordering.
On this view, the intellectual property system is rendered more administrable through
broadly defined rights.
Smith (2007) and Long (2004) each offer versions of this argument. Acknowledging the
theoretical case in favor of rewards or perfectly tailored rights, Smith (2007) concludes
that broadly defined rights of exclusion are more efficient in practice because they render
a complex innovation system wieldy by limiting the information required by officials
entrusted to administer it. On his view, rights of exclusion create a modular system in
which those who administer it must only attend to boundaries and need not gather
information necessary to govern activities taking place within the boundaries.
Long (2004) similarly concludes that the administrative benefits of uniform rights in
the patent system outweigh the benefits that more finely tailored rights would provide.

For example, she acknowledges that certain classes of patentable subject matter, such as
business methods and software, are qualitatively different from the paradigmatic inven-
tions that informed the structure of patent’s uniform rights. She recognizes that tailoring
might reduce the information costs associated with transactions and enforcement. She
nevertheless concludes that the trade-offs associated with tailoring rights may not be
cost-justified on information-cost grounds.
Others argue that the benefits of broadly defined rights are illusory, particularly with
respect to patent scope, and that tailoring could improve efficiency in multiple contexts
(Bessen and Meurer, 2009; Carrier, 2007; Sarnoff, 2004). All participants in this debate
agree that weighing the relative administrative costs, including transaction costs, of a
system of broadly defined or more tailored rights is largely an empirical matter.
Certain considerations of political economy also counsel in favor of uniformity as the
default position for domestic policy.1 The strongest argument is that domestic uniformity
increases the costs of rent-seeking by industries and other interest groups. Copyright and
patent legislation serves for some as a paradigm public-choice case because such legisla-
tion generally is the product of bargaining among industry groups with little or no user
or consumer representation because they cannot overcome the costs of collective action
(Litman, 2001; Patry, 1996). Commentators suggest that interest group involvement in
copyright and patent legislation has intensified in recent years (Burk and Lemley, 2003;
Merges, 2000). With uniform patents or copyrights, legislative change must submit to
what Olson (1989) calls the ‘iron law of consensus,’ by which all industries affected by the
law must agree for an amendment to pass through the many veto points in the legislative
process. The argument holds that even if a particular tailoring measure would produce
a superior partial equilibrium, legislative practice that would routinely grant additional
rewards or create special carve-outs for individual interest groups would intensify the
social costs resulting from successful rent-seeking already apparent in the process (Long,
2004).
One final argument a uniform rights proponent might rely on—advanced primarily
and independently by Abramowicz (2004, 2005, 2011) and Yoo (2004)—is that the overall
social costs of intellectual property, at least in copyright law, are less than often claimed
because monopolistic competition tends to dissipate copyright-generated rents and that
broader rights might better manage this competition while driving price close to marginal
cost. If this is correct, then even if uniformity cost remains a component of copyright’s
social cost, its magnitude is small enough to be tolerated. Bracha and Syed (2014) agree
that copyright may be attracting inefficient entry, but they assert that this is an ineluctable
1
The argument for the relative efficiency of domestic uniformity is distinct from the politi-
cal economy argument for uniform rights in international law. Evidence that would support the
political discipline argument for domestic legislative policy does not equally support the argument
for international harmonization. On the contrary, the average level of exclusion needed in any
particular economy will vary in part on the extent to which it is a net importer or exporter of goods
or services embodying protected information (Scotchmer, 2004). Thus, the push for international
harmonization, which imposes the same average level of exclusion and is now encoded in the
Agreement on Trade-Related Aspects of Intellectual Property Rights (TRIPS), can be viewed as
successful rent-seeking spearheaded by multinational rightsholding corporations headquartered in
the United States, Europe, and Japan (Braithwaite and Drahos, 2007; Correa, 2006; Rajec, 2013).

by-product of granting copyright owners rent-generating power over price. From the
uniform rights perspective, although these commentators draw opposing conclusions
about proper doctrinal responses, their proposals favor a general rather than tailored
approach to managing entry through the design of exclusive rights.
V. MINIMIZING UNIFORMITY COST
Other commentators are less sanguine about the relative efficiency of uniform rights.
Emphasizing the need to devise institutional design approaches that minimize uniformity
cost, these commentators have drawn attention to three undertheorized strategies already
found in contemporary patent and copyright law.
A. Tailoring Through Private Ordering
Implicit in the case for uniform exclusive rights is recognition of three features of market
exchange that mitigate uniformity cost: demand elasticity, price discrimination, and
Coasean bargaining (Carroll, 2009). As Arrow (1962) recognized, the social costs of
intellectual property rights arise only when there is demand for protected information
(Mansfield et al., 1977). Along with the other social costs of intellectual property, uni-
formity cost rises with demand (Carroll, 2009; Lunney, 1996).
Even when uniformity cost arises, under traditional economic analysis, perfect price
discrimination theoretically would eliminate the underdistribution of protected informa-
tion. That is, if intellectual property owners are able to engage fully in first-degree price
discrimination—selling or licensing to each user willing to pay more than marginal
cost—static deadweight loss would be zero (Demsetz, 1970; Kaplow, 1984; Shapiro and
Varian, 1999).
As others have shown, however, even as a matter of theory, perfect price discrimination
would not eliminate all social costs of intellectual property rights (Bhaskar and To, 2004;
Cohen, J., 2000; Edlin et al., 1998; Frischmann, 2005; Lunney, 2008; Meurer, 2001).
Moreover, even if perfect price discrimination would theoretically avoid reduction in
social value, perfect first-degree price discrimination in the intellectual property context
is a practical impossibility (Farber and McDonnell, 2003; Yoo, 2004).
Finally, when demand is positive and price discrimination is imperfect, the Coase
Theorem asserts that uniformity cost will affect allocative efficiency only if reallocation
or reapportionment of uniform entitlements by contract is too costly (Coase, 1960).
Through licensing and non-enforcement of intellectual property rights, those who need
to use another’s information will obtain access and the practical ability to use it.
Most commentators agree that difficulties in valuing patents and copyrights raise trans-
action costs to the point that allocative efficiency will depend upon how the subject matter,
scope, and duration of intellectual property entitlements are defined (Barnett, 2009; Burk
and McDonnell, 2007; Burstein, 2012; Carroll, 2009; Long, 2000). This is particularly true
because the externalities that justify patent and copyright law differ fundamentally from
those that inspired Coase (Lemley, 2005), and the law’s choice is not between granting
an entitlement to party A or to party B, but between granting an entitlement to party A
or to the public at large, comprised of an unknown and often unknowable proportion of

higher- and lower-valued users (Carroll, 2009). Consequently, allocative inefficiency in

intellectual property law potentially imposes a far more significant social cost than it does
with respect to tangible property (Frischmann and Lemley, 2007).
B. Standards Instead of Rules
Intellectual property law defines the exclusive rights in patent and copyright law through
a mix of rules and standards.2 Duration is defined primarily by rules, although some com-
mentators have shown that courts can use their discretion over scope to limit or enhance
the effective duration of protection (Hughes, 2003; Liu, 2002). As is discussed below, the
patent system mitigates the uniformity cost that arises from rules’ inflexibility by using
real options to vary the actual distribution of protection (Weeds, 2001). With respect to
scope, however, patent and copyright laws define rights with flexible standards that allow
relevant decisionmakers to minimize uniformity cost.
Legal standards confer interpretive discretion on adjudicators and, generally, the more
broadly a standard is stated, the more discretion adjudicators have. This interpretive dis-
cretion can be deployed ad hoc or systematically. With respect to the scope of intellectual
property rights, courts can choose to use flexible doctrines to strike the incentives-access
balance either on a per-work (Cooper, 1993) or per-invention basis, or more broadly along
industry-specific or technology-specific lines (Burk and Lemley, 2009; Wagner, 2003).
With regard to subject matter, for example, courts have a certain amount of discretion to
determine whether a work is sufficiently original or to draw the line between unprotected
idea and protected expression (Hardy, 2001). Similarly, determining whether a process is
protectable or whether a biological organism is a ‘machine,’ a ‘manufacture,’ or ‘composi-
tion of matter,’ requires the exercise of interpretive discretion through which the courts
can tailor protection.
i. Standards in patent law

Burk and Lemley (2009) argue that this is where the solution to the problem of uniform-
ity costs lies—at least in patent law. In sum, they argue: (1) that a purely unitary patent
system no longer fits the extraordinarily diverse needs of innovators in today’s technology
industries; (2) that the solution is not to split the patent system into industry-specific
statutes, but to tailor the unitary patent rules on a case-by-case basis to the needs of
different industries; and (3) that it is the courts, not Congress or the United States Patent
and Trademark Office (USPTO), that are best positioned to do this tailoring.
In patent law, the formally uniform statutory definition of patentable subject matter is
2
For present purposes, the following definitions make the point:
1. Rules—A legal directive is ‘rule’-like when it binds a decisionmaker to respond in a
determinate way to the presence of delimited triggering facts.
2. Standards—A legal directive is ‘standard’-like when it tends to collapse decisionmaking
back into the direct application of the background principle or policy to a fact situation
(Sullivan, 1992).
The rules/standards literature is substantial (Kaplow, 1992; Kelman, 1987; Kennedy, 1976;
Posner, 1997; Radin, 1989; Rose, 1988; Schauer, 1991; Schlag, 1985; Sunstein, 1995).

broadly stated. Courts could use this to identify categorical exclusions from patentable
subject matter, but courts have resisted invitations to do so other than for ‘abstract ideas,
natural phenomena, and laws of nature.’ As a result, in the name of uniformity, the courts
have extended patent protection to living organisms (Diamond v. Chakrabarty, 447 U.S.
303, 308–9 (1980)), methods of doing business (State Street Bank & Trust Co. v. Signature
Financial Group, Inc., 149 F.3d 1368, 1375 (Fed. Cir. 1998)), and software (AT&T Corp.
v. Excel Commc’ns, Inc., 172 F.3d 1352, 1356 (Fed. Cir. 1999)). Nonetheless, courts retain
discretion to reduce uniformity cost on a per-patent basis through flexible subject matter
doctrines including the limitation on patenting ‘abstract ideas’ (O’Reilly v. Morse, 56 U.S.
62 (1853)), the utility doctrine (Burk and Lemley, 2009), novelty’s requirement of a prior
‘public’ use (Bernhardt, L.L.C. v. Collezione Europa USA, Inc., 386 F.3d 1371, 1381 (Fed.
Cir. 2004)), and the non-obviousness standard (35 U.S.C. § 103). Application of other eli-
gibility doctrines, such as the disclosure requirements of enablement, and written descrip-
tion, as well as the triggers for statutory bars, all supply tools for the courts to assess and
reduce uniformity cost (Burk and Lemley, 2009). For example, non-obviousness and the
disclosure doctrines vary to a certain degree along technology-specific or industry-specific
lines because these are applied with reference to a ‘person having ordinary skill in the art’
(PHOSITA) (Burk and Lemley, 2002, 2009; Kesan and Gallo, 2006; Wagner, 2004).
When specifying eligibility through the PHOSITA device, the law requires a court or
a patent examiner to make a variety of judgments concerning the level of skill in the art
and the set of background knowledge that the PHOSITA would be able to rely upon when
drafting or reading a patent (Burk and Lemley, 2002, 2009). Allison and Tiller (2003)
note that ‘[w]hen one realizes that an ordinarily skilled practitioner may range from an
experienced mechanic or electrician to a person with a Ph.D. and much experience in
molecular biology or computer science, the conclusion is inescapable that not all rules can
be applied exactly the same in every case.’ The courts can and do vary patent eligibility
for different industries or technologies by the amount of information and the kinds of
technical skills that a patentee can incorporate by reference (In re GPAC, Inc., 57 F.3d
1573, 1579 (Fed. Cir. 1995)).
Reliance on the PHOSITA also provides a basis for tailoring patent scope, particularly
through the doctrines of non-obviousness, enablement, and written description (Burk
and Lemley, 2002, 2009; Eisenberg, 2004; Merges and Nelson, 1990). Patent scope also
can vary along industry-specific or technology-specific lines through application of the
doctrine of equivalents (Cohen, J., and Lemley, 2001; Cotropia, 2005). Finally, in the case
of patent injunctions, the flexibility in the standard for relief should lead to industry-
specific patterns because of industry-specific differences in facts that are salient under the
standard (Carroll, 2007; Sarnoff, 2009).
ii. Standards in copyright law

Copyright law also uses standards to define the subject matter and scope of its exclusive
rights to reduce uniformity cost. With respect to subject matter, copyright law provides
courts with even greater doctrinal flexibility than does patent law. Principally, these
doctrines are the idea/expression dichotomy, the functionality exception, and the merger
doctrine. Copyright applies only to the author’s original expression and not the abstract
ideas embodied in the copyrighted work (17 U.S.C. § 102). Likewise, facts are not copy-
rightable but an author’s expression in relating facts usually will be sufficiently original

to be copyrightable. Similar line-drawing difficulties arise and may be resolved differently

depending on subject matter. The merger doctrine holds that if there are limited means to
express ideas or facts, then the expression merges with the uncopyrightable element and
the whole of the author’s work is either uncopyrightable or the copyright in the expression
is unenforceable (Joyce et al., 2016). The functionality doctrine is related to the merger
doctrine and holds that protection for expressive sculptural, pictorial, and graphic works
that are combined with functional goods is limited only to expressive elements that are
physically or conceptually separable from the functional good (17 U.S.C. § 101).
The scope of rights under copyright is determined in relation to a number of
context-sensitive standards. For example, whenever the defendant’s work does not
literally reproduce the plaintiff’s work, the court must resolve whether the two works
are ‘substantially similar’ from the ‘ordinary observer’s’ perspective (Samuelson, 2013).
Both of these judgments are context-sensitive and can be applied to reduce uniformity
cost. The most notable example of a court using this flexibility is Computer Assocs. Int’l,
Inc. v. Altai, Inc., 982 F.2d 693 (2d Cir. 1992), which endorsed the use of an abstraction-
filtration-comparison method for determining substantial similarity in software cases.
Even when literal copying takes place, the copyright owner’s rights are limited by flexible
standards, such as fair use (17 U.S.C. § 107). This doctrine is flexible enough to grant
courts substantial tailoring discretion (Harper & Row, Publishers, Inc. v. Nation Enters.,
471 U.S. 539, 552 (1985)), as are less-frequently invoked infringement doctrines such
as de minimis use (Newton v. Diamond, 388 F.3d 1189, 1193–4 (9th Cir. 2004)), scènes
à faire (Walker v. Time Life Films, Inc., 784 F.2d 44, 50 (2d Cir. 1986)), and the useful
article doctrine. Moreover, courts have license to be flexible with the choice of a remedy
(17 U.S.C. § 502(a)).
C. Tailoring Through Real Option Pricing
To further minimize uniformity cost, intellectual property law supplements the private
tailoring of intellectual property rights through market exchange by using ‘real options’
in entitlement design. Policymakers have three choices when allocating entitlements: (1)
directly granting the entitlement to all eligible holders; (2) granting an option to acquire
the entitlement to all eligible holders (a call option); or (3) granting multi-tiered options
to acquire the entitlement, that is, an automatic grant of an option to acquire an option
to acquire the full entitlement, etcetera (Ayres, 2005). Whenever intellectual property law
requires potential owners to take affirmative, costly acts to acquire, maintain, or enforce
their rights, it effectively requires these potential owners to place an option value on the
prospect of protection. When the option price exceeds the owner’s expected value, the
owner will forgo or give up their rights, thereby reducing uniformity cost (Carroll, 2006).
Option prices also reveal information about the value of the entitlement. One goal
of entitlement design can be to force private actors to reveal their private valuations of
options regulated by legal rules (Fennell, 2005). In patent and copyright law, call options
serve two important economic functions: (1) limiting the number of entitlement holders
and thereby reducing social costs by tailoring the number of entitlements granted; and
(2) producing coarse-grained information about the private valuation of the entitlement.
Of course, the substantial cost of litigation also functions as a real option that can either
reduce or increase uniformity cost (Depoorter and Walker, 2013).

Patent law deploys call options along both the subject matter and duration dimensions.
With respect to subject matter, not every inventor of a new, useful, and non-obvious
process, machine, manufacture, or composition of matter receives a patent because a
potential patentee must undergo a time-consuming and expensive process to prosecute his
or her claim to a patent (Long, 2002). The potential patentee must assess the option value
or strike price of patent protection and compare that to the costs of exercising the option
through patent prosecution (Lemley and Sampat, 2008). The option value of patent
protection in a given case usually is comparative because the potential patentee generally
also has the options to keep trade secret protection or acquire the benefits of defensive
publication (Eisenberg, 2000; Denton and Heald, 2003). When a potential patentee for-
goes protection, society is spared the associated costs. Real options reduce uniformity cost
by filtering out low-value inventions across all industries (Long, 2002). Empirical research
indicates that real options also reduce uniformity cost because the value placed on patent
protection generally varies by industry (Cohen, W., et al., 2000; Levin et al., 1987).
Call options in the form of maintenance fees effectively tailor patent duration (Lunney,
2003). A patent owner’s decision not to pay the relatively modest maintenance fees is a
decision to dedicate the invention to the public domain. One study shows that the owners
of more than half of all patents choose to dedicate their inventions to the public domain
prior to the expiration of the full 20-year term (Moore, 2006). Data for fee payments
during the ten-year period from 1994 to 2003 show that, on average, 18 percent of patent
owners placed little value on their patents and permitted protection to lapse at the 3.5
year mark; 42 percent of patent owners who had proceeded past the first stage chose not
to extend protection at the 7.5 year mark; and of those patentees who previously had
purchased extended protection fully, 64 percent chose to end the patent term at the 11.5
year mark (Parchomovsky and Wagner, 2005). A more recent study of patents allowed
to expire between January 1, 2008 and December 31, 2012 found that among this group,
patent expiration generally coincided with the due date of the second maintenance
payment (mean age of 8.18 years), but that there was some variation by patent class
(O’Donoghue, 1998; Vishnubhakat, 2015).
By viewing these rules as filters, the uniformity-cost perspective reframes at least two
debates that have engaged economically oriented scholars. First, the ‘patent quality’
debate can be recast as a debate about setting the right price for the option of patent
protection. Most commentators appear to agree that some real option should be placed
on patent acquisition (Partnoy, 2001). There also seems to be consensus that the option
price should be relatively high, by requiring prosecution and examination rather than
mere registration (Duffy, 2002; Ghosh, 2004; Kieff, 2003). Most scholarly debate has
focused on whether the mesh of the current examination filter should be made smaller to
restrict the flow of invalid patents into the system (Ghosh and Kesan, 2004). Alternatively,
the examination option could be tiered to force greater revelation of an inventor’s private
valuation of the invention (Lemley et al., 2005). In general, improving quality control in
the USPTO would tend to increase the option value necessary to make pursuing patent
protection cost-justified.3
3
The portfolio strategy used by many firms suggests that the option value has to be calculated
not only in reference to potential revenues from the exploitation of individual inventions but also

Second, as is explained more fully in Chapter 13 of volume II, the extensive economic
literature on optimal patent duration generally overlooks the role of real options either by
assuming a uniform term or contemplating variability without analysis of how it might
be implemented (Love, 2018). Some analysts seek to make the case for the efficiency of
a uniform term, not recognizing that real options render actual terms heterogeneous
(McFetridge and Rafiquzzaman, 1986). The bulk of the literature, however, demonstrates
theoretically that uniform duration for all patents is inefficient because optimal patent
life is conditional (Berkowitz and Kotowitz, 1982; Nordhaus, 1969) and that patent terms
should be tailored to vary by industry (Gallini, 1992; Dore et al., 1993; Scherer, 1972;
Waterson, 1990; Wright, 1999) or per invention (Dasgupta and Stiglitz, 1980; DeBrock,
1985; Denicolò, 1996; Gilbert and Shapiro, 1990). The uniformity-cost perspective sug-
gests that future research on the patent system should analyse option pricing rather than
either assuming a uniform term or modeling per-invention variability (Scotchmer, 1999;
Duffy, 2003).
In copyright law, the uniformity-cost perspective reveals how recent changes that
eliminate or constrict real options have increased its social costs. Policymakers reduced
the effectiveness of the filtering function that the registration-and-notice requirements
played when the United States chose to adhere to the Berne Convention, which requires
that member states grant the entitlement itself rather than an option to acquire the entitle-
ment (17 U.S.C. § 102(a)). Current law, however, has not entirely abandoned real options.
Instead, authors of works in the United States must still register their respective claims
to copyright to enforce their rights in federal court (17 U.S.C. § 411(a)), and they have the
option to abandon protection through public dedication (Patry, 2016).
Matters are worse with respect to copyright duration. Until 1976, copyright law
divided duration into two terms, which served to vary the effective term of protection
because the renewal procedure acted as a real option similar to patent law’s maintenance
fees. The Copyright Act of 1976 removed this filter by adopting a life-plus-fifty term,
recently extended to life-plus-seventy (17 U.S.C. § 302). This change has rendered the
duration dimension of copyright law particularly insensitive to context, as was made
evident by the economists’ submissions to the Supreme Court in Eldred v. Ashcroft, 537
U.S. 186 (2003). The increase in social costs imposed by a substantively uniform term of
copyright protection led even Landes and Posner, who had once praised the life-plus-fifty
term as economically efficient (Landes and Posner, 1989), to call for reestablishing a real
option along copyright law’s duration dimension (Landes and Posner, 2003).
The full range of tailoring through real options as a policy tool to reduce uniformity
cost has not been explored in the literature. Although some commentators have wrestled
with the merits of registration versus examination procedures along the subject matter
dimension, and others have discussed the relative merits of renewable terms or main-
tenance fees along the duration dimension, few have discussed how the benefits of real
options could be realized along the scope dimension of copyright or patent entitlements.
Consider the scope of copyright law, for example. Copyright prohibits four kinds
from revenues associated with the marginal increase in portfolio value (Parchomovsky and Wagner,
2005). Portfolio value likely varies by industry, particularly because some individual pharmaceuti-
cal patents carry significant value (Allison et al., 2004).

of copying: (1) complete duplication; (2) partial duplication; (3) creation of a work
substantially similar to the whole; and (4) creation of a work with a substantially similar
part (Patry, 2016). At a minimum, a copyright owner should receive the right to prohibit
complete duplication as part of the initial grant if copyright is to serve as any kind of
solution to the appropriability problem. It would be possible to impose a call option
filter by conditioning the grant of the right to control the other three kinds of copying
on either payment of a modest fee and/or registration of the claim to control these uses.
Even though scope options potentially would reduce uniformity costs, the administrative
costs of implementation could be quite significant.
Bell and Parchomovsky (2014) have focused on the potential for scope options, arguing
that real options could completely mitigate uniformity cost. In their model, inventors
and authors would be required to register to opt in to patent or copyright protection.
The registrant would receive the current entitlements for a one-year period, after which
continued protection would require re-registration. At re-registration, the registrant
would be offered tailored packages of rights that would rely on current law but unbundle
the exclusive rights, types of infringers, period of protection, and remedies. The packages
would be paid for at the time of enforcement and priced at some percentage of the price of
infringement. In defending their claim, the authors rely upon, but partially misconstrue,
some of the uniformity-cost literature.4
In sum, real options are a useful mechanism for reducing uniformity cost, and there is
need for continued research into the performance of existing options as well as research
that would support greater use of options and option pricing as means of reducing
uniformity cost. Exploration of option designs that would affect the scope of intellectual
property rights seem to be particularly rich in possibility.
VI. TAILORING EXCLUSIVE RIGHTS
Explicitly tailoring intellectual property rights is a more direct strategy for minimizing
uniformity cost than relying on the application of standards or the filtering effects of
4
Bell and Parchomovsky (2014) position their model as a superior alternative to Burk and
Lemley’s (2009) argument for judicial tailoring in patent law or Carroll’s (2009) evidentiary
taxonomy for tailoring patent and copyright law. In this author’s view, Burk and Lemley do not
argue for judicial tailoring to the exclusion of other forms of uniformity-cost mitigation. This
author’s prior work is misconstrued in two ways. First, the economic justification for uniform
rights set forth in prior work and reprised above in this chapter is an attempt to extrapolate from
economic first principles the argument that would justify a one-size-fits-all policy response to a
problem known to vary significantly across and within industries. Although widely assumed in
the literature, the assumption has not been explicitly defended except in the work of Smith and
Long cited above. This is an attempt to give the best version of the explicit argument, but it is not
this author’s intent to persuade the reader with the argument. Second, identifying the economic
conditions favorable to tailoring, described in Carroll (2009), was part of a larger analysis of
existing policy strategies for mitigating uniformity cost. Bell and Parchomovsky treat the tailoring
framework in isolation without engaging with the discussion of this author’s endorsement of the
real option strategy—and the opportunity to explore scope options in particular—set forth in
prior work (Carroll, 2006).

real options. However, this strategy usually also requires a stronger justification than use
of flexible standards or real options because of the information asymmetries between
creators and policymakers. For this discussion, tailoring intellectual property rights
means varying the subject matter, scope, or duration of intellectual property rights for
some defined subject matter or for specific categories of creators (e.g., federally funded
researchers) or users (e.g., musicians making a sound recording of a musical composi-
tion). The policy tools to tailor rights include: (1) legislation; (2) judicial interpretation
and evidentiary presumptions; and (3) administrative rules, statutory interpretation, and
adjudication.
Tailoring proponents argue that the information-asymmetry objection is overstated
and that, in a dynamic system, rules change in response to lessons rulemakers learn about
the effects of prior decisions. The economic analysis of intellectual property law lacks a
model of a general information feedback framework for assessing and adjusting regula-
tion, by, for example, tailoring the law. But, even without such a model, a framework for
assessing when the government has sufficiently reliable information to justify tailoring can
be inferred from the work of a number of commentators.
All three forms of tailoring already are present in current law, but there is no general
theory of tailoring by which these measures can be assessed for their economic efficiency.
Unless one is committed to the view that all existing forms of tailoring must be the product
of successful rent-seeking, it is important to develop a framework for assessing the economic
effects of existing and proposed tailoring measures. Examples of existing legislative tailoring
measures include those that apply sui generis rights to certain subject matter (17 U.S.C.
§§ 901–14; Kastenmeier and Remington, 1985), as well as certain provisions of the Patent
Act applicable to the term of patent for certain pharmaceutical drugs (35 U.S.C. § 155), that
are aimed at overcoming differential treatment caused by regulatory approval processes
(Outlook, 2009) and making uniform the effective term of protection (Lunney, 2003).
In patent law, the Hatch-Waxman Act grants immunity for a generic drug manufac-
turer’s use of a patented invention to pursue regulatory approval for a drug to compete
with a patented drug six months prior to the patent’s expiration, making it likely the
most economically significant tailoring measure (35 U.S.C. § 271(e)). Also potentially
significant is the Bayh-Dole Act, which permits grantees to pursue patent protection for
inventions created with the support of federal funds but limits scope by providing the
government with ‘march-in’ rights (35 U.S.C. §§ 200–212). This tailored measure is specifi-
cally aimed at reducing uniformity cost. Federal grantees face differential appropriability
problems because the government has supplied both direct financial support and exclusive
rights to induce the investment. Congress also has tailored the scope of process patents for
medical method claims and business method claims in response to perceived uniformity
costs (35 U.S.C. §§ 273, 287(c)), and has indirectly eliminated patents on tax strategies
by declaring them to be in the prior art (Leahy-Smith America Invents Act, Pub. L. No.
112-29, 125 Stat. 284 § 14(a) (2011)).
In the Copyright Act, Congress has tailored the scope of protection, primarily by
replacing the right to exclude with statutory licenses for certain uses of certain classes
of works. Examples of these provisions include one that tailors rights in musical works
to permit garage bands and other musicians the right to record cover versions of their
favorite songs without the songwriter’s permission (17 U.S.C. § 115). Others tailor per-
formance rights to permit cable and satellite companies to retransmit network television

programming without prior consent of the copyright owners (17 U.S.C. §§ 111, 119, 122).
In addition, for certain authors, such as recognized visual artists, Congress has granted
additional rights (17 U.S.C. § 106A), while for other classes of authors, such as architects
(17 U.S.C. § 120) and authors of sound recordings (17 U.S.C. § 106(6)), Congress has
limited the exclusive rights available. Similarly, Congress has limited the scope of rights
in functional pictorial, graphical, or sculptural works (17 U.S.C. §§ 101, 113). Congress
also has tailored copyright scope by identifying certain privileged users, primarily persons
with visual disabilities, librarians, educators, and Internet service providers, who enjoy
certain additional limits on liability or available remedies (17 U.S.C. §§ 110, 121, 504(c),
512). Liu (2004) and Wu (2005) have remarked upon how underanalysed these features
of copyright law have been.
Judicial tailoring is a built-in feature of intellectual property law. For the purposes of
this chapter, judicial tailoring requires more systematic differentiation in the application
or interpretation of formally uniform rights. For example, copyright law treats books and
source code as literary works, which can be infringed by other works that are ‘substantially
similar’ where there is evidence of copying. But the scope of how much expression the
other work must borrow to fall within the zone of substantial similarity is noticeably
different between the two types of literary work. The effectiveness of judicial tailoring
for making intellectual property law more context-sensitive depends on the dimension of
rights being adapted.
Burk and Lemley (2009) argue in favor of judicial tailoring to reduce uniformity cost in
patent law, particularly when applied to software and biotechnology. They further assert
that the Federal Circuit already has applied the PHOSITA-based eligibility doctrines in
technology-specific fashion to software and biotechnology inventions (Burk and Lemley,
2004, 2009). They argue that the Federal Circuit has not explicitly chosen to tailor patent
law in this way, but that it should (Burk and Lemley, 2004, 2009). Some commentators
reinforce their argument (Laakmann, 2012), while others disagree with their reading of
the cases (Wagner, 2004). Others have extended the argument by identifying new flexible
doctrines, primarily fair use for patent law, that should be introduced to facilitate judicial
tailoring (O’Rourke, 2000; Strandburg, 2011), or how European patent law provides
analogous policy levers (van Overwalle, 2011).
Administrative tailoring has been implemented to a limited degree. Administrative
tailoring has greater potential effect in patent law because protection does not commence
until the USPTO has issued a patent, and tailoring can be accomplished during the
examination process. As with judicial tailoring, mere differential treatment—such as
the issuance of patents for obvious software inventions because of the absence of prior
art—does not amount to administrative policy to tailor the subject matter or scope of
protection to better balance incentives and access (Abramson, 2001). The USPTO also
arguably applies the Patent Act in tailored fashion. For example, evidence shows that
potential patentees in certain industries encounter more demanding prosecution than
others, and that this is a relatively recent development (Burk and Lemley, 2009). Indeed,
the USPTO’s examination guidelines for biotechnological inventions or business method
patents reflect a tailored interpretation of the requirements of patentability (Allison and
Tiller, 2003). Congress also responded to allegations of uniformity cost by delegating
authority to the USPTO to conduct post-grant review of ‘covered business methods’
(Leahy-Smith America Invents Act, Pub. L. No. 112-29, 125 Stat. 284, 18(a) (2011)).

In copyright law, Congress has delegated limited tailoring authority to the Copyright
Office (17 U.S.C. § 1201(a)(1)(B); 17 U.S.C. § 1201(a)(1)(C)). For example, the Copyright
Office’s determination that the deposit requirement for source code should be altered to
enable copyright owners to enjoy both copyright and trade secret protection (37 C.F.R.
§ 202.20(c)(2)(vii)) is a tailoring of copyright law’s disclosure function. The Copyright
Office has made the judgment that incentives are more important than access for software
and implemented that judgment within the discretion granted by the Copyright Act (17
U.S.C. § 408(c)). Some argue that expanding an agency’s power to tailor copyright law
would outperform legislative or judicial tailoring in some contexts (Greenberg, 2016).
The fact that patent and copyright law have been tailored by all three branches of
government suggests that the case for uniform intellectual property rights has some weak-
nesses. The cause for tailoring by policymakers must be either a response to rent-seeking
by special interest groups, an ignorant but well-intentioned response to perceived uni-
formity cost, or an informed response to real and substantial uniformity cost. The effects
of these measures must be that they: (1) increase the social costs of intellectual property
rights by harming incentives, reducing access, or imposing additional administrative
cost with no offsetting benefit; (2) do not affect the social costs of intellectual property
rights because the tailored rights are of relatively little economic significance or because
the degree of tailoring is minor enough to be immeasurable; or (3) make the intellectual
property strategy more efficient by establishing a better fit between the appropriability
problem and its solution either by accident or by design.
The case for uniform intellectual property rights set forth above predicts that the motive
to tailor is most likely rent-seeking, and that the likely effects have been either to make
matters worse or to be relatively meaningless. But, perhaps, policymakers in the dynamic
intellectual property system have made the law more efficient by responding to evidence of
uniformity cost. Analysts currently lack a framework for assessing existing and proposed
tailoring measures to ascertain their effects.
VII. A FRAMEWORK FOR TAILORING
From a range of commentary, one can glean a set of economic conditions that favor
tailoring intellectual property rights. Most commentary focuses on the mismatch between
the baseline uniform rights and the appropriability problem in different settings. These
conditions include:
A. Alternative Incentives to Create or Innovate and Distribute
The literature demonstrates that innovative activity has many complicated motivations,
and society may receive the benefits of certain forms of innovation even without extend-
ing rights sufficient to induce a rational, selfish actor to innovate (Amabile, 1996; Cohen,
J., 2007; Frey and Jegan, 2001). Moreover, even when one holds firm to the rational actor
thesis, in some cases anticipated prestige, notoriety, or other ‘nonpecuniary income’ would
serve as a sufficient return on the investment to induce initial production in the absence
of copyright or patent (Burk and Lemley, 2009; Oliar and Sprigman, 2008; Raustalia
and Sprigman, 2006; Rosenblatt, 2011; Shavell and van Ypersele, 2001). Alternatively,

the investment in initial production may serve as a loss leader to increase other revenue
streams, such as speaker’s fees (Landes and Posner, 1989). Finally, firms in a competitive
economy are under constant pressure to innovate to differentiate their products and
services from their competitors even without the promise of exclusive rights (Baker, 2007).
In copyright, for example, scholars and researchers do not receive royalties for their
journal articles, and it is likely that they would continue to research and to write even
without copyright in their articles because they receive direct compensation to do research
and there are a variety of indirect benefits that flow from publication (Tuckman and
Leahy, 1975; Shavell and van Ypersele, 2001). In patent law, there is a real question about
whether the extension of patent rights to business methods was necessary on incentive
grounds (Meurer, 2002; Samuelson and Schultz, 2011). Software is protected by patents,
copyrights, and trade secrets. It is a favored target of tailoring proposals on multiple
grounds, including the alternative incentive ground (Bitzer et al., 2007; Lerner and Tirole,
2002; Menell, 1986, 1989, 1994; Samuelson et al., 1994; Samuelson, 2011).
B. Alternative Appropriability Mechanisms
Even when innovators and their investors require appropriability, the availability of
alternative mechanisms to intellectual property rights may justify tailoring these rights. It
is likely that tailoring will have to be done on an industry-specific or technology-specific
basis (Levin et al., 1987). One alternative that commentators draw attention to is direct
cost subsidies for innovators. Government funding is the most common form of these. In
the United States, the federal government funds approximately 25 percent of research and
development, and in OECD countries, public sector investment approaches 50 percent
(Menell and Scotchmer, 2007). Tailoring protection for intellectual property arising from
public investment is likely to be a reasonable response (Rai and Eisenberg, 2003; Golden,
2001). The Bayh-Dole amendments to the Patent Act (35 U.S.C. §§ 200–212) are the clear-
est signal that policymakers are at least nominally responsive to evidence-based arguments
concerning the effects of direct subsidies. Responding to arguments that commercializers
lacked sufficient incentives to build on unpatented discoveries made by federal grantees,
Congress made clear that these grantees could seek and receive patent protection subject to
some tailoring of scope to reduce the social costs of monopoly under particular conditions.
In response to these developments, the patent literature has seen the emergence of two
related cottage industries. One is specifically focused on proposals to revise the Bayh-Dole
Act (de Larena, 2007; Eisenberg, 1996; Mireles, 2006; Rai and Eisenberg, 2003; Kieff,
2001; Mowery et al., 2004; Rai, 1999; Strandburg, 2005), and the other embraces a range
of related proposals to tailor patent law for universities (Bagley, 2006; Bouchard, 2007;
Lemley, 2008; Osenga, 2007; Rowe, 2006). The argument in both cases is that the fact
of direct subsidies is the kind of evidence likely to give rise to substantial uniformity
cost because baseline patent rights are premised on an assumption that innovators need
to recoup most of their costs in the market through the promise of monopoly pricing
(Menell and Scotchmer, 2007).
The other form of direct subsidy that should be squared with uniform intellectual
property rights is prizes and rewards. As discussed above, direct compensation in the
form of prizes or rewards is a well-established policy option. Many of the modern
implementations of this option, however, use it as a supplement rather than a substitute

for intellectual property protection (National Research Council. 2007. Innovation Prizes
at the National Science Foundation).
Commentators also identify certain market-specific economic features, primarily lead-
time advantage and network effects, that can supply a sufficient degree of appropriability
to justify tailoring intellectual property rights. Research indicates that the value of lead
time often is industry-specific (Lieberman and Montgomery, 1998; Robinson et al., 1994),
and that in product markets with patentable goods incumbents often enjoy significant
market share advantages even after competitors have entered a market (Barnett, 2004).
The value of the lead-time advantage is affected not only by its duration but also by
competitors’ copying costs. One study shows that imitation costs for patentable goods can
run at about 65 percent of the costs of innovation (Mansfield et al., 1981). However, about
70 percent of the goods studied were patented and ‘imitation costs’ included the costs of
inventing around the patent. Consequently, this data does not translate immediately into
the costs of competition in markets without intellectual property rights. Additionally,
when one accounts for the monetary value of time, reflected as a competitor’s opportunity
costs, the necessary level of protection would be further reduced. If expected profits
derived from this lead-time or first-mover advantage are sufficient to recoup the costs of
initial production, the case for government intervention largely disappears (Reichman,
1994). The best case for tailoring based on lead-time profits would be one in which intel-
lectual property rights should be treated as superfluous on both sides of the ledger, but
rightsholders use the rights to engage in strategic litigation that squeezes additional rents
from competitors and consumers with no offsetting social benefits (Bessen and Meurer,
2009).
Strong network effects can also serve as an alternative source of appropriability.5
Economists distinguish between markets involving actual networks, virtual networks, and
positive feedback effects, and these distinctions could impact the strength of a tailoring
proposal (Lemley and McGowan, 1998).
Some argue that evidence of network externalities may support a shift from a property
to a liability rule to provide a form of rate-of-return regulation of a ‘natural’ monopoly
(Lemley and Weiser, 2007). Unauthorized copying can serve to strengthen the market
share of an information provider in a ‘tippy’ market (Shapiro and Varian, 1999; Takeyama,
1994). Even where network effects are not strong enough to induce a desired level of
investment in information production, network effects can amplify the market power
that exclusive rights can confer. Lemley and McGowan (1998) suggest that evidence of
a market exhibiting strong network effects because of demand for standardization may
support proposals to permit reverse engineering to allow competitive entry. Patent law
currently does not recognize a reverse engineering defense, although some scholars have
offered proposals either for a generalized fair use defense that could be adapted to market
5
Network effects, as manifested by the ‘superstar’ effect, is a version of the phenomenon.
Although Rosen (1981) explains the skew distribution of popularity in the cultural sphere as
reflecting a skewed distribution of talent in the population, a more convincing account would focus
on the signaling function that certain forms of consumption play. Once momentum builds behind
a particular book, movie, song, entertainer, athlete, or fashion design, consumers’ purchasing deci-
sions will be influenced more by the importance of signaling membership in the herd than by any
subjective evaluation of the good’s quality (Frank and Cook, 1995).

circumstances or more tailored proposals to permit reverse engineering (O’Rourke, 2000;

Samuelson and Scotchmer, 2002; Strandburg, 2011).
C. Overlapping Rules and Rights
Two common legal features that intellectual property owners factor into their appropri-
ability calculus is the availability of other legal rights of exclusion, such as trade secret,
trademark, or contract law, and regulatory requirements in the marketing and distribution
of creations or innovations. Product differentiation strategies supported by trademark law
supply an important source of power over price. Some scholars argue that the effects of
trademark and trade secret protection may be sufficient to induce the desired level of
investment even in the absence of copyright or patent rights in some cases (Kieff, 2006).
Moreover, the exclusivity provided by copyright or patent rights facilitates the producer’s
ability to establish strong, highly distinctive marks (Parchomovsky and Siegelman, 2002).
In markets in which this effect is particularly strong, the level of protection may be
reduced by, for example, reducing the term of protection without significantly reducing
the incentive effects the protection supplies.
D. Demand-Side Features—Positive Spillovers
Some argue that efficient patent and copyright law must be limited and leaky to encourage
or allow certain types of uncompensated demand-side sharing of valuable information
(Frischmann, 2005, 2009; Frischmann and Lemley, 2007; Lohr, 2003). Their arguments
call for tailoring to ensure that the law is particularly permissive with respect to patented
or copyrighted information that functions as ‘infrastructure.’ In particular, Frischmann
(2012) argues that information should be managed as a commons rather than through
private, exclusive rights when: (1) the resource may be consumed nonrivalrously; (2)
social demand for the resource is driven primarily by downstream productive activity that
requires the resource as an input; and (3) the resource may be used as an input into a wide
range of goods and services, including private goods, public goods, and nonmarket goods.
But some information that would fall within the subject matters of patent and
copyright, respectively, also may function as nonrivalrous, generic inputs that supply
social and public goods, such as public health or public education, for which markets
are either absent or incomplete. In such cases, the opportunity cost of an exclusive right
may be greater than its benefit (Lessig, 2005). In these cases, the argument for solving the
appropriability problem through some combination of prize, reward, or tax strategy is
likely to be particularly strong. Alternatively, a tailoring proponent may seek to propose
limitations or exceptions for the producers of public goods while leaving exclusive rights
intact as against the producers of private goods. Much of the recent commentary calling
for a revitalized experimental use exception to patent infringement or for tailoring of
patent law with respect to university researchers follows this line (Hagelin, 2006). The
proponent should also be prepared to answer the argument advanced by Wagner that even
‘complete’ rights of exclusion are unable to prevent positive spillovers, and—to the extent
that more robust, uniform rights encourage production of additional information—the
total amount of spillovers will increase with broad exclusive rights (Wagner, 2003).

VIII. INDUSTRIAL INNOVATION DECISION STRUCTURES
Economic analysis to date suggests that the magnitude of appropriability problems

varies by industry. Merges and Nelson (1990) and Wu (2006) independently suggest that
policymakers should recognize the effect on industry structure and degree of competitive
entry influenced by the subject matter and scope of intellectual property rights and
that policymakers should tailor rights to modulate the degree of entry depending upon
industry maturity and other competitive conditions.
Merges and Nelson (1990) made a significant contribution to the patent literature first
by calling attention to the role courts must play in tailoring the scope of patents through
application of the law’s flexible scope doctrines and second by ‘show[ing] that the issues
at stake regarding patent scope depend on the nature of technology in an industry. This
dependence includes two characteristics: the relationship between technical advances in
the industry, and the extent to which firms license technologies to each other.’
By studying and categorizing the effects of patent scope on follow-on invention,
Merges and Nelson (1990) generally reject the ‘prospect’ theory of patent scope that
would delegate control over follow-on innovation to early inventors in favor of greater
entry tailored to the characteristics of what they label ‘cumulative technologies,’ ‘chemical
industries,’ and ‘science-based industries.’
Wu (2006) makes a related argument concerning the role that the presence or absence
of intellectual property rights and the delineation of their scope will shape the ‘decision
architecture’ for innovation within particular industries. In short, he argues that where
intellectual property rights are robust, innovation decisions are likely to be made within
hierarchal firms that own these rights, and the willingness of these firms to grant licenses
to follow-on innovators who may become competitors is suspect. In contrast, where
intellectual property rights are subject to significant limitations or exceptions, innovation
decisions are likely to be made polyarchically. Wu (2006) argues that policymakers should
employ presumptions that favor limited intellectual property rights in new industries to
favor decentralized development unless the risk of misappropriation is so significant that
investments in new development will be deterred. In contrast, he argues that policymakers
should be more solicitous of claims for more robust rights applicable to ‘dead’ industries
unless overpropertization was one of the causes of death (Wu, 2006).
Others argue that tailoring should increase available patent rights when lead time is
insufficient to encourage market experimentation (Abramowicz and Duffy, 2008) or when
the costs of commercialization justify additional protection (Sichelman, 2010).
IX. CONCLUSION
Decisions about intellectual property policy should be evidence-based. Posner is correct

to say that gathering and assessing evidence about the performance of uniform patents
and copyrights is difficult. But gathering and assessing evidence about the social costs of
uniform rights and how these could be tailored to perform better is far more plausible and
effective. This should be a primary focus for the economic analysis of intellectual property.
While the general problem of uniformity cost has been recognized relatively recently,
early analysis by Plant (1934) and Breyer (1970) in copyright law laid the groundwork

for the above analysis. In sum, uniformity cost can be reduced through private ordering,
entitlement design, and the use of real options in place of direct entitlement grants. The
types of evidence most likely to support a tailoring proposal are those showing innovator
incentives that depart from the standard rationale for intellectual property rights, the
availability of alternative appropriability mechanisms, the role of overlapping legal rights
that influence appropriability, the magnitude of positive spillovers generated by certain
types of innovative activity, and industry-specific effects on follow-on innovation from
uniform rights.
REFERENCES
Abramowicz, Michael. 2004. ‘An Industrial Organization Approach to Copyright Law,’ 46 William and Mary
Abramowicz, Michael. 2005. ‘A Theory of Copyright’s Derivative Right and Related Doctrines,’ 90 Minnesota
Abramowicz, Michael. 2011. ‘A New Uneasy Case for Copyright,’ 79 George Washington Law Review 1644–711.
Abramowicz, Michael, and John F. Duffy. 2008. ‘Intellectual Property for Market Experimentation,’ 83 New
York University Law Review 337–410.
Abramson, Bruce. 2001. ‘Promoting Innovation in the Software Industry: A First Principles Approach to
Intellectual Property Reform,’ 8 Boston University Journal of Science & Technology 75–156.
Adams, Wendy A. 2002. ‘Intellectual Property Infringement in Global Networks: The Implication of Protection
Ahead of the Curve,’ 10 International Journal of Law and Information Technology 71–131.
Allison, John R., and Emerson H. Tiller. 2003. ‘The Business Method Patent Myth,’ 18 Berkeley Technology
Allison, John R., Mark A. Lemley, Kimberly A. Moore, and R. Derek Trunkey. 2004. ‘Valuable Patents,’ 92
Georgetown Law Journal 435–79.
Amabile, Teresa M. 1996. Creativity in Context: The Social Psychology of Creativity. Boulder, CO: Westview Press.
Arrow, Kenneth J. 1962. ‘Economic Welfare and Allocation of Resources for Invention,’ in National Bureau of
Economic Research, ed., The Rate and Direction of Inventive Activity: Economic and Social Factors. Princeton,
NJ: Princeton University Press, 609–26.
Ayres, Ian. 2005. Optional Law: The Structure of Legal Entitlements. Chicago, IL: University of Chicago Press.
Bagley, Margo A. 2006. ‘Academic Discourse and Proprietary Rights: Putting Patents in Their Proper Place,’
47 Boston College Law Review 217–74.
Baker, Jonathan B. 2007. ‘Beyond Schumpeter vs. Arrow: How Antitrust Fosters Innovation,’ 74 Antitrust Law
Journal 575–602.
Barnett, Jonathan M. 2004. ‘Private Protection of Patentable Goods,’ 25 Cardozo Law Review 1251–313.
Yale Law Journal 456.
Bell, Abraham, and Gideon Parchomovsky. 2014. ‘Reinventing Copyright and Patent,’ 113 Michigan Law
Review 231–78.
Berkowitz, Michael K., and Yehuda Kotowitz. 1982. ‘Patent Policy in an Open Economy,’ 15 Canadian Journal
Bessen, James E., and Michael J. Meurer. 2009. Patent Failure: How Judges, Bureaucrats, and Lawyers Put
Bhaskar, V., and Ted To. 2004. ‘Is Perfect Price Discrimination Really Efficient? An Analysis of Free Entry,’ 35
The RAND Journal of Economics 775–6.
Bitzer, Jürgen, Wolfram Schrettl, and Philipp J.H. Schröder. 2007. ‘Intrinsic Motivation in Open Source
Software Development,’ 35 Journal of Comparative Economics 160–69.
Boldrin, Michele, and David Levine. 2008. Against Intellectual Monopoly. New York, NY: Cambridge
University Press.
Bouchard, Ron A. 2007. ‘Balancing Public and Private Interests in the Commercialization of Publicly Funded
Medical Research: Is There a Role for Compulsory Government Royalty Fees?,’ 13 Boston University Journal
of Science and Technology Law 120–91.
Bracha, Oren, and Talha Syed. 2014. ‘Beyond the Incentive-Access Paradigm? Product Differentiation &

Braithwaite, John, and Peter Drahos. 2007. Information Feudalism: Who Owns the Knowledge Economy? New
York City, NY: The New Press.
Burk, Dan L., and Mark A. Lemley. 2002. ‘Is Patent Law Technology-Specific?,’ 17 Berkeley Technology Law
Journal 1155–1206.
Burk, Dan L., and Mark A. Lemley. 2004. ‘Biotechnology’s Uncertainty Principle,’ 54 Case Western Law Review
691–742.
Burk, Dan L., and Mark A. Lemley. 2009. The Patent Crisis and How the Courts Can Solve It. Chicago, IL:
Burk, Dan L., and Brett H. McDonnell. 2007. ‘The Goldilocks Hypothesis: Balancing Intellectual Property
Rights at the Boundary of the Firm,’ 2007 University of Illinois Law Review 575–636.
Burstein, Aaron, Will Thomas DeVries, and Peter S. Menell. 2004. ‘The Rise of Internet Interest Group Politics,’
Burstein, Michael J. 2012. ‘Exchanging Information Without Intellectual Property,’ 91 Texas Law Review
227–82.
Carrier, Michael A. 2007. ‘Why Modularity Does Not (and Should Not) Explain Intellectual Property,’ 117 Yale
Law Journal, Pocket Part 95–100.
Carroll, Michael W. 2006. ‘One for All: The Problem of Uniformity Cost in Intellectual Property Law,’ American
Carroll, Michael W. 2007. ‘Patent Injunctions and the Problem of Uniformity Cost,’ 13 Michigan
Telecommunications and Technology Law Review 421–43.
Carroll, Michael W. 2009. ‘One Size Does Not Fit All: A Framework for Tailoring Intellectual Property Rights,’
Ohio State Law Journal 1361–434.
Coase, R.H. 1960. ‘The Problem of Social Cost,’ 3 The Journal of Law and Economics 1–44.
Cohen, Julie E. 2007. ‘Creativity and Culture in Copyright Theory,’ 40 U.C. Davis Law Review 1151–205.
Cohen, Julie E., and Mark A. Lemley. 2001. ‘Patent Scope and Innovation in the Software Industry,’ 89
Cohen, Wesley M., Richard R. Nelson, and John P. Walsh. 2000. ‘Protecting Their Intellectual Assets:
Appropriability Conditions and Why U.S. Manufacturing Firms Patent (or Not),’ National Bureau of
Economic Research, Working Paper No. 7552, accessed March 20, 2019 at http://www.nber.org/papers/
w7552.
Cooper, Richard M. 1993. ‘Legislative Patent Extensions,’ 48 Food & Drug Law Journal 59–85.
Cornelli, Francesca, and Mark Schankerman. 1999. ‘Patent Renewals and R&D Incentives,’ 30 The RAND
Correa, Carlos M. 2006. ‘Implications of Bilateral Free Trade Agreements on Access to Medicines,’ 84 Bulletin
of the World Health Organization 399–404.
Cotropia, Christopher A. 2005. ‘“After-Arising,” Technologies and Tailoring Patent Scope,’ 14 Annual Survey
of American Law 151–201.
Cotropia, Christopher A. 2009. ‘Describing Patents as Real Options,’ 34 Journal of Corporation Law 1127–49.
Cotropia, Christopher A., and James Gibson. 2010. ‘The Upside of Intellectual Property’s Downside,’ 57 UCLA
Dasgupta, Partha, and Joseph Stiglitz. 1980. ‘Uncertainty, Industrial Structure, and the Speed of R&D,’ 11 The
Bell Journal of Economics and Management Science 1–28.
DeBrock, Lawrence M. 1985. ‘Market Structure, Innovation, and Optimal Patent Life,’ 28 The Journal of Law
and Economics 223–44.
de Larena, Lorelei Ritchie. 2007. ‘The Price of Progress: Are Universities Adding to the Cost?,’ 43 Houston Law
Review 1373–444.
Demsetz, Harold. 1970. ‘The Private Production of Public Goods,’ 13 The Journal of Law and Economics
293–306.
Denicolò, Vincenzo. 1996. ‘Patent Races and Optimal Patent Breadth and Length,’ 44 The Journal of Industrial
Economics 249–65.
Denton, F. Russell, and Paul J. Heald. 2003. ‘Random Walks, Non-Cooperative Games, and the Complex
Mathematics of Patent Pricing,’ 55 Rutgers Law Review 1175–288.
Depoorter, Ben, and Robert Kirk Walker. 2013. ‘Copyright False Positives,’ 89 Notre Dame Law Review 319–64.
DeVany, Arthur S., and W. David Walls. 2004. ‘Motion Picture Profit, the Stable Paretian Hypothesis, and the
Curse of the Superstar,’ 28 Journal of Economic Dynamics and Control 1035–57.
DiMasi, Joseph A., Henry G. Grabowski, and Ronald W. Hansen. 2016. ‘Innovation in the Pharmaceutical
Industry: New Estimates of R&D costs,’ 47 Journal of Health Economics 20–33.

Dinwoodie, Graeme B. 2010. ‘“One Size Fits All”: Consolidation and Difference in Intellectual Property Law,’
in Annette Kur, ed., Horizontal Issues in Intellectual Property Law, Uncovering the Matrix. Cheltenham, UK
Dore, M.H.I., J. Kushner, and I. Masse. 1993. ‘The Optimal Length of a Patent with Variable Output Elasticity
and Returns to Scale in R&D,’ 21 Atlantic Economic Journal 10–26.
Duffy, H. 2002. ‘Harmony and Diversity in Global Patent Law,’ 17 Berkeley Technology Law Journal 685–726.
Duffy, John F. 2003. ‘A Minimum Optimal Patent Term,’ Working Paper, accessed March 20, 2019 at http://
ssrn.com/abstract=354282.
Duffy, John F. 2005. ‘Intellectual Property Isolationism and the Average Cost Thesis,’ 83 Texas Law Review
1077–95.
Edlin, Aaron S., Mario Epelbaum, and Walter P. Heller. 1998. ‘Is Perfect Price Discrimination Really Efficient?
Welfare and Existence in General Equilibrium,’ 66 Econometrica 897–922.
Eisenberg, Rebecca S. 1996. ‘Public Research and Private Development: Patents and Technology Transfer in
Government-Sponsored Research,’ 82 Virginia Law Review 1663–727.
Eisenberg, Rebecca S. 2000. ‘The Promise and Perils of Strategic Publication to Create Prior Art: A Response
to Professor Parchomovsky,’ 98 Michigan Law Review 2358–70.
Eisenberg, Rebecca S. 2004. ‘Obvious to Whom? Evaluating Inventions from the Perspective of PHOSITA,’ 19
Farber, Daniel A., and Brett M. McDonnell. 2003. ‘Why (and How) Fairness Matters at the IP/Antitrust
Interface,’ 87 Minnesota Law Review 1817–70.
Fennell, Lee Anne. 2005. ‘Revealing Options,’ 118 Harvard Law Review 1399–488.
Fisher, William W. III. 1998. ‘Property and Contract on the Internet,’ 73 Chicago-Kent Law Review 1203–56.
Fisher, William W. III. 2004. Promises to Keep: Technology, Law, and the Future of Entertainment. Stanford,
Fisher, William, and Talha Syed. 2006. ‘Global Justice in Health Care: Developing Drugs for the Developing
World,’ 40 U.C. Davis Law Review 581–678.
Frank, Robert H., and Phillip J. Cook. 1995. The Winner-Take-All Society: Why the Few at the Top Get So Much
More than the Rest of Us. New York, NY: Penguin Books.
Frey, Bruno S., and Reto Jegan. 2001. ‘Motivation Crowding Theory,’ 15 Journal of Economic Surveys 589–611.
Frischmann, Brett M. 2000. ‘Innovation and Institutions: Rethinking the Economics of U.S. Science and
Technology Policy,’ 24 Vermont Law Review 347–416.
Frischmann, Brett. 2001. ‘Privatization and Commercialization of the Internet Infrastructure: Rethinking
Market Intervention into Government and Government Intervention into Market,’ 2 Columbia Science and
Frischmann, Brett M. 2005. ‘An Economic Theory of Infrastructure and Commons Management,’ 89
Minnesota Law Review 917–1030.
Frischmann, Brett. 2009. ‘Spillovers Theory and Its Conceptual Boundaries,’ 51 William and Mary Law Review
801–24.
Frischmann, Brett. 2012. Infrastructure: The Social Value of Shared Resources. Oxford: Oxford University
Press.
Frischmann, Brett M., and Mark A. Lemley. 2007. ‘Spillovers,’ 107 Columbia Law Review 257–301.
Gallini, Nancy T. 1992. ‘Patent Policy and Costly Imitation,’ 23 The RAND Journal of Economics 52–63.
Ghosh, Shubha. 2004. ‘Patents and the Regulatory State: Rethinking the Patent Bargain Metaphor After
Eldred,’ 19 Berkeley Technology Law Journal 1315–88.
Ghosh, Shubha, and Jay Kesan. 2004. ‘What Do Patents Purchase? In Search of Optimal Ignorance in the Patent
Office,’ 40 Houston Law Review 1219–64.
Gilbert, Richard, and Carl Shapiro. 1990. ‘Optimal Patent Length and Breadth,’ 21 The RAND Journal of
Economics 106–12.
Golden, John M. 2001. ‘Biotechnology, Technology Policy, and Patentability: Natural Products and Invention
in the American System,’ 50 Emory Law Journal 101–91.
Greenberg, Brad. 2016. ‘Rethinking Technology Neutrality,’ 100 Minnesota Law Review 1495–562.
Hagelin, Ted. 2006. ‘The Experimental Use Exemption to Patent Infringement: Information on Ice, Competition
on Hold,’ 58 Florida Law Review 483–560.
Hall, Bronwyn H., and Rosemarie Ziedonis. 2001. ‘The Patent Paradox Revisited: An Empirical Study of
Patenting in the U.S. Semiconductor Industry,’ 32 The RAND Journal of Economics 103–28.
Hardy, Trotter. 2001. ‘The Copyrightability of New Works of Authorship: “XML Schemas” as an Example,’ 38
Hemel, Daniel J., and Lisa Larrimore Ouellette. 2013. ‘Beyond the Patents-Prizes Debate,’ 92 Texas Law Review
304–82.
Hughes, Justin. 2003. ‘Fair Use across Time,’ 50 UCLA Law Review 775–804.
Janis, Mark D. 2001. ‘Patent Abolitionism,’ 17 Berkeley Technology Law Journal 899–952.

Joyce, Craig, Tyler Ochoa, Michael Carroll, Marshall Leaffer, and Peter Jaszi. 2016. Copyright Law (10th ed.).
Durham, NC: Carolina Academic Press.
Kaplow, Louis. 1984. ‘The Patent-Antitrust Intersection: A Reappraisal,’ 97 Harvard Law Review 1813–92.
Kaplow, Louis. 1992. ‘Rules versus Standards: An Economic Analysis,’ 42 Duke Law Journal 557–629.
Karjala, Dennis S. 2003. ‘Distinguishing Patent and Copyright Subject Matter,’ 35 Connecticut Law Review
439–524.
Kastenmeier, Robert W., and Michael J. Remington. 1985. ‘The Semiconductor Chip Protection Act of 1984: A
Swamp or Firm Ground?,’ 70 Minnesota Law Review 417–70.
Kelman, Mark A. 1987. Guide to Critical Legal Studies. Boston, MA: Harvard University Press.
Kennedy, Duncan. 1976. ‘Form and Substance in Private Law Adjudication,’ 89 Harvard Law Review 1685–778.
Kesan, Jay P., and Andres A. Gallo. 2006. ‘Why “Bad” Patents Survive in the Market and How Should We
Change? The Private and Social Costs of Patents,’ 55 Emory Law Journal 61–140.
Kesan, Jay P., and Andres A. Gallo. 2009. ‘The Political Economy of the Patent System,’ 87 North Carolina Law
Review 1341–419.
Kieff, F. Scott. 2001. ‘Facilitating Scientific Research: Intellectual Property Rights and the Norms of Science:
A Response to Rai and Eisenberg,’ 95 Northwestern University Law Review 691–705.
Kieff, F. Scott. 2003. ‘The Case for Registering Patents and the Law and Economics of Present Patent-Obtaining
Rules,’ 45 Boston College Law Review 55–123.
Kieff, F. Scott. 2006. ‘Coordination, Property and Intellectual Property: An Unconventional Approach to
Anticompetitive Effects and Downstream Access,’ 56 Emory Law Journal 327–438.
Kitch, Edmund W. 2000. ‘Elementary and Persistent Errors in the Economic Analysis of Intellectual Property,’
53 Vanderbilt Law Review 1727–41.
Knight, Frank H. 1921. Risk, Uncertainty and Profit. Ithaca, NY: Cornell University Library.
Kremer, Michael. 1998. ‘Patent Buyouts: A Mechanism for Encouraging Innovation,’ 113 Quarterly Journal of
Laakmann, Anna B. 2012. ‘An Explicit Policy Lever for Patent Scope,’ 19 Michigan Telecommunications and
Landes, William M., and Richard A. Posner. 1989. ‘An Economic Analysis of Copyright Law,’ 18 The Journal
of Legal Studies 325–63.
Landes, William M., and Richard A. Posner. 2003. The Economic Structure of Intellectual Property. Cambridge,
MA: Belknap Press.
Lemley, Mark A. 2005. ‘What’s Different about Intellectual Property?,’ 83 Texas Law Review 1097–104.
Lemley, Mark A. 2008. ‘Are Universities Patent Trolls?,’ 18 Fordham Intellectual Property Media & Entertainment
Lemley, Mark A., and David McGowan. 1998. ‘Legal Implications of Network Economic Effects,’ 86 California
Lemley, Mark A., and Bhaven Sampat. 2008. ‘Is the Patent Office a Rubber Stamp?,’ 58 Emory Law Journal
181–206.
Lemley, Mark A., and Philip J. Weiser. 2007. ‘Should Property or Liability Rules Govern Information?,’ 85
Texas Law Review 783–841.
Lemley, Mark A., Douglas Lichtman, and Bhaven N. Sampat. 2005. ‘What To Do About Bad Patents,’ 28
Regulation 10–13.
Lerner, Joshua, and Jean Tirole. 2002. ‘Some Simple Economics of Open Source,’ 50 The Journal of Industrial
Lessig, Lawrence. 2001. The Future of Ideas: The Fate of the Commons in a Networked World. New York, NY:
Vintage Books.
Lessig, Lawrence. 2005. ‘Re-Marking the Progress in Frischmann,’ 89 Minnesota Law Review 1031–43.
Levin, Richard C., Alvin K. Klevorick, Richard R. Nelson, and Sidney G. Winter. 1987. ‘Appropriating the
Returns from Industrial Research and Development,’ 1987 Brookings Papers on Economic Activity 783–831.
Lichtman, Douglas Gary. 1997. ‘Pricing Prozac: Why the Government Should Subsidize the Purchase of
Patented Pharmaceuticals,’ 11 Harvard Journal of Law and Technology 123–39.
Lieberman, Marvin B., and David B. Montgomery. 1998. ‘First-Mover (Dis)Advantages: Retrospective and
Link with the Resource-Based View,’ 19 Strategic Management Journal 1111–25.
Litman, Jessica. 2001. Digital Copyright: Protecting Intellectual Property on the Internet. Amherst, NY:
Prometheus Books.
Liu, Joseph P. 2002. ‘Copyright and Time: A Proposal,’ 101 Michigan Law Review 409–81.
Liu, Joseph P. 2004. ‘Regulatory Copyright,’ 83 North Carolina Law Review 87–166.
Lohr, Steve. 2003. ‘Steal This Book? A Publisher Is Making It Easy,’ The New York Times C4 (Jan. 13, 2003),
accessed March 20, 2019 at http://www.nytimes.com/2003/01/13/business/technology-steal-this-book-a-publi
sher-is-making-it-easy.html.
Long, Clarisa. 2000. ‘Proprietary Rights and Why Initial Allocations Matter,’ 49 Emory Law Journal 823–36.

Love, Brian J. 2018. ‘Patent Duration,’ in Peter S. Menell and David L. Schwartz, eds., Research Handbook
on the Economics of Intellectual Property, vol. II. Cheltenham, UK and Northampton, MA, USA: Edward
Elgar Publishing.
Love, James, and Tim Hubbard. 2007. ‘The Big Idea: Prizes to Stimulate R&D for New Medicines,’ 82 Chicago-
Kent Law Review 1519–54.
Lunney, Glynn S. Jr. 1996. ‘Reexamining Copyright’s Incentives-Access Paradigm,’ 49 Vanderbilt Law Review 557.
Lunney, Glynn S. Jr. 2003. ‘Patent Law, the Federal Circuit, and the Supreme Court: A Quiet Revolution,’ 11
Supreme Court Economic Review 1–80.
Lunney, Glynn S. Jr. 2008. ‘Copyright’s Price Discrimination Panacea,’ 21 Harvard Journal of Law & Technology
387–456.
Lunney, Glynn S. Jr. 2009. ‘Copyright, Private Copying, and Discrete Public Goods,’ 12 Tulane Journal of
Technology & Intellectual Property 1–34.
Macaulay, Thomas Babington. 1900. ‘Speech in the House of Commons,’ in Speeches and Legal Studies: The
Complete Works of Thomas Babington Macaulay. Boston, MA: Houghton, Mifflin, Harcourt.
Machlup, Fritz. 1958. ‘An Economic Review of the Patent System,’ Study No. 15, The Subcommittee on Patents,
Trademarks and Copyrights, Committee on the Judiciary, U.S. House of Representatives, 85th Congress, 2d
Session.
Machlup, Fritz, and Edith Penrose. 1950. ‘The Patent Controversy in the Nineteenth Century,’ 10 The Journal
of Economic History 1–29.
Mansfield, Edwin, John Rapoport, Anthony Romeo, Samuel Wagner, and George Beardsley. 1977. ‘Social and
Private Rates of Return from Industrial Innovations,’ 91 Quarterly Journal of Economics 221–40.
Mansfield, Edwin, Mark Schwartz, and Samuel Wagner. 1981. ‘Imitation Costs and Patents: An Empirical
Study,’ 91 Economic Journal 907–18.
McFetridge, D.G., and M. Rafiquzzaman. 1986. ‘The Scope and Duration of the Patent Right and the Nature
of Research Rivalry,’ 8 Research in Law and Economics 91–129.
Menell, Peter S. 1989. ‘An Analysis of the Scope of Copyright Protection for Application Programs,’ 45 Stanford
Menell, Peter S. 1994. ‘The Challenges of Reforming Intellectual Property Protection for Computer Software,’
94 Columbia Law Review 2644–54.
Menell, Peter S., and Suzanne Scotchmer. 2007. ‘Intellectual Property,’ in A. Mitchell Polinsky and Steven
Shavell, eds., Handbook of Law and Economics. Oxford: North-Holland.
Merges, Robert P. 2000. ‘One Hundred Years of Solicitude: Intellectual Property Law, 1900–2000,’ 88 California
Merges, Robert P., and Richard Nelson. 1990. ‘On the Complex Economics of Patent Scope,’ 90 Columbia Law
Review 839–916.
Meurer, Michael. 2002. ‘Business Method Patents and Patent Floods,’ 8 Washington University Journal of Law
and Policy 309–40.
Mireles, Michael S. Jr. 2006. ‘States as Innovation System Laboratories: California, Patents, and Stem Cell
Technology,’ 28 Cardozo Law Review 1133–212.
Moore, Kimberly A. 2006. ‘Worthless Patents,’ 20 Berkeley Technology Law Journal 1521–52.
Mowery, David C., Richard R. Nelson, Bhaven N. Sampat, and Arvids A. Ziedonis. 2004. Ivory Tower and
Industrial Innovation: University-Industry Technology Transfer Before and After the Bayh-Dole Act. Stanford,
O’Donoghue, Ted, Suzanne Scotchmer, and Jacques-François Thisse. 1998. ‘Patent Breadth, Patent Life, and
the Pace of Technological Progress,’ 7 Journal of Economics & Management Strategy 1–32.
Oliar, Dotan, and Christopher Sprigman. 2008. ‘There’s No Free Laugh (Anymore): The Emergence of
1787–868.
Olson, Thomas P. 1989. ‘The Iron Law of Consensus: Congressional Responses to Proposed Copyright Reforms
Since the 1909 Act,’ 36 Journal of the Copyright Society of the U.S.A.109–37.
O’Rourke, Maureen A. 2000. ‘Toward a Doctrine of Fair Use in Patent Law,’ 100 Columbia Law Review
1177–250.
Osenga, Kristen. 2007. ‘Rembrandts in the Research Lab: Why Universities Should Take a Lesson from Big
Business to Increase Innovation,’ 59 Maine Law Review 407–38.

Outlook. 2009. Tufts Center for the Study of Drug Development. Boston, MA: Tufts University.
Parchomovsky, Gideon, and Peter Siegelman. 2002. ‘Towards an Integrated Theory of Intellectual Property,’ 88
Review 1–78.
Partnoy, Frank. 2001. ‘Finance and Patent Length,’ University of San Diego Law and Economic Research Paper
No. 19, accessed March 20, 2019 at https://papers.ssrn.com/sol3/papers.cfm?abstract_id=285144.
Patry, William F. 1996. ‘Copyright and the Legislative Process: A Personal Perspective,’ 14 Cardozo Arts &
Patry, William F. 2016. Patry on Copyright. New York, NY: Thomson West Publishing.
Plant, Arnold. 1934. ‘The Economic Aspects of Copyright in Books,’ 1 Economica 167–95.
Polanvyi, Michael. 1944. ‘Patent Reform,’ 11 The Review of Economic Studies 61–76.
Posner, Eric A. 1997. ‘Standards, Rules, and Social Norms,’ 21 Harvard Journal of Law and Public Policy 101–18.
Radin, Margaret Jane. 1989. ‘Reconsidering the Rule of Law,’ 69 Boston University Law Review 781–822.
Rai, Arti Kaur. 1999. ‘Regulating Scientific Research: Intellectual Property Rights and the Norms of Science,’
94 Northwestern University Law Review 77–152.
Rai, Arti K., and Rebecca S. Eisenberg. 2003. ‘Bayh-Dole Reform and the Progress of Biomedicine,’ 66 Law
Rajec, Sarah R. 2013. ‘Evaluating Flexibility in International Patent Law,’ 65 Hastings Law Journal 153–210.
Reichman, Jerome H. 1994. ‘Legal Hybrids Between the Patent and Copyright Paradigms,’ 94 Columbia Law
Review 2432–558.
Robinson, William T., Gurumurthy Kalyanaram, and Glen L. Urban. 1994. ‘First-Mover Advantages from
Pioneering New Markets: A Survey of Empirical Evidence,’ 9 Review of Industrial Organization 161–78.
Roin, Benjamin N. 2014. ‘Intellectual Property Versus Prizes: Reframing the Debate,’ 81 University of Chicago
Rose, Carol M. 1988. ‘Crystals and Mud in Property Law,’ 40 Stanford Law Review 577–610.
Rosen, Sherwin. 1981. ‘The Economics of Superstars,’ 71 American Economic Review 845–58.
Rosenblatt, Elizabeth L. 2011. ‘A Theory of IP’s Negative Space,’ 34 Columbia Journal of Law and the Arts 317–66.
Rowe, Elizabeth A. 2006. ‘The Experimental Use Exception to Patent Infringement: Do Universities Deserve
Special Treatment?,’ 57 Hastings Law Journal 921–54.
Samuelson, Pamela. 2011. ‘The Uneasy Case for Software Copyright Revisited,’ 79 George Washington Law
Review 1746–82.
Samuelson, Pamela. 2013. ‘A Fresh Look at Tests for Nonliteral Copyright Infringement,’ 107 Northwestern
Samuelson, Pamela, and Jason Schultz. 2011. ‘“Clues” for Determining Whether Business and Service
Innovations are Unpatentable Abstract Ideas,’ 15 Lewis & Clark Law Review 109–31.
Samuelson, Pamela, and Suzanne Scotchmer. 2002. ‘The Law and Economics of Reverse Engineering,’ 111 Yale
Samuelson, Pamela, Randall Davis, Mitchell D. Kapor, and J.H. Reichman. 1994. ‘A Manifesto Concerning the
Legal Protection of Computer Programs,’ 94 Columbia Law Review 2308–431.
Sarnoff, Joshua D. 2004. ‘Abolishing the Doctrine of Equivalents and Claiming the Future after Festo,’ 19
Sarnoff, Joshua D. 2009. ‘Flexible Application of Injunctive Relief in Intellectual Property Enforcement,’
in Edward Xuan Li and Carlos M. Correa, eds., Intellectual Property Enforcement. Cheltenham, UK and
Saulo, Eleanor C., Birger C. Forsberg, Zul Premji, Scott M. Montgomery, and Anders Björkman. 2008.
‘Willingness and Ability to Pay for Artemisinin-Based Combination Therapy in Rural Tanzania,’ 7 Malaria
Journal 185–96.
Schauer, Frederick. 1991. Playing by the Rules: A Philosophical Examination of Rule-Based Decisionmaking in
Law and in Life. Oxford: Clarendon Press.
Scherer, F.M. 1972. ‘Nordhaus’ Theory of Optimal Patent Life: A Geometric Reinterpretation,’ 62 American
Scherer, F.M. 2001a. ‘Economics of Innovation and Technological Change,’ in Neil J. Smelser and Paul B. Baltes,
eds., International Encyclopedia of the Social & Behavioral Sciences. Oxford: Pergamon Press.
Scherer, F.M. 2001b. ‘The Innovation Lottery,’ in Rochelle Cooper Dreyfuss, Diane Leenheer Zimmerman,
and Harry First, eds., Expanding the Boundaries of Intellectual Property: Innovation Policy for the Knowledge
Society. Oxford: Oxford University Press.
Scherer, F.M., Dietmar Harrhoff, and Jörg Kukies. 2000. ‘Uncertainty and the Size Distribution of Rewards
from Innovation,’ 10 Journal of Evolutionary Economics 175–201.

Schlag, Pierre. 1985. ‘Rules and Standards,’ 33 UCLA Law Review 379–430.
Scotchmer, Suzanne. 1999. ‘On the Optimality of the Patent Renewal System,’ 30 The RAND Journal of
Economics 181–96.
Scotchmer, Suzanne. 2004. ‘The Political Economy of Intellectual Property Treaties,’ 20 Journal of Law,
Economics, and Organization 415–37.
Shapiro, Carl, and Hal Varian. 1999. Information Rules: A Strategic Guide to the Network Economy. Boston,
MA: Harvard Business Review Press.
Shavell, Steven. 2004. Foundations of Economic Analysis of Law. Cambridge, MA: Belknap Press.
Shavell, Steven, and Tanguy van Ypersele. 2001. ‘Rewards versus Intellectual Property Rights,’ 44 The Journal
Shaver, Lea. 2014. ‘Copyright and Inequality,’ 92 Washington University Law Review 117–68.
Sichelman, Ted M. 2010. ‘Commercializing Patents,’ 62 Stanford Law Review 341–413.
Smith, Henry E. 2007. ‘Intellectual Property as Property: Delineating Entitlements in Information,’ 116 Yale
Stiglitz, Joseph. 2006a. ‘Give Prizes Not Patents,’ New Scientist (Sept. 16, 2006).
Stiglitz, Joseph. 2006b. ‘Scrooge and Intellectual Property Rights: A Medical Prize Fund Could Improve the
Financing of Drug Innovations,’ 333 British Medical Journal 1279–80.
Strandburg, Katherine J. 2005. ‘Curiosity-Driven Research and University Technology Transfer,’ in Gary D.
Libecap, ed., University Entrepreneurship and Technology Transfer: Process, Design, and Intellectual Property.
Bradford: Emerald Group Publishing.
Strandburg, Katherine J. 2011. ‘Patent Fair Use 2.0,’ 1 UC Irvine Law Review 265–306.
Sullivan, Kathleen M. 1992. ‘The Justices of Rules and Standards,’ 106 Harvard Law Review 22–123.
Sunstein, Cass R. 1995. ‘Problems with Rules,’ 83 California Law Review 953–1026.
Takeyama, Lisa N. 1994. ‘The Welfare Implications of Unauthorized Reproduction of Intellectual Property in
the Presence of Demand Network Externalities,’ 42 The Journal of Industrial Economics 155–66.
Tuckman, Howard P., and Jack Leahy. 1975. ‘What Is an Article Worth?,’ 83 Journal of Political Economy
951–68.
Wasserman, Melissa Feeney, and Michael Frakes. 2013. ‘Does Agency Funding Affect Decisionmaking? An
Empirical Assessment of the PTO’s Granting Patterns,’ 66 Vanderbilt Law Review 67–148.
Van Overwalle, Geertrui. 2011. ‘Policy Levers Tailoring Patent Law to Biotechnology: Comparing U.S. and
European Approaches,’ 1 UC Irvine Law Review 433–514.
Vishnubhakat, Saurabh. 2015. ‘Expired Patents,’ 64 Catholic University Law Review 419–62.
Wagner, R. Polk. 2003a. ‘Information Wants to Be Free: Intellectual Property and the Mythologies of Control,’
103 Columbia Law Review 995–1034.
Wagner, R. Polk. 2003b. ‘(Mostly) against Exceptionalism,’ in F. Scott Kieff, ed., Perspectives on Properties of
the Human Genome Project. San Diego, CA: Elsevier Academic Press.
Wagner, R. Polk. 2003c. ‘Of Patents and Path Dependency: A Comment on Burk and Lemley,’ 18 Berkeley
Wagner, R. Polk. 2004. ‘Exactly Backwards: Exceptionalism and the Federal Circuit,’ 54 Case Western Reserve
Waterson, Michael. 1990. ‘The Economics of Product Patents,’ 80 American Economic Review 860–69.
Weeds, Helen. 2001. ‘Strategic Delay in a Real Options Model of R&D Competition,’ 69 The Review of
Wright, Donald J. 1999. ‘Optimal Patent Breadth and Length with Costly Imitation,’ 17 International Journal
of Industrial Organization 419–36.
Wu, Timothy. 2005. ‘Copyright’s Communications Policy,’ 103 Michigan Law Review 278–366.
Wu, Tim. 2006. ‘Intellectual Property, Innovation, and Decision Architectures,’ 92 Virginia Law Review 123–48.
Ziedonis, Rosemarie Ham. 2004. ‘Don’t Fence Me In: Fragmented Markets for Technology and the Patent
Acquisition Strategies of Firms,’ 50 Management Science 804–20.
17 U.S.C. § 101, § 102, § 106, § 107, § 110, § 111, § 113, § 115, § 119, § 120, § 121, § 122, § 302, § 408, § 411, § 502,
§ 504, § 512, §§ 901–14, § 1201.
35 U.S.C. § 103, § 155, §§ 200–212, § 271, § 273, § 287.
37 C.F.R. § 202.20.

Cases
AT&T Corp. v. Excel Commc’ns, Inc., 172 F.3d 1352 (Fed. Cir. 1999).
Bernhardt, L.L.C. v. Collezione Europa USA, Inc., 386 F.3d 1371 (Fed. Cir. 2004).
Computer Assocs. Int’l, Inc. v. Altai, Inc., 982 F.2d 693 (2d Cir. 1992).
Diamond v. Chakrabarty, 447 U.S. 303 (1980).
Eldred v. Ashcroft, 537 U.S. 186 (2003).
Harper & Row, Publishers, Inc. v. Nation Enters., 471 U.S. 539 (1985).
In re GPAC, Inc., 57 F.3d 1573 (Fed. Cir. 1995).
Newton v. Diamond, 388 F.3d 1189 (9th Cir. 2004).
O’Reilly v. Morse, 56 U.S. 62 (1853).
State Street Bank & Trust Co. v. Signature Financial Group, Inc., 149 F.3d 1368 (Fed. Cir. 1998).
Walker v. Time Life Films, Inc., 784 F.2d 44 (2d Cir. 1986).

15. Intellectual property enforcement costs
Ben Depoorter*
Contents
I. Introduction
II. Costly Intellectual Property Enforcement
III. The Effect of Enforcement Costs
A. Negative Effects
B. Benefits of Enforcement Costs
C. Distributive Effects of Enforcement Costs
IV. Adjusting Intellectual Property Rights in Light of Costly Enforcement
V. Enforcement of Copyright Law
A. Innovation Spillovers of Copyright Enforcement
B. Social Norm Complications
C. Statutory Damages
D. Small Claims Courts
VI. Enforcement of Patent Law
VII. Enforcement of Trademark Law
VIII. Conclusion
References
I. INTRODUCTION
Conventional economic wisdom is that an appropriation problem plagues markets for

information goods (Posner, 2005, 2014).1 Because information goods can be easily
copied and distributed illegally,2 creative and innovative markets are at risk.
By providing a time-limited monopoly right, intellectual property laws seek to
protect creators and innovators against competition from copyright pirates, trademark
counterfeiters, and patent infringers who did not incur the cost of creation or innova-
tion. Ideally, intellectual property (IP) enforcement discourages intellectual property
* Max Radin Chair and Distinguished Professor of Law, University of California, Hastings
College of the Law; Affiliate Scholar, Stanford Law School, Center for Internet and Society;
CASLE, Ghent University. I express my gratitude to participants of the roundtables convened
by Northwestern University Law School, U.C. Berkeley School of Law (BCLT), and U.C.
Hastings College of the Law for useful suggestions and comments. Contact: depoorter@uchast
ings.edu.
1
On the utilitarian philosophical theory of intellectual property law and the economic account
of copyright law more specifically, see Hadfield (1992), Goldstein (1995).
2
For a discussion, see Syed and Bracha (2014).
407

violations ex ante while compensating rights holders for any lost profits due to infringe-
ments ex post.3
Intellectual property enforcement has become a matter of great controversy over the
past few decades. Concerns with the private and social costs of enforcing IP rights have
instigated intense discussions and a substantial body of literature. Part II first describes
the various costs of IP enforcement. Part III reviews the various negative, positive, and
distributive effects of costly IP enforcement. Part IV discusses the optimal scope of IP
rights in light of enforcement costs. Parts V, VI, and VII review several area-specific
enforcement issues. Part VIII concludes.
II. COSTLY INTELLECTUAL PROPERTY ENFORCEMENT
Enforcing IP rights is big business in a globalized marketplace. Since information goods

and services are so valuable in information markets, the potential gains from infringing
IP rights have increased over time. As the scale of infringements expands over time, so do
the enforcement efforts by right holders who stand to lose more from infringing activities.
There is a widespread perception that enforcement has increased exponentially in the
area of patent law in the past two decades. The growth in the number of patent lawsuits
is instructive. Studies that control for the rapid increase in patenting itself and a shift
toward more litigious technology areas confirm that patent litigation has increased in
impressive numbers (Lanjouw and Schankerman, 2004).4 Several studies indicate that
patent litigation is becoming increasingly difficult to avoid in many innovative sectors
(Eisenberg, 1989; Hall and Ziedonis, 2001; Shapiro, 2001). Likewise, in other areas of IP
law, enforcement has evolved in an upward direction. Enhanced by digital technologies,
global piracy and counterfeiting markets have elevated enforcement efforts by the copy-
right owners (Depoorter et al., 2011) and trademark holders (Grinvald, 2011; Port, 2008).
Enforcing IP rights is an expensive affair. Survey evidence indicates that the median
litigation costs (including, primarily, attorney fees and expenses related to evidentiary
burdens) for patent infringement suits on claims below $1 million are $650,000. When $1
million to $25 million is at stake, total litigation costs are $2.5 million. For a claim with
over $25 million in controversy, the median legal cost is $5 million. The median cost to
litigate a copyright infringement lawsuit with less than $1 million at stake through appeal
is $350,000 (AIPLA, 2011, p. 35). Another study estimated that the average duration of IP
litigation in a home jurisdiction is 3 years with an average cost of $475,000. For litigation
in another jurisdiction, the average duration and costs increases to 3.5 years with legal
fees of just over $850,000 (de Castor and Schallnau, 2013).
The overall level and cost of enforcing IP rights are influenced by several factors, some
of which are endogenous to the legal system, while others are exogenous.
First, the scope of legal protection influences enforcement costs. The potential for
3
Infringers are discouraged ex ante by disgorging the infringer’s profits and imposing puni-
tive measures.
4
Lanjouw and Schankerman (2004) also find that litigation is concentrated in important
ways in firms and patents with particular characteristics.

Intellectual property enforcement costs 409
enforcement increases as the level of IP protection expands. In patent law, for instance, if
restrictions on patentability are relaxed, the increase of patents is likely to induce more
enforcement, as there will be more rights holders and closely related patents that are likely
to overlap. In this regard, practices at the US patent and trademark office (USPTO) affect
enforcement levels.
Second, doctrinal complexity and legal uncertainty about IP protection contribute
to enforcement costs. In copyright law, for example, the use of open-ended standards
(Ehrlich and Posner, 1974; Kaplow, 1995) increases uncertainty. Open-ended standards
provide courts with the much-needed flexibility to adapt copyright doctrine to technologi-
cal advances (Depoorter, 2009; Menell, 2002) but increase ex ante uncertainty and litiga-
tion costs. For instance, the open-ended fair use standard leaves more room for optimism
and disputes among potential litigants. Similarly, doctrines in patent and trademark law
that leave more ambiguity are likely to increase litigation costs.
Third, enforcement costs are influenced by substantive and procedural doctrinal
developments. For instance, when federal courts rejected the broad application of
indirect liability for copyright infringement in cases involving internet service providers,
this rejection reduced the ability of copyright holders to economize on enforcement
costs by targeting the middleman (Lichtman and Landes, 2003). By reducing the role of
gate-keeping liability, the burden to identify infringement remains with the rights holders.
Although IP rights holders can more easily identify infringing content, the removal of
infringing material and identification of infringers is handled more efficiently by the host-
ing party. Similarly, doctrinal developments in civil procedure often have a major effect on
enforcement costs. For instance, when courts rejected the application of joinder concern-
ing peer-to-peer file-sharing infringements, lawsuits against multiple unrelated parties
had to be filed separately, ramping up enforcement costs (Sag, 2015; Sag and Haskell,
2018; Balganesh, 2013). Similarly, when the American Invents Act revised the rules on
joinder, the aggregation of infringement lawsuits became more difficult, rendering it more
burdensome to engage in patent trolling (35 U.S.C. § 299 (2012)).
Other enforcement costs determinants are exogenous to legal doctrine. For instance,
technological advances have a major effect on enforcement costs. Some technologies
are enforcement cost ‘loading’ while others are enforcement cost ‘dumping.’5 Some
technological changes clearly increase enforcement costs. In the area of copyright law,
for instance, the internet and digital technology vastly increased the number of infringe-
ments. In response to a more decentralized ‘digital’ piracy, enforcement costs ramped
up substantially (Menell, 2002). Additionally, the geographical distance and relative
anonymity of online IP infringers have increased overall enforcement costs. Similarly,
in the area of trademark law, the internet’s erosion of geographic restrictions increased
potential confusion between trademarks globally, increasing the need for trademark
policing and other enforcement costs.
In other instances, new technologies may reduce enforcement costs. For instance, web
search tools and automated bots enhance copyright and trademark enforcement6 by
5
For an application in the context of tort liability, see Grady (1988). For a formal model, see
Depoorter and De Mot (2011).
6
On the potential for and costs of scaling copyright enforcement, see Bridy (2010).

making search and notice efforts more comprehensive and less costly. Digital enforcement
tools provide IP rights holders a means to counteract the massive scale of non-commercial
infringements online. Such instruments are criticized, however, for producing false posi-
tives and public backlash (Depoorter and Walker, 2013; Perel and Elkin-Koren, 2016),
On another dimension, digital rights management technologies (such as encryption and
other piracy protection tools) potentially reduce the need for legal enforcement by making
infringements pragmatically more difficult (Hughes, 2016).
III. THE EFFECT OF ENFORCEMENT COSTS

A. Negative Effects
Since IP infringements are the basis for private action by rights holders, the costs of
enforcement are an important determinant of the overall level of IP enforcement.
While enforcement costs entail various elements, including the costs of identifying
infringements, the economic literature on legal enforcement focuses on litigation costs in
the broad sense (costs of discovery, legal fees to lawyers, and so on).
The basic economic model of litigation describes a plaintiff (p) with a potential claim
against a defendant (d). The plaintiff’s estimate of the net expected value (NEV) of the
claim consists of his or her estimation of the probability (P) that he or she will succeed
in trial (Pp) multiplied by the expected award (Dp), minus the costs of litigation of the
plaintiff (Cp) (Gould, 1973; Landes and Posner, 1975, 1979).7
The net expected value (NEV) of a lawsuit is the primary determinant of whether or not
a plaintiff will file a suit to enforce their rights. Although the value of a lawsuit is deter-
mined primarily by the expected award in court (Dp), the benefits of filing suit might also
follow from other factors, such as the precedential value of the lawsuit, the expectation
that the defendant will settle in order to avoid the publicity (Daughety and Reinganum,
1999), the costs of trial (Rosenberg and Shavell, 1985), and the political mobilizing effect
of a lawsuit (Depoorter, 2013).
Litigation costs may create what economists term ‘negative value claims.’ Even though
an infringement claim might have merit, litigation costs might render the plaintiff’s threat
to litigate non-credible. For instance, even if the probability of the plaintiff prevailing in
court is 80 percent and the likely damage award is $10,000, the plaintiff has no financial
incentive to pursue the claim if the litigation costs exceed $8,000. The value of the claim
is negative for the plaintiff if litigation costs are $9,000 (-$1,000). Negative value claims
are not a credible basis for litigation.
Moreover, negative value suits create a potential gap between the private and social
incentives to enforce (Shavell, 1982). Although the social benefits of enforcement might
be substantial, an individual rights holder may lack the incentive to bear the costs of
7
Conversely, the defendant’s estimated expected loss (EL) equals his or her estimate of the
probability the plaintiff will prevail (Pd) times the expected award (Dd) plus the defendant’s costs
of litigation (Cd).

enforcement (Hazard, 1989; Kaplow, 1986; Menell, 1983; Priest, 1982; Rose-Ackerman
and Geistfeld, 1987; Schwartz and Tullock, 1975; Williams and Williams, 1994).
Given that enforcement increases the overall deterrent effect of the legal system, private
enforcement can be conceived as a public good subject to a collective action problem.
For instance, in the context of IP enforcement, rights holders as a group benefit from the
enforcement of infringement claims. Each individual rights holder, however, would prefer
that another incurs the expense of the positive external effects of enforcement. This collective
action problem creates a potential problem of under-enforcement. More generally, this diver-
gence between the private and social incentives of enforcement undermines the deterrent
effect of IP laws and remedies. Public enforcement of IP rights can fill the gap. Public pros-
ecutions of IP law violations are relatively uncommon, however, except for some high-profile
cases (Haber, 2015). Moreover, the efficacy of cross-border enforcement of trademark and
patent violations remain a topic of contention and widespread dissatisfaction (Yu, 2012).
Another negative effect of enforcement costs is the potential for opportunistic enforce-
ment practices by plaintiffs (Bebchuk, 1988, 1996; Bone, 1997; Katz, 1990; Miceli, 1993;
Rosenberg and Shavell, 1985). When litigation costs are high, aggressive litigants might
obtain generous settlements. Although defendants might have a high probability of suc-
cessfully challenging the infringement claim, the litigation costs might be prohibitive to
many defendants due to, for instance, liquidity constraints or negative publicity effects,
and so on. As such, the cost of defending against dubious enforcement actions are fertile
soil for opportunistic enforcement practices by copyright and patent trolls (Sag, 2015).
When based on dubious claims, such enforcement practices operate as a tax on creative
and innovative practices (Sag, 2015; Sag and Haskell, 2018).
B. Benefits of Enforcement Costs
The analysis above is based on the premise that IP protection is set at socially optimal levels.
In the utilitarian perspective, the scope of IP rights should optimize the access-incentive
trade-off: at a level where the marginal incentive benefits of additional protection equal
the marginal costs of any further reduction of access to the public and follow-up creators
and innovators (Syed and Bracha, 2014).
Given the inherent difficulties in measuring incentive effects in creative and innovative
industries, as well as new evidence on alternative drivers of creativity and innovation,8 the
notion that IP laws drive creativity is called into question. Some observers have argued
that the exclusive rights granted by IP laws have become overly strong (Jaffe and Lerner,
2004; Maskus and Reichman, 2004).9
8
A growing literature shows, however, that creativity is often driven by motives unaffected
by copyright protection. For a collection of recent work, see Darling and Perzanowski (2017).
Similarly, an ongoing literature has pointed to ways in which authorship can be commercialized
outside of a copyright framework. Seminal contributions include Plant (1934) and Hurt and
Schuchman (1966). Empirical accounts of the minimal effects of copyright law on creative output
include Waldfogel (2012) and Ku et al. (2009). But see on the commercialization side of IP rights,
Sichelman (2010) and Barnett (2013).
9
Potential causes include the creation of a specialized patent court (Posner and Landes, 2003)
and the political rent-seeking in copyright law (Litman, 1996).

In this regard, costly enforcement might have the unintended yet beneficial effect of
reducing the overall level of enforcement and, as a result, bringing intellectual property
protection closer to the socially optimal level. For instance, it has been stated that under-
enforcement ‘operates as an important safety valve that introduces an informal breathing
space into copyright’s functioning’ (Balganesh, 2013, p. 723). Such gaps between the rights
conveyed (‘de jure’) and what is enforced (in practice) have been documented in various
areas of IP law. In the area of copyright, digital technologies and the internet facilitated
a wide variety of uses that are copyright infringing in theory yet tolerated pragmatically
by right holders in practice (Tehranian, 2007; Wu, 2008). In patent law, many innovators
tend to ignore existing patents since most patents remain unenforced anyway and/or have
limited validity (Lemley, 2008). Similarly, in trademark law some rights are conveyed ‘de
jure’ but enforced only in limited circumstances.10 Finally, empirical work reveals a major
gap between law in the books and law in action in the area of criminal enforcement of IP
law, indicating that criminal copyright prosecutions are relatively rare, especially given the
amount of criminal copyright legislation in effect (Haber, 2015).
C. Distributive Effects of Enforcement Costs
On a basic level, enforcement costs harm all holders of IP rights. Upon closer inspection,
however, enforcement costs impact some IP stakeholders more than others. As a general
observation, legal rights are ‘relationally contingent,’ providing protection primarily
against challengers who face higher litigation costs than the rights holders. Conversely,
‘challengers who can litigate more cheaply than a right holder can force the right holders
to forfeit the right and thereby render the right ineffective’ (Parchomovsky and Stein,
2012). Like in other areas of law, larger stakeholders can scale enforcement costs and gain
an expertise in enforcement techniques as repeat players (Galanter, 1974).11
In copyright law, for instance, some have observed that the costs of exercising
statutorily defined rights can be prohibitive and disproportionally affect non-professional
creators who do not have the capacity to scale out the costs of exercising these rights. In
this vein, it has been observed that many smaller creators and innovators are more risk
averse, inducing them to engage in generous licensing concessions rather than vigorously
defend rights preserved to the public (Gibson, 2007). Over time, such ‘settlements outside
of the law’ (Depoorter, 2010) may become part of a trend of rights accretion since courts
will recognize such customary licensing practices as accepted within an industry (Gibson,
2007; Rothman, 2007).
Similarly, in patent law, repeat players and large firms have certain advantages in manag-
ing IP enforcement costs. Large, innovative firms can protect patent rights and defend
against patent infringement claims more easily because (1) their reputation as repeat players
may induce bargaining and settlement (Bernheim and Whinston, 1990; Siegelman and
Waldfogel, 1999; Tirole, 1994) and (2) a larger patent portfolio enables countersuits and may
10
For instance, Long (2006) describes how, prior to the amendments of the Federal Trademark
Dilution Act (15 U.S.C. § 1125(c), 1127), federal courts were relatively unreceptive to enforcing the
doctrine of trademark dilution.
11
For empirical evidence on the effect of repeat player on litigation, see Siegelman and
Waldfogel (1999).

work as a deterrent in litigation. Empirical work by Lanjouw and Schankerman (2004) finds
that a larger patent portfolio reduces the probability of litigation on any individual patent in
the portfolio, suggesting there are beneficial ‘enforcement spillovers within a given firm in
enforcing its different patents’ (p. 4). This finding suggests that it is less costly to protect any
given patent when it is a fragment of a bigger patent portfolio. As a result, the litigation risk
is more pronounced for patents owned by individuals and firms with small patent portfolios.
The distributive effect of enforcement costs may have a significant dynamic effect.
Enforcement costs can have the inadvertent consequence of moving innovation away
from smaller firms. For example, Lerner (1995) shows that small firms avoid R&D areas
where the threat of litigation from larger firms is high. Lanjouw and Lerner (2001) argue
that the use of preliminary injunctions by large firms can negatively affect the incentives
of small firms to undertake R&D (Lanjouw and Schankerman, 2004).12 The threat of
enforcement may also induce more frequent trading of IP and, therefore, create a desire
to obtain patents as bargaining chips in future negotiations (Hall and Ziedonis, 2001).
IV. ADJUSTING INTELLECTUAL PROPERTY RIGHTS IN

LIGHT OF COSTLY ENFORCEMENT
Economic theory posits that private property rights adapt to changes in underlying costs and
benefits. In Harold Demsetz’s words, ‘[p]roperty rights develop to internalize externalities
when the gains of internalization become larger than the cost of internalization’ (Demsetz,
1967, p. 350) In his seminal account of the Montagnes Indians of the Labrador Peninsula,
Demsetz described how private rights in land and forest animals developed in response
to heightened opportunities in the commercial fur trade. Because overhunting presented
a serious problem once fur became commercially valuable, a strong incentive existed to
internalize costs by way of property rights protection. This story of the maturation of
property rights emphasizes that, with increases in value and economic activity, property
rights evolve to become stronger and increasingly more private. Applied to IP rights, this
positive economic theory of property rights suggests that the observed expansion of IP
over time aligns with the increase of the economic value of information goods because of
new technologies and the maturation of information markets (Depoorter, 2005).
Enforcement costs figure prominently as a factor of influence in the overall evolution of
property rights. For instance, in his classic writing, Demsetz attributes the relative absence
of private property rights on the Southwestern plains to the high costs of containing wide
range, migratory animals. For Indians of the Labrador Peninsula, by contrast, fencing
forest animals was relatively less expensive. Variance in the degree of private property
rights protection can be explained in relation to the costs involved in the ‘fencing’ of those
assets (Demsetz, 1967, p. 353).13
12
On the basis of data regarding patent suits and settlements during 1978–99, linking infor-
mation from the US patent office, the federal courts, and industry sources, the authors find that
litigation risk is much higher for patents owned by individuals and firms with small patent portfolios.
13
For other analytical accounts on the costs of excluding and enforcing rights for the emer-
gence of private property, see Field (1989), Ellickson (1993), Lueck (2002), Anderson and Hill
(1975, 1983), and Libecap and Smith (2002).

In the context of copyright law, for instance, technological changes have led to calls to
expand as well as reduce the scope of IP protection. On the one hand, when digital rights
management and encryption technologies promised seamless exclusion and enforcement
of copyrighted content, it was held that copyright exceptions should be constricted (Bell,
1998; Hardy, 1996; Kitch, 1999). On the other hand, technological change may also bring
about a move from private property back to open-access arrangements (Levmore, 2002).
For instance, file-sharing technologies made it substantially more difficult to enforce
copyright online. In light of these enforcement complications, copyright law becomes
costlier to society. Various steps toward an ‘information commons’ have been advocated
in the literature, including the exemption of personal, non-commercial uses of content
(Litman, 1996) reducing the statutory damage range as it applies to non-commercial
infringements (Samuelson and Wheatland, 2009), and a tolerance for small value infringe-
ments online (Tehranian, 2007; Wu, 2008).
V. ENFORCEMENT OF COPYRIGHT LAW
Traditionally, copyright enforcement has focused on commercial pirates and bootleg

recorders that produced and/or distributed creative works on a large scale. However, the
digital revolution fundamentally altered the modern landscape of copyright enforcement
(Cohen, 2006). Faced with rampant online infringements, copyright holders aggressively
protect their rights, targeting internet service providers, software developers, amateur
artists, home copyists, or others who previously did not come to the attention of copyright
industries. This Part describes several copyright-specific enforcement issues.
A. Innovation Spillovers of Copyright Enforcement
Many uses of copyrighted content—both legitimate and unauthorized uses—are enabled

by technologies. Whenever a new technology emerges, copyright holders are alarmed by
the potentially negative impact of these technologies on already existing business models
(Lemley, 2011). To halt, or at least control, infringing uses of these technologies, copyright
industries target the developers of new technologies based on intermediary liability doc-
trines (Ginsburg, 2002; Lichtman and Landes, 2003). In doing so, copyright enforcement
influences the path of development of new technologies—its availability, permitted uses,
and commercial future. For instance, qualitative evidence suggests that the ex ante effect
of copyright enforcement costs runs much deeper, affecting investment in new technolo-
gies going forward. In-depth interviews with chief executive officers, company founders,
and vice presidents from technology companies and venture capital firms suggest that the
injunction in the Napster case had a debilitating effect on venture capital investments in
technologies involving copyrighted content (Carrier, 2012).
B. Social Norm Complications
Motivated by rampant online infringements, copyright holders have waged an aggressive

legal campaign against online piracy since the early 2000s. In doing so, the record industry
made an unprecedented shift in enforcement by targeting individual, non-commercial

infringers for direct copyright infringement (Depoorter and Van Hiel, 2015; Giblin,
2011). The phenomena of massive amounts of low value infringements on P2P networks
presented a vexing enforcement dilemma for copyright holders, however. On the one
hand, to attain some level of deterrence, stringent sanctions are needed to offset the rather
low probability of the enforcement; especially given the vast number of infringers and the
pragmatic obstacles to enforcement. On the other hand, in the effort to establish some
measure of deterrence, enforcement may come to be regarded as being disproportionate
or excessive. Social psychologists have shown that legal obedience is ‘morality-based’ and/
or ‘legitimacy-based’ (Carroll, 1987; James et al., 2001; Wenzel, 2005). Individuals tend to
observe laws more readily when doing so is perceived as the right thing to do (Tyler, 1990).
Moreover, individuals are more likely to disregard legal commands if they believe that the
law is out of touch or unjust. Some individuals will ignore the risks of noncompliance
more easily and some may even disobey the legal rule as a matter of principle. In the
context of tax compliance, for example, several studies indicate that stringent enforce-
ment measures may backfire, inducing increased tax evasion (Carroll, 1987; Grasmick
and Green, 1980; Lederman, 2003; Smith, 1990). This normative aspect of compliance
likely complicated enforcement of copyright law in the P2P era. Empirical evidence has
documented the personal beliefs and social norms that sharing private content is socially
desirable (Feldman and Nadler, 2006; Gervais, 2004). Experimental evidence suggests
that the stringent enforcement of copyright law may have had the unintended effect of
reinforcing and strengthening the belief that copyright law is unjust, especially if legal
sanctions are perceived as excessive relative to the infringing behavior (Depoorter and
Vanneste, 2005, 2011; Depoorter and Van Hiel, 2015).14
C. Statutory Damages
In pursuing a copyright infringement claim, the Copyright Act does not require that
all plaintiffs provide evidence of injury from infringement. Once infringement has been
established, copyright holders of registered works can elect that the jury sets a statutory
damage award (17 U.S.C. § 504(c)). This litigation cost-reducing effect of statutory dam-
ages helps reduce the issue of negative value suits and, in doing so, fosters deterrence even
for violations of works by small artists.
In recent years, however, many commentators claim that the statutory damage system
is understood to be out of step with the digital age (Samuelson and Wheatland, 2013).
At the time of the 1976 Copyright Act, piracy typically involved the sale of multiple
copies of the same work. Since the advent of P2P networks and BitTorrent applications,
however, online infringements typically involve infringers that copy many different works.
Since the 1976 Copyright Act provides a statutory award for each infringed work, statu-
tory damages can quickly add up for online infringements. For instance, one file-sharer
was ordered to pay $222,000 in statutory damages for sharing 24 songs online (Capitol
Records, Inc. v. Thomas-Rasset, 692 F.3d 899 (8th Cir. 2012)). In another case, a jury
14
For a discussion of the alternative, soft ‘nudge’ approach to mass copyright infringements in
the United States, see Bridy (2012). For an empirical examination of the graduate response systems,
see Depoorter and Van Hiel (2015).

imposed $675,000 in statutory damages for the sharing of 30 songs (Sony BMG Music
Entm’t v. Tenenbaum, 719 F.3d 67 (1st Cir. 2013)). Critics observe that the system of
statutory damages facilitates copyright trolling, enabling plaintiffs to assert rights they do
not have and making poorly substantiated claims (Sag, 2015). Although dubious infringe-
ment claims are likely to be negated in litigation, many defendants prefer to settle the
dispute and avoid incurring the costs of combatting the claim in court. Empirical analysis
of dockets and case law reveal that statutory damages claims are commonplace in virtually
all areas of copyright law (Depoorter, 2019a). Plaintiffs assert enhanced damages in 81
percent of all copyright disputes in the examined period, yet courts awarded enhanced
damages in less than 2 percent of all cases that moved to verdict. One possible explanation
is that the claims for enhanced damages serve the goal of subduing alleged infringers into
making settlement concessions. These findings support reforming the statutory damage
framework (Samuelson and Wheatland, 2009).
D. Small Claims Courts
There is a widespread belief that the area of copyright has a small claims problem.
Litigation in federal court is prohibitively expensive for most copyright stakeholders. The
time, effort, and legal costs involved with litigation outweigh the resources available to
many copyright holders; especially given the modest amounts at stake in most disputes.
One promising solution to copyright’s problem of negative value suits is to establish a
small claims court (Lemley and Reese, 2004, 2005; Depoorter, 2019b). A small claims pro-
cess can provide a smoother, more cost-effective forum to decide minor, straightforward
infringements. Ideally, such a process enables sympathetic plaintiffs, such as individual
photographers and independent fashion designers, to obtain relief against infringers,
while also enabling accused infringers to mount a more efficient defense against dubious
accusations by opportunistic plaintiffs.
VI. ENFORCEMENT OF PATENT LAW
It is well documented that the cost of enforcing patents is often very steep. According to
a 2009 economic survey by the American Intellectual Property Law Association, average
infringement litigation costs exceed $3 million when the amount in dispute is between $1
million and $25 million (AIPLA, 2011).
Litigation costs are especially high for information technology industries, where the
boundaries between potentially infringing patents are often unclear or hard to disentangle
from an information cost perspective (Bessen and Meurer, 2009; Menell and Meurer, 2014).
The landscape of patent enforcement has changed significantly since 2000–08. During
that period, practicing entities and industry competitors filed a majority of lawsuits.
Since 2008, however, patent holders who do not manufacture products themselves have
been initiating a high number of patent lawsuits.15 This heterogeneous group of patent
15
A number of studies document the spectacular rise of lawsuits initiated by patent assertion
companies early this decade, especially in the information technology (IT) industries (Chien, 2009,

holders (referred to as non-practicing entities, (NPEs), patent assertion entities (PAEs),

patent monetization entities (PMEs), or patent trolls) includes universities, failed start-
ups, speculators who purchase patents from others, and companies formed by venture
capitalists in order to exploit the inventions by others (Allison et al., 2009; Chien, 2012;
Cotropia et al., 2014; Risch, 2012).
Patent owners whose primary business is to obtain income from alleged infringers of
their patents are a topic of intensive discussion; both in the public arena and the academic
literature. In the most benign version of this enforcement model, NPEs simply set about
a more efficient model of enforcing patents. Specializing in patent enforcement, these
entities can scale enforcement and gain expertise in evaluating allegations of infringement.
Along these lines, the presence of NPEs provides an avenue for inventors to monetize
innovative activity. For instance, individual inventors who do not otherwise commercial-
ize their patents can obtain some return on their inventions by selling their patent rights
to enforcement entities. In this positive outlook on enforcement intermediaries, NPEs
pursue meritorious claims that yield recoveries in excess of out-of-pocket litigation
expenses (Shrestha, 2010).
In the more negative interpretation of specialized enforcement entities, litigants lever-
age the defendant’s prospective litigation costs in order to obtain a settlement. The most
socially wasteful variation involves NPEs that assert frivolous and extortionary claims.16
Such ‘bottom-feeder’ trolls seek quick settlements far below the expensive cost of patent
litigation, which typically costs millions of dollars (Lemley and Melamed, 2013). 17
As a general observation, NPEs can be expected to engage in aggressive forms of patent
enforcement. Specialized enforcers benefit from obtaining a reputation as ferocious
litigants and tough negotiators. An operation or manufacturing company, by contrast,
may be more likely to moderate its enforcement strategies in light of long-term business
relationships with potential infringers, or out of fear of countersuits for infringement.
The empirical evidence on litigation by specialized enforcement entities is somewhat
mixed. Several empirical studies have examined the characteristics of enforcement by
patent trolls and NPEs.18 NPEs are documented to win both larger judgments and larger
settlements than do practicing entities (PwC, 2012). And trolls do so despite complaints
2012; Feldman and Ewing, 2012) as well as the increasingly more complex and organized practices
of so-called ‘super-trolls’ or ‘troll aggregators’ in the gathering and asserting or licensing of massive
amounts of patents (Feldman and Ewing, 2012; Allison et al., 2009).
16
Using the decline of stock values as a benchmark, one study calculated that trolls cost
society approximately $30 billion per year (Bessen and Meurer, 2014) and have cost a total of
$500 billion over the past 20 years (Bessen et al., 2012). More generally, patent enforcement costs
are stated to outweigh the benefits of patent rights more generally in areas, such as IT industries,
where potentially overlapping rights are hard to disentangle from an information cost perspective
(Bessen and Meurer, 2009). In these circumstances, the suggestion is that, due to the enforcement
costs, society and innovators would be better off without patents (Boldrin and Levine, 2009), unless
enforcement costs can be moderated.
17
Lemley and Melamed (2013) associate patent trolling with patent fragmentation issues
(Heller and Eisenberg, 1998; Lemley and Shapiro, 2007). The disaggregation of complementary
patents into different hands, as opposed to the aggregation by non-practicing entities as such,
amplifies the problems associated with patent trolls.
18
On the predictability of patent litigation, see Mazzeo et al. (2013).

that they often assert weak patents (Merges, 2009),19 despite some evidence that troll-
owned patents are more likely than other patents to lose in court (Allison et al., 2011;
Miller, 2013).20 Several valuable studies have made strides, carefully analyzing various
types of specific litigants within the broad and diverse group of NPEs (Cotropia et al.,
2014; Lemley and Melamed; 2013; Schwartz, 2012).21
VII. ENFORCEMENT OF TRADEMARK LAW
Trademark law incentivizes trademark owners to engage in enforcement. Besides the

common law equity principles of laches and estoppel that also apply to other areas of IP
law, trademark holders are under doctrinal pressure to enforce their rights. For instance,
because failure to control the use of a trademark by third parties can lead to a loss of
rights, strong or famous trademarks receive stronger protection (Beebe and Hemphill,
2017), and the doctrine of genericide encourages very aggressive policing (Desai and
Rierson, 2007; Dreyfuss, 1990).22
Although active enforcement of trademark law is the modus operandi of trademark hold-
ers, trademark holders are sometimes accused of crossing the line from aggressive to abusive
practices of trademark enforcement. Trademark bullying has been defined as enforcement
of ‘an unreasonable interpretation by a large corporation of its trademark rights against a
small business or individual through the use of intimidation tactics’ (Grinvald, 2011; Greene,
2004). Empirical research on trademark litigation documents potentially opportunistic uses
of strike suits to deter market entrants (Port, 2008) or to obtain settlements on weak infringe-
ment claims (Gallagher, 2012). From an economic perspective, these trademark enforcement
tactics mimic strategies by incumbents to raise the costs of market entry for prospective
competitors (Milgrom and Roberts, 1982; Salop and Scheffman, 1986; Waldman, 1987).
In light of these developments, scholars have proposed to implement stronger trademark
infringement defenses (Dinwoodie, 2009; Grynberg, 2009; McGeveran, 2008) and doctrines
of IP misuse (Ridgway, 2006). Such measures might help counteract the negative effects of
aggressive enforcement on competition and free speech (Ramsey, 2003; Schlosser, 2001).
VIII. CONCLUSION
Intellectual property enforcement has become a matter of great controversy over the
past few decades. Concerns with the private and social costs of enforcing IP rights have
instigated intense discussions and a substantial literature.
19
But see Miller (2010) (citation-based finding of higher-quality patents in litigation by NPEs).
20
But see, for evidence to the contrary, Miller (2013).
21
For a further distinction among various types of mass aggregators and their practices, see
Cotropia et al. (2014) and Risch (2014). For an empirical description of mass aggregators, see
Ewing and Feldman (2012).
22
Some substantive rules in trademark law have been explained in light of the goal of reducing
enforcement costs and the costs of administration of justice overall (Bone, 2004). A similar argu-
ment has been made in the context of copyright law (Lichtman, 2003).

At least three major trends feature prominently in the literature. First, as markets
have developed more fully on a global scale, IP infringements have exploded over time,
reaching more consumers of infringing materials. The geographic dispersion of infringers
significantly increases enforcements costs overall. Second, technologies have a double-
edged effect on enforcement costs. On the one hand, technologies facilitate the ease and
appeal of infringing activities. On the other hand, technological advances enable scaled
enforcement. Third, especially regarding patents, the explosion of intellectual property
rights granted over time has prompted enforcement acceleration by enforcement inter-
mediaries. These new entities have ramped up enforcement considerably. The nuances
of these enforcement practices are being teased out and the overall effect on innovation
remain topics of further exploration.
REFERENCES
AIPLA (American Intellectual Property Law Association). 2011. Report of the Economic Survey.
Allison, John R., Mark A. Lemley, and Joshua Walker. 2009. ‘Extreme Value or Trolls on Top: The
Characteristics of the Most-Litigated Patents,’ 158 University of Pennsylvania Law Review 1–38.
Allison, John R., Mark A. Lemley, and Joshua Walker. 2011. ‘Patent Quality and Settlement among Repeat
Patent Litigants,’ 99 Georgetown Law Journal 677–712.
Anderson, Terry L., and Peter J. Hill. 1975. ‘The Evolution of Property Rights: A Study of the American West,’
18 Journal of Law & Economics 163–80.
Anderson, Terry L., and Peter J. Hill. 1983. ‘Privatizing the Commons: An Improvement?,’ 50 Southern
Economic Journal 438–50.
Balganesh, Shyamkrishna. 2013. ‘The Uneasy Case Against Copyright Trolls,’ 86 Southern California Law
Review 723–82.
Barnett, Jonathan M. 2013, ‘Copyright Without Creators,’ 9 Review of Law & Economics 389–438.
Bebchuk, Lucian A. 1988. ‘Suing Solely to Extract a Settlement Offer,’ 17 Journal of Legal Studies 437–50.
Bebchuk, Lucian A. 1996. ‘A New Theory Concerning the Credibility and Success of Threats to Sue,’ 25 Journal
Beebe, Barton, and C. Scott Hemphill. 2017. ‘The Scope of Strong Marks: Should Trademark Law Protect the
Strong More Than the Weak,’ 92 New York University Law Review 1339–98.
Bell, Tom W. 1998. ‘Fair Use vs. Fared Use: The Impact of Automated Rights Management on Copyright’s Fair
Use Doctrine,’ 76 North Carolina Law Review 557–620.
Bernheim, B. Douglas, and Michael D. Whinston. 1990, ‘Multimarket Contact and Collusive Behavior,’ 21 The
Bessen, James, and Michael J. Meurer. 2014. ‘The Direct Costs from NPE Disputes,’ 99 Cornell Law Review 387–424.
Bessen, James E., Jennifer Laurissa Ford, and Michael J. Meurer. 2012. ‘The Private and Social Costs of Patent
Trolls,’ 34 (4) Regulation 26–35.
Boldrin, Michele, and David K. Levine. 2009. Against Intellectual Monopoly. Cambridge: Cambridge University
Press.
Bone, Robert G. 1997. ‘Modeling Frivolous Suits,’ 145 University of Pennsylvania Law Review 519–605.
Bone, Robert G. 2004. ‘Enforcement Costs and Trademark Puzzles,’ 90 Virginia Law Review 2099–186.
Bridy, Annemarie. 2010. ‘Is Online Copyright Enforcement Scalable?,’ 13 Vanderbilt Journal of Entertainment
& Technology Law 695–737.
Bridy, Annemarie. 2012. ‘Graduated Response American Style: “Six Strikes” Measured against Five Norms,’ 23
Fordham Intellectual Property Media & Entertainment Law Journal 1–67.
Carrier, Michael A. 2012. ‘Copyright and Innovation: The Untold Story,’ 2012 Wisconsin Law Review 891–962.
Carroll, John S. 1978, ‘A Psychological Approach to Deterrence: The Evaluation of Crime Opportunities,’ 36
Journal of Personality & Social Psychology 1512–20.
Carroll, John S. 1987. ‘Compliance with the Law: A Decision-Making Approach to Taxpaying,’ 11 Law and
Human Behavior 319–35.
Chien, Colleen V. 2009. ‘Of Trolls, Davids, Goliaths, and Kings: Narratives and Evidence in the Litigation of
High-Tech Patents,’ 87 North Carolina Law Review 1571–616.

Chien, Colleen V. 2012. ‘Patent Assertion Entities,’ Presentation to the December 10, 2012 DOJ/FTC Hearing
on PAEs 1–86, accessed at https://digitalcommons.law.scu.edu/facpubs/590/.
Cohen, Julie E. 2006. ‘Pervasively Distributed Copyright Enforcement,’ 95 Georgetown Law Journal 1–48.
Cotropia, Christopher A., Jay P. Kesan, and David L. Schwartz. 2014. ‘Unpacking Patent Assertion Entities
(PAEs),’ 99 Minnesota Law Review 649–704.
Darling, Kate, and Aaron Perzanowski. 2017. Creativity Without Law: Challenging the Assumptions of
Intellectual Property. New York, NY: New York University Press.
Daughety, Andrew F., and Jennifer F. Reinganum. 1999. ‘Hush Money,’ 30 The RAND Journal of Economics
661–78.
de Castor, Ignacio, and Judith Schallnau. 2013. ‘What Does It Cost to Defend Your IP Rights?,’ 3/2013, WIPO
Magazine, accessed April 3, 2019 at http://www.wipo.int/wipo_magazine/en/2013/03/article_0006.html.
Demsetz, Harold. 1967. ‘Toward a Theory of Property Rights,’ 57 The American Economic Review 347–59.
Depoorter, Ben. 2005. ‘The Several Lives of Mickey Mouse: The Expanding Boundaries of Intellectual Property
Law,’ 9 Virginia Journal of Law & Technology 1–68.
Depoorter, Ben. 2009. ‘Technology and Uncertainty: The Shaping Effect on Copyright Law,’ 157 University of
Pennsylvania Law Review 1831–68.
Depoorter, Ben. 2010. ‘Law in the Shadow of Bargaining: The Feedback Effect of Civil Settlements,’ 95 Cornell
Depoorter, Ben. 2013. ‘The Upside of Losing,’ 113 Columbia Law Review 817–62.
Depoorter, Ben. 2019a. ‘When the Remedy is the Wrong: Copyright Enforcement in the Digital Age,’ forthcom-
ing, 66(2) UCLA Law Review 400–47.
Depoorter, Ben. 2019b. ‘When You Build It, They Will Come: The Promises and Pitfalls of a Small Claims
Process,’ 33 Berkeley Technology Law Journal 711–34.
Depoorter, Ben, and Jef De Mot. 2011. ‘Technology & Torts: A Theory of Memory Costs, Nondurable
Precautions and Interference Effects,’ 31 International Review of Law and Economies 284–304.
Depoorter, Ben, and Alain Van Hiel. 2015. ‘Copyright Alert Enforcement: Six Strikes and Privacy Harms,’ 39
Columbia Journal of Law & the Arts 233–80.
Depoorter, Ben, and Sven Vanneste. 2005. ‘Norms and Enforcement: The Case Against Copyright Litigation,’
84 Oregon Law Review 1127–80.
Depoorter, Ben, and Robert Kirk Walker. 2013. ‘Copyright False Positives,’ 89 Notre Dame Law Review 319–60.
Depoorter, Ben, Alain Van Hiel, and Sven Vanneste. 2011. ‘Copyright Backlash,’ 84 Southern California Law
Review 1251–92.
Desai, Deven R., and Sandra L. Rierson. 2007. ‘Confronting the Genericism Conundrum,’ 28 Cardozo Law
Review 1789–856.
Dinwoodie, Graeme B. 2009. ‘Developing Defenses in Trademark Law,’ 13 Lewis & Clark Law Review 99–154.
Dreyfuss, Rochelle Cooper. 1990. ‘Expressive Genericity: Trademarks as Language in the Pepsi Generation,’ 65
Notre Dame Law Review 397–424.
Ehrlich, Isaac, and Richard A. Posner. 1974. ‘An Economic Analysis of Legal Rulemaking,’ 3 Journal of Legal
Studies 257–86.
Eisenberg, Rebecca S. 1989. ‘Patents and the Progress of Science: Exclusive Rights and Experimental Use,’ 56
Ellickson, Robert C. 1993. ‘Property in Land,’ 102 Yale Law Journal 1315–400.
Feldman, Robin, and Tom Ewing. 2012, ‘The Giants Among Us,’ 2012 Stanford Technology Law Review 1–47.
Feldman, Yuval, and Janice Nadler. 2006. ‘The Law and Norms of File Sharing,’ 43 San Diego Law Review
577–618.
Field, Barry. C. 1989. ‘The Evolution of Property Rights,’ 42 Kyklos International Review for Social Sciences
319–45.
Galanter, Marc. 1974. ‘Why the Haves Come out Ahead: Speculations on the Limits of Legal Change,’ 9 Law
& Society Review 95–160.
Gallagher, William T. 2012. ‘Trademark and Copyright Enforcement in the Shadow of IP Law,’ 28 Santa Clara
Computer & High Technology Law Journal 453–98.
Gervais, Daniel. 2004. ‘The Price of Social Norms: Towards a Liability Regime for File-Sharing,’ 12 Journal of
Intellectual Property 39–74.
Giblin, Rebecca. 2011. Code Wars: 10 Years of P2P Software Litigation. Cheltenham, UK and Northampton,
Gibson, James. 2007. ‘Risk Aversion and Rights Accretion in Intellectual Property Law,’ 116 Yale Law Journal
882–951.
Ginsburg, Jane C. 2002. ‘How Copyright Got a Bad Name for Itself’, 26 Columbia Journal of Law & the Arts
61–74.
Goldstein, Paul. 1995. Copyright’s Highway: From Gutenberg to the Celestial Jukebox. Stanford, CA: Stanford
University Press.

Gould, John P. 1973. ‘The Economics of Legal Conflicts,’ 2 Journal of Legal Studies 279–300.
Grady, Mark F. 1988. ‘Why Are People Negligent? Technology, Nondurable Precautions, and the Medical
Malpractice Explosion,’ 82 Northwestern University Law Review 293–334.
Grasmick, Harold G., and Donald E. Green. 1980. ‘Legal Punishment, Social Disapproval and Internalization
as Inhibitors of Illegal Behavior,’ 17 Journal of Criminal Law & Criminology 325–35.
Greene, K.J. 2004. ‘Abusive Trademark Litigation and the Incredible Shrinking Confusion Doctrine: Trademark
Abuse in the Context of Entertainment Media and Cyberspace,’ 27 Harvard Journal of Law & Public Policy
609–42.
Grinvald, Leah Chan. 2011. ‘Shaming Trademark Bullies,’ 2011 Wisconsin Law Review 625–90.
Grynberg, Michael. 2009. ‘Things Are Worse than We Think: Trademark Defenses in a Formalist Age,’ 24
Haber, Eldar. 2015. ‘The Criminal Copyright Gap,’ 18 Stanford Technology Law Review 247–88.
Hadfield, Gillian. 1992, ‘The Economics of Copyright: An Historical Perspective,’ 38 Copyright Law Symposium
(ASCAP) 1–46.
Hall, Bronwyn, and Rosemarie Ziedonis. 2001. ‘The Patent Paradox Revisited: An Empirical Study of Patenting
in the Semiconductor Industry, 1979–1999,’ 32 The RAND Journal of Economics 101–28.
Hardy, Trotter. 1996. ‘Property (and Copyright) in Cyberspace,’ 1996 University of Chicago Legal Forum 217–60.
Hazard, Jr., Geoffrey C. 1989. ‘Authority in the Dock,’ 69 Boston University Law Review 469–76.
Hughes, Justin. 2016. ‘Motion Pictures, Markets, and Copylocks,’ 23 George Mason Law Review 941–66.
Hurt, Robert M., and Robert M. Schuchman. 1966. ‘The Economic Rationale of Copyright,’ 56 The American
Jaffe, Adam B., and Josh Lerner. 2004. Innovation and Its Discontents: How Our Broken Patent System Is
Endangering Innovation and Progress, and What to Do About It. Princeton, NJ: Princeton University Press.
James, Simon, John Hasseldine, Peggy Hite, and Marika Toumi. 2001. ‘Developing a Tax Compliance Strategy
for Revenue Services,’ 55 Bulletin for International Fiscal Documentation 158–64.
Kaplow, Louis. 1986. ‘Private versus Social Costs in Bringing Suit,’ 15 Journal of Legal Studies 371–85.
Kaplow, Louis. 1995. ‘A Model of the Optimal Complexity of Legal Rules,’ 11 Journal of Law, Economics and
Katz, Avery. 1990. ‘The Effect of Frivolous Lawsuits on the Settlement of Litigation,’ 10 International Review
Kitch, Edmund W. 1999. ‘Can the Internet Shrink Fair Use,’ 78 Nebraska Law Review 880–90.
Ku, Raymond Shih Ray, Jiayang Sun, and Yiying Fan. 2009. ‘Does Copyright Law Promote Creativity: An
Empirical Analysis of Copyright’s Bounty,’ 62 Vanderbilt Law Review 1667–746.
Landes, William M., and Richard A. Posner. 1975. ‘The Private Enforcement of Law,’ 4 Journal of Legal Studies
1–46.
Landes, William M., and Richard A. Posner. 1979. ‘Adjudication as a Private Good,’ 8 Journal of Legal Studies
235–84.
Lanjouw, Jean O., and Josh Lerner. 2001. ‘Tilting the Table? The Predatory Use of Preliminary Injunctions,’
44(2) The Journal of Law and Economics 573–603.
Lanjouw, Jean O., and Mark Schankerman. 2004. ‘Protecting Intellectual Property Rights: Are Small Firms
Handicapped?,’ 47 Journal of Law and Economics 45–74.
Lederman, Leandra. 2003. ‘The Interplay Between Norms and Enforcement in Tax Compliance,’ 64 Ohio State
Lemley, Mark A. 2001. ‘Rational Ignorance at the Patent Office,’ 95 Northwestern Law Review 1–34.
Lemley, Mark A. 2008. ‘Ignoring Patents,’ 2008 Michigan State Law Review 19–34.
Lemley, Mark A. 2011. ‘Is the Sky Falling on the Content Industries?,’ 9 Journal on Telecommunications & High
Technology Law 125–36.
Lemley, Mark A., and Douglas Melamed. 2013. ‘Missing the Forest for the Trolls,’ 113 Columbia Law Review
2117–90.
Lemley, Mark A., and Anthony R. Reese. 2004. ‘Reducing Digital Copyright Infringement without Restricting
Innovation,’ 56 Stanford Law Review 1345–426.
Lemley, Mark A., and Anthony R. Reese. 2005. ‘A Quick and Inexpensive System for Resolving Digital
Copyright Disputes,’ 23 Cardozo Arts & Entertainment Law Journal 1–20.
1991–2050.
Lerner, Joshua, 1995. ‘Patenting in the Shadow of Competitors,’ 38 Journal of Law and Economics 463–96.
Levmore, Saul. 2002. ‘Two Stories about the Evolution of Property Rights,’ 31 Journal of Legal Studies 421–52.
Libecap, Gary D., and James L. Smith. 2002. ‘The Economic Evolution of Petroleum Property Rights in the
United States,’ 31 Journal of Legal Studies 589–608.

Lichtman, Douglas. 2003. ‘Copyright as a Rule of Evidence,’ 52 Duke Law Journal 683–744.
Lichtman, Douglas, and William M. Landes. 2003. ‘Indirect Liability for Copyright Infringement: An
Economic Perspective,’ 16 Harvard Journal of Law & Technology 395–410.
Litman, Jessica. 1996. ‘Revising Copyright Law for the Information Age,’ 75 Oregon Law Review 19–48.
Long, Clarissa. 2002. ‘Patent Signals,’ 69 University of Chicago Law Review 625–80.
Long, Clarissa. 2006. ‘Dilution,’ 106 Columbia Law Review 1029–78.
Lueck, Dean. 2002. ‘The Extermination and Conservation of the American Bison,’ 31 The Journal of Legal
Studies 609–52.
Maskus, Keith E., and Jerome H. Reichman. 2004. ‘The Globalization of Private Knowledge Goods and the
Privatization of Global Public Goods,’ 7 Journal of International Economic Law 279–320.
Mazzeo, Michael J., Jonathan Hillel, and Samantha Zyontz. 2013. ‘Explaining the “Unpredictable”: An
Empirical Analysis of U.S. Patent Infringement Awards,’ 2013 International Review of Law & Economics 58–72.
McGeveran, William. 2008. ‘Rethinking Trademark Fair Use,’ 94 Iowa Law Review 49–124.
Menell, Peter S. 1983. ‘A Note on Private versus Social Incentive to Sue in a Costly Legal System,’ 12 Journal of
Menell, Peter S. 2002. ‘Envisioning Copyright Law’s Digital Future,’ 46 New York Law School Law Review
63–200.
Analysis 1–60.
Merges, Robert P. 2009. ‘The Trouble with Trolls: Innovation, Rent-Seeking, and Patent Law Reform,’ 24
Miceli, Thomas J. 1993. ‘Optimal Deterrence of Nuisance Suits by Repeat Defendants,’ 13 International Review
Milgrom, Paul, and John Roberts. 1982. ‘Predation, Reputation, and Entry Deterrence,’ 27 Journal of Economic
Theory 280–312.
Miller, Shawn P. 2010. ‘Patent “Trolls”: Rent-Seeking Parasites or Innovation-Facilitating Middlemen?,’
accessed April 3, 2019 at http://papers.ssrn.com/abstract=1885538.
Miller, Shawn P. 2013. ‘Where’s the Innovation: An Analysis of the Quantity and Qualities of Anticipated and
Obvious Patents,’ 18 Virginia Journal of Law & Technology 1–58.
Parchomovsky, Gideon, and Alex Stein. 2012. ‘The Relational Contingency of Rights,’ 98 Virginia Law Review
1313–72.
Perel, Maayan, and Niva Elkin-Koren. 2016. ‘Accountability in Algorithmic Copyright Enforcement,’ 19
Plant, Arnold. 1934. ‘The Economic Aspects of Copyright in Books,’ 1 Economica 167–95.
Port, Kenneth L. 2008. ‘Trademark Extortion: The End of Trademark Law,’ 65 Washington and Lee Law Review
585–638.
Posner, Richard A. 2005. ‘Intellectual Property: The Law and Economics Approach,’ 19 Journal of Economic
Posner, Richard A. 2014. Economic Analysis of Law. New York, NY: Wolters Kluwer Law & Business. 9th
edition.
Posner, Richard A., and William A. Landes. 2003. The Economic Structure of Intellectual Property Law.
Cambridge, MA: Belknap Press.
PwC. 2012. Patent Litigation Study.
Priest, George L. 1982. ‘Regulating the Content and Volume of Litigation: An Economic Analysis,’ 1 Supreme
Court Economic Review 163–83.
Ramsey, Lisa P. 2003. ‘Descriptive Trademarks and the First Amendment,’ 70 Tennessee Law Review 1095–176.
Ridgway, William E. 2006. ‘Revitalizing the Doctrine of Trademark Misuse,’ 21 Berkeley Technology Law
Journal 1547–88.
Risch, Michael. 2012. ‘Patent Troll Myths,’ 42 Seton Hall Law Review 457–500.
Risch, Michael. 2014. ‘Licensing Acquired Patents,’ 21 George Mason Law Review 979–1014.
Rose-Ackerman, Susan, and Mark Geistfeld. 1987. ‘The Divergence between Social and Private Incentives to
Sue: A Comment on Shavell, Menell, and Kaplow,’ 16 Journal of Legal Studies 483–91.
Rosenberg, David, and Steven Shavell. 1985. ‘A Model in Which Suits are Brought for Their Nuisance Value,’ 5
International Review of Law and Economics 3–13.
Rothman, Jennifer E. 2007. ‘The Questionable Use of Custom in Intellectual Property,’ 93 Virginia Law Review
1899–982.
Sag, Matthew. 2015. ‘Copyright Trolling, an Empirical Study,’ 100 Iowa Law Review 1105–48.
Sag, Matthew, and Jake Haskell. 2018. ‘Defense against the Dark Arts of Copyright Trolling,’ 103 Iowa Law
Review 571–662.
Salop, Steven C., and David T. Scheffman. 1986. ‘Raising Rivals’ Costs,’ 16 Journal of Reprints for Antitrust

Samuelson, Pamela, and Tara Wheatland. 2009. ‘Statutory Damages in Copyright Law: A Remedy in Need of
Reform,’ 51 William and Mary Law Review 439–512.
Samuelson, Pamela, Phil Hill, and Tara Wheatland. 2013. ‘Statutory Damages: A Rarity in Copyright Laws
Internationally—But for How Long?,’ 60 Journal of the Copyright Society 529–53.
Schlosser, Sarah Mayhew. 2001. ‘The High Price of (Criticizing) Coffee: The Chilling Effect of the Federal
Trademark Dilution Act on Corporate Parody,’ 43 Arizona Law Review 931–64.
Schwartz, David L. 2012. ‘The Rise of Contingent Fee Representation in Patent Litigation,’ 62 Alabama Law
Review 335–88.
Schwartz, Warren F., and Gordon Tullock. 1975. ‘The Costs of a Legal System,’ 4 Journal of Legal Studies
75–82.
Shapiro, Carl. 2001. Navigating the Patent Thicket: Cross Licenses, Patent Pools and Standard-Setting.
Shavell, Steven. 1982. ‘The Social Versus the Private Incentive to Bring Suit in a Costly Legal System,’ 11 Journal
Shrestha, Sannu K. 2010. ‘Trolls or Market-Makers: An Empirical Analysis of Nonpracticing Entities,’ 110
Columbia Law Review 114–60.
Sichelman, Ted M. 2010. ‘Commercializing Patents,’ 62 Stanford Law Review 341–413.
Siegelman, Peter, and Joel Waldfogel. 1999. ‘Toward a Taxonomy of Disputes: New Evidence through the Prism
of the Priest/Klein Model,’ 18 Journal of Legal Studies 101–30.
Smith, Kent W. 1990. ‘Integrating Three Perspectives on Noncompliance: A Sequential Decision Model,’ 17
Criminal Justice & Behavior 350–69.
Syed, Talha, and Bracha Oren. 2014. ‘Beyond the Incentive-Access Paradigm: Product Differentiation &
Tehranian, John. 2007. ‘Infringement Nation: Copyright Reform and the Law/Norm Gap,’ 2007 Utah Law
Review 537–50.
Tirole, Jean. 1994. The Theory of Industrial Organization. Cambridge, MA: The MIT Press. 
Tyler, Tom R. 1990. Why People Obey the Law. Princeton, NJ: Princeton University Press.
Waldfogel, Joel. 2012. ‘Music Piracy and Its Effects on Demand, Supply, and Welfare,’ 12 Innovation Policy and
the Economy 91–110.
Waldman, Michael. 1987. ‘Noncooperative Entry Deterrence, Uncertainty, and the Free Rider Problem,’ 54 The
Review of Economic Studies 301–10.
Wenzel, Michael. 2005. ‘Motivation or Rationalization? Causal Relations Between Ethics, Norms and Tax
Compliance,’ 26 Journal of Economic Psychology 491–508.
Williams, Philip L., and Ross A. Williams. 1994. ‘The Cost of Civil Litigation: An Empirical Study,’ 14
International Review of Law and Economics 73–86.
Wu, Tim. 2008. ‘Tolerated Use,’ 31 Columbia Journal of Law & the Arts 617–36.
Yu, Peter K. 2012. ‘Enforcement, Enforcement, What Enforcement?,’ 52 IDEA Journal 239–84.
Case List
Capitol Records, Inc. v. Thomas-Rasset, 692 F.3d 899 (8th Cir. 2012).
Sony BMG Music Entm’t v. Tenenbaum, 719 F.3d 67 (1st Cir. 2013).
17 U.S.C. § 504(c) (1976).

35 U.S.C. § 299 (2012).
15 U.S.C. § 1125(c), 1127 (1995).

16. Economic analysis of intellectual property notice
and disclosure
Peter S. Menell* 23
Contents
I. Notice Functions
A. Knowledge Dissemination
B. Signaling
C. Freedom to Operate
D. Cumulative Creativity
II. Notice Frameworks
A. Costs
B. Rules and Institutions
C. Incentives and Pathologies
1. Costs of describing and disclosing intangible resources
2. Notice externalities
3. Interplay with remedies
4. Subversion of public policy
III. Policy Analysis
A. Utility Patent
1. Disclosure timing
2. Claim clarity
3. Ownership transparency
4. Systematic judicial procedures for determining patent scope
5. Internalizing and counteracting adjustments
6. Fostering licensing
7. Depropertization
B. Copyright
1. Mandatory registration
2. Optimal specificity
3. Preclearance institutions
4. Systematic judicial procedures for determining copyright scope
5. Substantive reforms
6. Voluntary licensing regimes
7. Depropertization
* Koret Professor of Law and Director of the Berkeley Center for Law and Technology,
Berkeley School of Law, University of California. I am grateful to Michael Meurer for his collabo-
ration on related work and to Amit Elazari, Concord Cheung, and Megan McKnelly for research
assistance.
424

Economic analysis of intellectual property notice and disclosure 425
C. Design Patent
D. Trademark
E. Trade Secret
References
Notice of intellectual property content, ownership, boundaries, scope of rights (and

limitations), enforcement institutions, and remedial consequences plays a central role in
resource planning and other economic and social functions. Developers of resources—
whether tangible or intangible—ideally seek to know all potential encumbrances in
advance of investing their time, effort, and money into a project. Just as contracting par-
ties would ideally operate with fully specified contracts, resource developers would like to
know all contingencies affecting their choices before they incur development costs. Such
information includes the boundaries and holders of neighboring properties as well as the
extent to which those rights affect ‘neighboring’ resources and activities—for example,
nuisance law for non-trespassory activities, copyright law’s fair use doctrine, and patent
law’s doctrine of equivalents.
Legislatures, jurists, and intellectual property scholars have long looked to real property
law and institutions as models and a toolbox for designing intellectual property rules and
institutions (Van Houweling, 2019). Landowners record property interests with govern-
ment registries, which inform the public of property boundaries, owners, restrictions, and
encumbrances. Prospective purchasers, neighbors, land developers, contractors, lenders,
property tax assessors, zoning authorities, and other interested parties consult these
records to plan resource use and fund government activities. Thus, public notice of real
estate ownership functions as a critical governance institution. It has become more reliable
and less costly over time as a result of advances in surveying technologies, geographical
information systems, and publicly available computer databases.
There is much to be gained from looking to real property notice rules and institutions
in analyzing notice rules and institutions for intangible resources. Inventors and authors
seeking to appropriate a return on their investments wish to attract investors, users, and
consumers. Furthermore, developers of new technologies and works of authorship need
to determine whether their projects can be commercialized without running afoul of the
rights of others. Disclosure of technological knowledge and creative works promotes
further development of knowledge.
Yet notice in the intellectual property realm introduces distinctive challenges (Van
Houweling, 2019; Menell, 2011). Although patent and copyright law principally rely
upon market mechanisms to promote innovation and expressive creativity, various tech-
nological, economic, and social factors require a nuanced mix of regulatory adjustments
as well as government oversight. Trademark law serves primarily as a complement to free
competition by protecting source-identifying symbols against confusing use by others.
Thus, notice of the connection between such indicia and associated products and services
plays a critical role in the functioning of trademark law. In contrast to patent, copyright,
and trademark law, trade secret protection is destroyed by public disclosure. Therefore,
it depends upon those obtaining access to trade secrets to keep proprietary information
confidential.
Unlike land, intangible resources cannot typically be mapped onto two-dimensional
grids. Inventors draft patent claim boundaries using words, which introduces inherent

linguistic ambiguities. Inventors, authors, artists, and musicians draw upon the works of
others in inventing new technologies and expressing their own creativity. Marketers often
seek to associate common terms and shapes with their wares and services as a means
to attract consumers. Thus, the various intellectual property regimes face substantial
challenges in informing the public about the contours and limits of intellectual property
rights.
This chapter examines the function, design, and economic effects of intellectual
property notice rules and institutions. Section I discusses the principal functions served
by notice of intellectual property notice and disclosure. Section II surveys the economic
frameworks for analyzing intellectual property notice costs, institutions, pathologies, and
policy approaches. Section III traces the ramifications of these frameworks for patent,
copyright, trademark, and trade secret regimes, highlighting reforms for promoting the
efficacy of these regimes as well as broader social goals.
I. NOTICE FUNCTIONS
Notice of the content, ownership, boundaries/scope, and legal standards governing

intellectual property protection serve several functions: (A) knowledge dissemination; (B)
signaling; (C) freedom to operate; and (D) cumulative creativity. These functions vary
among the various stakeholders: inventors/creators, investors, competitors, complemen-
tors (aftermarket entrants/participants), potential collaborators (including alliances and
standard setting organizations (SSOs)), purchasers, employees/contractors, potential
whistleblowers, the general public, and the government.
A. Knowledge Dissemination
In contrast to real and chattel (personal) property protections, which flow from state
statutes and common law, patent and copyright law protections derive from the US
Constitution’s conferral of legislative power ‘To promote the Progress of Science and
useful Arts, by securing for limited Times to Authors and Inventors the exclusive Right
to their respective Writings and Discoveries’ (U.S. Const. art I, § 8, cl. 8). The Framers
believed that dissemination of scientific and technological knowledge and creative
expression were critical to encourage advancements and the flourishing of society. Thus,
knowledge dissemination serves a central role in the provision of patent and copyright
protection (Menell, 2007a).
Such dissemination is often characterized as the quid pro quo for the grant of patent
protection, which ‘allows the [USPTO] to examine applications effectively; courts to
understand the invention, determine compliance with the statute, and construe the claims;
and the public to understand and improve upon the invention and to avoid the claimed
boundaries’ (Ariad Pharmaceuticals v. Eli Lilly and Co., 598 F.3d 1336, 1345 (Fed. Cir.
2010)). Burk (2008) suggests that the disclosure function of the patent system can be
conceived more capaciously as a means for codifying what would otherwise be tacit
knowledge. In order to obtain a patent, the applicant is required to describe ‘the invention,
and of the manner and process of making and using it, in such full, clear, concise, and
exact terms as to enable any person skilled in the art to which it pertains, or with which it

is most nearly connected, to make and use the same . . .’ (35 U.S.C. § 112(a)). The Patent
Act further requires that the applicant conclude the patent specification ‘with one or
more claims particularly pointing out and distinctly claiming the subject matter which
the inventor . . . regards as the invention’ (35 U.S.C. § 112(b)).
Several scholars question whether the patent system serves this function well. Devlin
(2010) contends that patent disclosure is largely ineffective and ought to take a back
seat to promoting ex ante incentives to innovate and commercialization. Fromer (2009)
argues that knowledge dissemination is central to the effective functioning of the patent
system, although in desperate need of invigoration. Based on a survey of nanotechnology
researchers and several case studies, Larrimore Ouellette (2012) finds that patents provide
valuable technical information for many early-stage researchers, although the value of
patent disclosures could be improved to better align patent system disclosure with general
scientific disclosure norms. Seymore (2010) proposes the greater use of working examples
to enhance the teaching function of patent specifications.
Although copyright law also derives from the same constitutional clause, most expres-
sive works typically function as disclosure.1 Nonetheless, the scope of such protection
is often difficult to discern due to the many limitations on copyright protection (Menell,
2016b).
Similarly, trademarks are also readily apparent on their face.
By contrast, trade secret law is premised on keeping knowledge secret. Unlike patent
protection, however, trade secrecy does not afford the holder exclusive rights, only rights
against misappropriation through improper means or breach of confidence. Therefore,
those who reverse engineer or independently develop the knowledge are free to use and
disclose the information. Therefore, patent law provides stronger protection, but subject
to the quid pro quo of public disclosure.
B. Signaling
Intangible resource developers often seek to disseminate knowledge as a means of

signaling promising investment opportunities and business collaborations. As President
Abraham Lincoln eloquently captured, the patent system ‘added the fuel of interest to the
fire of genius[] in the discovery and production of new and useful things’ (Lincoln, 1953).
The ‘fuel of interest’ refers to venture finance.
To attract the capital investment necessary to pursue and commercialize technological
inventions, inventors must disclose their ideas to potential investors. As Arrow (1962)
recognized, such investors will want to know the details of the technology prior to making
a financial commitment. But once exposed to the knowledge, they have the information
goods and can proceed, in the absence of intellectual property protection, to commercial-
ize the technology (or expressive work) without compensating the inventor (or author).
Patent and copyright protection can solve Arrow’s Information Paradox by establishing
rights in the knowledge. Patents provide a means for credibly publicizing information
1
Computer software can be an exception, where the know-how underlying the software can
be obscured through release or registration of only the binary code and not human-decipherable
source code.

(Long, 2002). Patents can reduce informational asymmetries between inventors and
prospective funders. Lemley (2000) and Conti et al. (2013) highlight the signaling role in
modern venture capital. Acquisition of a patent signals the potential for an excludable
intangible resource as well as the fact that the Patent Office deems the claimed invention
novel, nonobvious, and adequately disclosed.
Contractual agreements, such as nondisclosure agreements (NDAs) and agreements to
compensate the inventor for use of an idea, can also solve Arrow’s Information Paradox.
(Burstein, 2012; Anton and Yao, 2002). Courts have developed various doctrines for this
purpose (Smith v. Dravo, 203 F.2d 369 (7th Cir. 1953) (recognizing implied confidential
relationships as a basis for trade secret protection); Restatement (Third) of Unfair
Competition § 41; Desny v. Wilder, 46 Cal. 2d 715 (Cal. S. Ct. 1956) (recognizing implied-
in-fact contract); Blaustein v. Burton, 88 Cal. Rptr. 319 (Cal. Ct. App. 1970) (extending
Desny doctrine)). Nonetheless, the costs of negotiating express agreements can be pro-
hibitive. Many venture capitalists refuse to sign NDAs. Furthermore, start-up inventors
often have difficulty gaining access to major funders and established corporations. Patent
protection enables them to establish protection for implementations of ideas and signal
the potential to exclude competitors.
Inventors might also use strategic disclosure of knowledge to communicate with
potential business partners and competitors. The patent system provides a database for
technology and finance enterprises to study the intangible resource landscape and explore
business opportunities. Patents provide assets that can be bought, sold, and cross-licensed
to realize gains from trade and collaboration.
Patents also play a critical role in standard setting activities, which are critically
important to network industries (Menell, 2019). SSOs seek to lessen the tension between
employing the best technological solutions in industry standards and ensuring wide-
spread access to standards by requiring members to disclose standard-essential patents
(SEPs) and license them on fair, reasonable, and non-discriminatory (FRAND) terms
(Contreras, 2019). Thus, disclosure of patents plays a critical role in the development of
technology standards, which can be critical to promoting interoperability, coordination,
and collaboration.
C. Freedom to Operate
Like real property, patents and copyrights can encumber competitors’ freedom to operate.
Inventors risk substantial liability, including injunctions barring their operations, even if
they unknowingly infringe patented technologies. Hence, inventors and investors often
engage in freedom to operate searches of patent records before introducing products and
services. Thus, patent disclosures and the searchability of Patent Office records play a
critical role in technology ventures. They also come into play during due diligence searches
accompanying corporate mergers and acquisitions. Furthermore, recordation of transfers
of patent assets protects purchasers of patents from the risk of loss of ownership due to
subsequent purchases (Menell, 2007b). The patent system’s recordation of transfer regime
largely mirrors transfer recordation of real estate and chattels.
Copyright law also operates to constrain freedom to operate. Unlike patent law,
however, copyright liability requires actual copying as an element of infringement.
Therefore, independent development serves as a defense to copyright infringement

(Sheldon v. Metro-Goldwyn Pictures Corp., 81 F.2d 49, 54 (2d Cir. 1936)). This defense,
however, is unlikely to succeed where a later developed work is substantially similar to a
widely available expressive work. Subconscious copying is actionable and courts will infer
actual copying where there is a high objective basis for access to the protected work and
probative similarity between the works (Bright Tunes Music Corp. v. Harrisongs Music,
Ltd., 420 F. Supp. 177, 181 (S.D.N.Y. 1976); Sheldon v. Metro-Goldwyn Pictures Corp.,
81 F.2d 49, 54 (2d Cir. 1936) (L. Hand, J.) (‘Yet unconscious plagiarism is actionable quite
as much as deliberate’)).
Copyright registration is not required for protection, although it is recommended
and commonly obtained to facilitate enforcement and expand potential remedies.
Nonetheless, Copyright Office registration records can be difficult to track, especially for
works that lack clearly marked titles (such as photographs). Furthermore, older works
may have forfeited protection through publication without proper notice or failure to
renew. Thus, conducting freedom to operate searches for copyrighted works can be costly
and unreliable.
As with patents (and real and chattel property interests), recordation of transfers of
copyright assets, including particular rights (such as the right to prepare a derivative
work), protects purchasers of copyright interests and financiers from the risk of loss of
ownership due to subsequent transfers. The copyright system’s recordation of transfer
regime largely mirrors transfer recordation of real and chattel property (Menell, 2007b).
Film financiers, for example, conduct careful searches of the various copyrighted works—
novels, scripts, musical compositions, sound recordings, images, props—featured in film
projects. Tracing the provenance of these works can be a complex task.
Companies introducing new goods and services risk trademark liability for using terms
that are confusingly similar to established trademarks. Thus, before marketing, manu-
facturers routinely conduct trademark searches. Such searches are complicated by the
fact that trademark registration is optional in the United States. Both federally registered
trademarks and common law trademarks are enforceable.
D. Cumulative Creativity
Intellectual property notice plays a vital role in promoting cumulative innovation and
creativity. (Menell and Scotchmer, 2007). Rather than re-invent the wheel or create an
entirely new genre, inventors and creators can usefully build upon the work of those who
have come before them. Intellectual property licenses as well as building on unprotected
elements of the prior art reduces wasteful duplicative research and promotes collaboration.
Thus, clear and accessible notice of prior knowledge, the scope of rights, and the person
or persons with whom a cumulative creator must license rights promotes transactional
efficiency, productive collaboration, technological progress, and cumulative creativity.
II. NOTICE FRAMEWORKS
While conceptually straightforward, the provision of resource notice, especially of

intangible resources, can be complex. As background for analyzing intellectual property
notice regimes, this section contrasts notice costs for tangible and intangible resources,

summarizes intangible resource notice rules and institutions, and analyses the incentives
(and disincentives) to provide accurate, accessible, and timely notice.
A. Costs
Resource notice can entail five types of social costs: (1) describing and disclosing resource
ownership and scope; (2) determining owners of potentially conflicting property rights;
(3) ascertaining boundaries of those properties; (4) assessing the scope of those property
rights and the ramifications of infringing those rights; and (5) dispute resolution costs
(Menell, 2016b; Menell and Meurer, 2013). The first type of notice cost is borne by the
resource claimant. The second, third, and fourth types—what can be characterized as
tracing costs—are borne by neighboring developers or, in the case of intangible resources,
cumulative creators. The fifth is shared by resource claimants and those who might violate
the claimants’ rights.
In some contexts, such as conventional real estate development, these costs are manage-
able because the resources are easily described, the number and complexity of ‘nearby’
property rights is small, and reliable publicly accessible registries record the resources.
Thus, the cost of evaluating relevant property rights is a small fraction of the value of
the development project. Notice costs rise when the resources become more difficult to
describe (as is the case with many intangible resources), the number of ‘neighbors’ grows
(making it more difficult to identify relevant counterparties), and the legal rights and
remedies are more difficult to evaluate. Notably, notice costs also tend to rise with the
value of the development project because the returns to enforcement of property rights
tends to grow with value. Unsuccessful projects rarely attract enforcement whereas com-
mercially successful projects often do. The familiar Hollywood quip—‘where there’s a hit,
there’s a Writ’—reflects this dispute selection phenomenon (Menell and Depoorter, 2014).
Blockbusters attract more lawsuits than box office flops. Inefficient notice regimes raise
development costs and generate wasteful litigation.
When a land or factory developer sets out to build a structure or a business considers
introducing a new product, they would ideally like to know the full range of potential
impediments to their projects. The land developer does not want to encroach neighbors’
lands. The factory developer does not want to cause a nuisance. And the product developer
does not want to infringe patents, copyrights, or trademarks. Each of these developers will
likely retain a resource specialist (e.g., land surveyor, land use consultant, environmental
lawyer, patent adviser, copyright clearance professional, trademark consultant) to deter-
mine whether the proposed project invades or infringes the rights of others. They would
like to know, to the extent feasible, the rights of others before sinking their investment.
But like the contract theorist’s fully specified contract, the economist’s world of costless
transactions, and the physicist’s vacuum, full knowledge of all potential constraints on
resource development might not be achievable in the real world. Neither property law
nor property deeds can fully specify the scope of all potential non-trespassory invasions
(nuisances). Nor can copyright law or copyright registrations fully delineate all potential
fair uses of the protected work. Yet this baseline provides a valuable construct for assess-
ing the sources of notice failure and a goal for policy intervention.
We can usefully divide the range of notice issues into two categories: (1) ‘deed’
information—factual information relating to the actual resource boundaries and the

identity of the owner(s); and (2) scope of property rights—information relating to the
legal scope of the rights pertaining to the particular resource claim (such as nuisance, fair
use, and so on).
The first category comports with a more conventional definition of notice information
for land and chattels in large part because the rights associated with tangible property
are relatively straightforward. Land and chattel boundaries are physical, measurable,
and directly observable. Default rules, such as the ad coelum doctrine which affords a
landowner control over the three-dimensional column defined by the deed, provide added
clarity. Standardization and modularity of real estate further enhance boundary precision
(Merrill and Smith, 2000). The adverse possession doctrine and good faith improver
statutes provide mechanisms to resolve disputes over clouded title (Dwyer and Menell,
1998, pp. 76–84, 292–6). Uncertainty about the scope of tangible property resources rarely
arises in modern experience2 outside of nuisance issues, which are resolved through the
application of a murky balancing standard (Dwyer and Menell, 1998, pp. 296–318).
The tangible, geographic, and rivalrous characteristics of land provide relatively
reliable and inexpensive means for verifying potential real property conflicts (Smith,
2011; Merrill and Smith, 2000 (extolling the virtues of modularity in property systems)).
Thanks to advances in land surveying technology, land claims can be and typically are
specified with numerical geographic precision. Signs, fences, and other structures are also
available to communicate boundary and ownership information efficiently. Moreover,
most rectangular land parcels have relatively few abutting property, often no more than
four to eight, thereby limiting the number of abutting neighbors with whom most land
developers must interact in planning projects and making it relatively easy to identify
owners of neighboring parcels. Furthermore, mandatory public recording systems and
related institutions (such as finance and title insurance) promote easy availability of land
notice information.3
By contrast, the boundaries and scope of intangible rights are frequently far more
difficult to establish, mark, and communicate. Ideas, which form the basis for intangible
2
Prior to the establishment of the Public Land Survey System (based on a rectangular grid) in
the western territories shortly after the Revolutionary War, surveying was relatively expensive and
imprecise, leading to frequent boundary disputes on the American frontier. The lack of an overall
framework for coordinating land demarcation, the vagueness of boundaries, and the irregular
shapes of parcels were major contributors to property disputes (Libecap and Lueck, 2011a, 2011b).
3
This is not to suggest that notice problems do not arise in modern real property settings
(Heller, 2008). The principal problems, however, relate to assessing permissible uses of land in more
densely populated communities. Neighbors can object to development projects that negatively
affect the value of their land. Zoning seeks to encourage coordinated and compatible land use.
Nuisance law also addresses conflicting land use by granting neighboring land owners protection
against noxious uses of land. These rules introduce some uncertainty into land use planning.
But even so, notice costs are usually not severe and zoning institutions enable developers to
determine whether a development project passes muster before the major construction costs are
incurred. Furthermore, some uncertainty about permissible uses is not necessarily an obstacle to
early, efficient resolution of disputes among affected parties. Land developers can typically identify
relevant counterparties relatively easily. Uncertainty about liability by itself usually does not cause
bargaining failure. Rather, breakdowns tend to be driven by private information about a dispute or
divergent beliefs about the likely trial outcomes. Such breakdowns might be related to notice, but
are more commonly attributable to other factors.

resources, can be multi-dimensional and amorphous. Unlike land resources, intangible

resources cannot generally be mapped to two-dimensional grids.4 Patents generally rely
upon textual descriptions, which opens boundary drawing to interpretation. Furthermore,
patent claims often use ambiguous terminology. In addition, outside of a few markets
(such as some pharmaceutical products), there is not a one-to-one relationship between a
patent and a product. Inventors and companies frequently develop large patent portfolios
surrounding their products, further complicating the task of mapping protected intan-
gible resources (Parchomovsky and Wagner, 2005). Moreover, many inventors and some
companies patent technologies that they do not ultimately commercialize. Therefore, the
fact that no product embodies a technology does not mean that the intangible space is
not occupied. Only by searching the patent registry can a new entrant assess freedom
to operate. And even then, the fact that patent applications are typically not publicly
available until 18 months after filing creates unknowable infringement risks. Thus, a
smartphone or app developer must sift through a bewildering, opaque, multi-dimensional
minefield of patented technologies to assess freedom to operate. They might also need to
assess copyright protection of application program interfaces and trademark protection
for compatibility standards.
While the outer ‘boundaries’ of copyrighted works are readily apparent—for exam-
ple, a manuscript, painting, musical composition, computer program, architectural
drawing—the scope of copyright protection is complicated by the unprotectability of
elements within the larger work. Thus, it can be difficult to determine what aspects of
copyrighted work are not protected, and hence free for others to use. Copyright law’s
protection of original compilations (selection and arrangement) of unprotected elements
further complicates the determination of the scope of copyright protection. Design
patents pose similar challenges as their scope excludes the functional features of articles
of manufacture (Menell and Yablon, 2019).
Moreover, unlike physical resources, intangible resources are generally not rivalrous.
Unlike a plot of land or an ice cream cone, multiple people can share an idea (like a
wireless email system) or a story (like the adventures of a young wizard and his friends)
without interfering with the use of that idea or information by others. Hence, intangible
resource conflicts are far less observable than trespass upon land or theft of personal
property. An independent inventor of a wireless email system would not necessarily know
that another inventor claimed that technology, whereas a land owner can use fences, direct
observation, and cameras to investigate and monitor trespassory invasions. Even though
patented technologies are examined by the Patent Office and contained in registries, the
classification of patented technologies is typically far more difficult to search than land
registries. Furthermore, copyrights and trademarks need not be registered to garner
protection. In addition, copyright registries can be difficult to trace. While trademark
registries for word marks are relatively straightforward to search, the same cannot be said
for trade dress, graphic image marks, and design patents.
Finally, the legal rules establishing and limiting protection for intangible resources are
far more complex than those applicable to land and other tangible resources. In contrast to
4
Chemical compositions can be classified systemically by reference to standardized nomen-
clature, molecular structure, and the Periodic Table.

the broad, uniform, and intuitive land rights bundle, intellectual property regimes employ
far more variegated and limited bundles of rights (Menell, 2007c). These more complex
rights structures reflect a range of countervailing legislative and constitutional purposes:
(1) promoting cumulative creativity; (2) restricting monopoly exploitation; (3) avoiding
impairment of free expression; (4) accommodating access by underserved constituencies;
and (5) limiting overreaching by publishers. As a result, intellectual property regimes
employ numerous doctrines that involve complexity, balancing, and inherent subjectiv-
ity, such as: utility patent law’s doctrine of equivalents, exhaustion doctrine, durational
limits, remedial standards, and misuse doctrine; design patent law’s ornamentality/
non-functionality limitation; copyright’s infringement standard (substantial similarity of
protected expression), first-sale doctrine, termination of transfer provision, compulsory
licenses, statutory exemptions, fair use defense, and statutory damages regime; and trade-
mark law’s infringement standard (likelihood of confusion) and defenses. In addition,
intellectual property law deploys various complex channeling doctrines (idea/expression
dichotomy, useful article doctrine, functionality doctrine, preemption rules) to prevent
interference among the modes of intellectual property protection and harmonize them
with contract and antitrust law.
B. Rules and Institutions
Notice of property interests can arise through the acts of resource owners as well as pursu-
ant to legal requirements. Furthermore, the nature of resources affects resource owners’
options for informing the public about the resource claim.
Land can be physically marked with signposts or fences to communicate ownership
claims. Land claims can and are legally required to be registered with local recording
offices using prescribed registration systems (grantor/grantee or geographic indices).
As noted above, the boundaries of land claims are typically specified with geographic
precision on tract indices with geographic coordinates (such as rectangular plat maps) or
through metes and bounds (measurements from particular corner point).5 Registration
of ownership claims (including security interests such as mortgages and liens) protects
bona fide purchasers of property interests from the claims of prior grantees who have not
recorded their interests. In addition, land title insurers and other private actors maintain
their own registries, some of which are private and others that are public or accessible for
a fee. Landowners can also provide notice of ownership by inspecting their property and
excluding trespassers.
Similarly, owners of chattels can mark their personal property and control access.
Pursuant to Article 9 of the Uniformed Commercial Code, states maintain registries of
security interests in personal property. Those who lend money secured against personal
5
For example, ‘beginning with a corner at the intersection of two stone walls near an apple
tree on the north side of Muddy Creek road one mile above the junction of Muddy and Indian
Creeks, north for 150 rods [a rod is a surveyor tool measuring 5.5 yards] to the end of the stone wall
bordering the road, then northwest along a line to a large standing rock on the corner of the prop-
erty now or formerly belonging to John Smith, thence west 150 rods to the corner of a barn near
a large oak tree, thence south to Muddy Creek road, thence down the side of the creek road to the
starting point’ (Metes and Bounds, Wikipedia, https://en.wikipedia.org/wiki/Metes_and_bounds).

property can register their claim with the state recording office, thereby providing notice
to subsequent security interest claimants.
Intangible resources vary in the practicality of marking, fencing, and physical control.
By distributing products protected by patents, copyrights, or trademarks with marking of
the intellectual property claims, intellectual property owners can provide direct notice to
those who come in contact with goods bearing intellectual property markings.
Furthermore, patent law, copyright law, and trademark law have public recordation
systems. The patent system requires examination of patent applications to obtain protec-
tion. Patent protection attaches upon issuance of the patent and entry of the issued patent
in the Patent Office’s official, publicly available records. Notice of the patent is published
in the Official Gazette of the United States Patent and Trademark Office (USPTO). In
addition, the patent system encourages marking of patented products by authorizing
recovery of damages from infringers for up to six years of infringement prior to the date
of the lawsuit so long as the infringer was actually aware of the patent or if patented
products bore the patent number.
Notwithstanding mandatory patent examination and recordation of issued patents,
those searching patent records cannot be assured that they can find all patents affecting
freedom to operate due to the inherent imprecision of patent claims and limitations
of patent database classification. The digitization of patent records and availability of
Boolean search have improved access (Schwartz and Sichelman, 2019). Furthermore,
private entities have developed additional search tools, some of which are public or
accessible for a fee.
The only way to determine patent scope definitively is through adjudication. This can
arise through patent enforcement proceedings or where there is a substantial controversy
between parties having adverse legal interests of sufficient immediacy to warrant the issu-
ance of a declaratory judgment (MedImmune, Inc. v. Genentech, Inc., 549 U.S. 118, 127
(2007)). In such litigation, district courts typically conduct a pre-trial hearing, commonly
referred to as a Markman hearing (Markman v. Westview Instruments, Inc., 517 U.S.
370 (1996)), to construe the patent claims (Menell et al., 2010). Many district courts have
established through local rules and standing orders a detailed procedure for preparing for
and conducting Markman hearings (Northern District of California Patent Local Rules;
Menell et al., 2016, ch. 5.1). The claim construction ruling can provide the basis for sum-
mary judgment determinations. It also serves a critical role in trial preparation. Expert
technical witnesses and damages experts rely on the court’s claim construction ruling in
preparing their reports. In jury trials, the Markman order also serves as a critical part of
the jury instructions, on which the jury bases its infringement and validity determinations.
Although copyright protection arises without formal registration of claims, US
copyright law encourages registration through procedural, evidentiary, enforcement, and
remedial inducements (Menell et al., 2018, vol. II, ch. IV). Nonetheless, searching the
Copyright Office’s registry can be expensive, difficult, and time-consuming. Those seeking
to determine the copyright status must know how to identify the work. Works registered
prior to 1978 can only be found in the Copyright Public Records Reading Room. For
post-1978 registrations, the Copyright Office provides an online database for identifying
registrations, but searches of records can be expensive. Some private registries exist for
particular copyright industries (such as music publishing). A few firms maintain private
databases of copyright ownership information and conduct searches.

The Music Modernization Act of 2018 addresses this collective action problem. This
legislation directs the U.S. Copyright Office to designate a mechanical licensing collective.
Among its principal responsibilities will be to ‘establish and maintain a database contain-
ing information relating to musical works (and shares of such works) and, to the extent
known, the identity and location of the copyright owners of such works (and shares
thereof) and the sound recordings in which the musical works are embodied.’ (17 U.S.C.
§ 115(d)(3)(E)(i)). The database shall include the title of the work, the copyright owner(s)
and ownership percentage(s), contact information for the copyright owner (if known),
and ‘to the extent reasonably available,’ ‘the international standard musical work code’ for
the work, and ‘identifying information for sound recordings in which the musical work is
embodied,’ including titles, featured artists, sound recording copyright owners, producers,
and international standard recording code (17 U.S.C. § 115(d)(3)(E)(ii)).
Even after a copyrighted work is traced to its owner and copyright subsistence is veri-
fied, the scope of copyright protection is often difficult to assess. Beyond determining
whether a work is registrable, the Copyright Office does not provide guidance on the scope
of protection of particular works. Unlike patent law’s pre-trial claim construction process,
the courts do not follow a standardized process for assessing the scope of copyright
protection.
Similarly, registration of trademarks is not required to establish protection. US trade-
mark protection arises from use of marks in commerce. The USPTO and state trademark
offices, however, play a significant role in recording valid trademarks. As with federal
copyright registration, federal recordation of trademarks offers various procedural,
evidentiary, enforcement, and remedial advantages (Menell et al., 2018, vol. II, ch. V). The
USPTO examines trademark applications (and intent-to-use applications) for compliance
with statutory requirements. After examination, the USPTO trademark applications
are published in the Official Gazette for opposition. Following that process (and use in
interstate commerce in the case of intent-to-use applications), valid trademarks are placed
on the Principal Register, which can be searched through an online portal.6 In addition,
several private search firms provide a range of search-related services.
C. Incentives and Pathologies
At first blush, there is good reason to believe that most resource claimants would prefer
clear boundaries and accessible notice information. First, claimants will want to secure
their property and will therefore seek to comply with applicable legal requirements. For
example, claimants will specify land boundaries using the Public Land Survey System,
write claim language in a patent application, or include drawings in a design patent
application. Second, claimants prefer to invest in boundary information when that will
increase the value of their resources—for example, through sale, financing, collaboration
with other developers, rental, or licensing. Land developers are likely to find that subdivi-
sions with clear boundaries and covenants will maximize the sale price of the land parcels
contained in the development. In addition, the developer may feel competitive pressure to
6
The USPTO also maintains a Supplemental Register which allows registration of descrip-
tive terms that have not yet acquired meaning among consumers.

make efficient investments in notice (Smith, 2003). Third, resource claimants sometimes
produce boundary information as a byproduct of efforts to deter trespass or infringement
via self-help measures like fences, locks, and encryption. Yet, there are circumstances in
which notice costs, opportunistic behavior, and unintended doctrinal rules override these
salutary instincts.
1. Costs of describing and disclosing intangible resources

As discussed above, intangible resources are often difficult to characterize. Unlike land,
which can typically be characterized precisely by reference to geographic coordinates
within a two-dimensional map, and chattels, which have objective physical boundaries,
intangible resources are multi-dimensional and judged by reference to the claimant’s
description, prior art, and other contextual factors.
Patented inventions are expressed as multi-dimensional, linguistic representations. The
boundaries and scope of patents are based on the patent claim language, the specification,
and prosecution history. US patents are typically claimed using a ‘peripheral’ format
whereby the drafter delineates the outer boundaries of the claimed invention. As an exam-
ple, U.S. Patent No. 5,205,473 claims an insulated recyclable beverage container sleeve as:
A recyclable, insulating beverage container holder, comprising a corrugated tubular member

comprising cellulosic material and at least a first opening therein for receiving and retaining
a beverage container, said corrugated tubular member comprising fluting means for contain-
ing insulating air; said fluting means comprising fluting adhesively attached to a liner with a
recyclable adhesive. (U.S. Patent No. 5,205,473 col. 4-5 (filed April 27, 1993))
Like other peripheral claims, this claim begins with a preamble (‘A recyclable, insulating
beverage container holder’) followed by a transitional phrase (‘comprising’) and the
claim body setting forth the claim restrictions (or elements). The open transitional phrase
‘comprising’ signifies that the patentee claims all structures containing the claim restric-
tions and anything else (Menell et al., 2010, pp. 759–60).
In general, patent claims are not limited to the particular embodiments or prophetic
examples except in particular circumstances, such as means-plus-function claim forms (35
U.S.C. § 112(f)). And even in that circumstance, the claim element extends to ‘equivalents’
of the embodiments contained in the patent specification.
Exacerbating the challenge of claiming patented inventions with precision, the patent
system rewards the first applicant to file for protection.7 This encourages applicants to
file claims as early as possible, in some cases before the inventor has a complete grasp of
the invention. Such premature filing, or gun-jumping, contributes to the vagueness of
patent claims. This motivation, as well as the inherent ambiguity of language and other
strategic considerations discussed in the following sub-section, makes patent claim scope
especially difficult to assess (Anderson and Menell, 2014).
The scope of copyright protection suffers from a somewhat different pathology. Like
7
Until March 15, 2013, the US based priority on a first-to-invent system, although inventors
seeking protection outside of the United States were subject to a first-to-file standard (Menell
et al., 2018, vol. I, ch. II). Hence, they jeopardized their non-US rights by not filing as early as
possible. And even for those filing only in the US, there were substantial difficulties in proving an
invention date prior to the filing date (Lemley and Chien, 2003).

chattels, copyrighted works have a physical form. To be protectable, a work of authorship

must be fixed in a tangible medium of expression (17 U.S.C. § 102(a)). Thus, the physical
boundaries of copyrighted works can be directly perceived. Yet, unlike a real estate parcel,
a copyrighted work does not exclude other authors and artists from copying elements
from within the work. And unlike patent law, which protects the claimed combination of
elements, copyright law allows for even the full combinations of elements to be used by
others under various circumstances. Copyright owners cannot exclude subsequent creators
from ‘any idea, procedure, process, system, method of operation, concept, principle, or
discovery’ (17 U.S.C. § 102(b)). Nor can they control elements of their own work that lack
originality. Nor can they bar any similarity or overlap, only those resulting in ‘substantial
similarity’ of protected expression. And even where another work is substantially similar
to protected expression, the copyright owner cannot prevent others from commenting
on, parodying, or sufficiently transforming their protected work (17 U.S.C. § 107). The
bounds of these limitations can be murky. Furthermore, the litigation process adds further
uncertainty. Outcomes depend on how a jury or judge perceives the works. Thus, even
though the public at large is on notice that a work is protected by copyright law, cumulative
creators cannot readily determine from registration materials or the underlying work which
particular elements or compilations of elements are protected and to what extent.
Whereas a real estate deed or patent claim can be analogized to a bounded parcel from
which interlopers are excluded, the copyright claim is more like a wedge, or in some cases
a thin slice, of Swiss cheese (Menell, 2016b). Cumulative creators have a clear under-
standing of the block or narrow slice as a whole. They are on notice that reproducing
the entirety of the block or slice risks infringement, although subject to a fair use claim.
But what the cumulative creator often cannot easily evaluate are the fermentation holes
in the cheese. They may freely operate within those holes. And even some compilations
of those holes can be fair game to the extent that the holes lack originality. Nonetheless,
uncertainty about the scope of what is not protected within a copyrighted work creates a
minefield for cumulative creators.
In theory, a notice system would apprise would-be developers or creators of which
resources or portions of resources are off-limits without authorization and where they
may freely operate. The copyright system does not, however, require the copyright owner
to designate those elements of their work that are unprotected.8 To do so would be
impractical. Almost all works reflect prior art ideas and concepts. For example, a graphic
work depicting the human body undoubtedly contains arms, legs, and torso, all of which
are found in human anatomy and hence are not entirely original to the artist’s expression.
8
The copyright registration process provides a limited exception. Registration forms require
applicants to identify preexisting material from which a work is derived and briefly explain the
original material added to the work for which copyright is claimed if the registered work is a
‘derivative work’ (Copyright Office, Form VA, line 6). The Copyright Office does not require
registrants to identify aspects of a claimed work which are not protected by copyright.
By contrast, the trademark registration directly addresses this issue by encouraging registrants to
disclaim unregistrable portions of marks. Trademark applicants include statements that ‘No claim is
made to the exclusive right to use “______” apart from the mark as shown’ (U.S. Patent and Trademark
Office, https://www.uspto.gov/trademark/laws-regulations/how-satisfy-disclaimer-requirement).
Common disclaimers refer to generic terms. Failure to include such disclaimers will lead the trademark
examiner to refuse registration. The disclaimer statement will appear on the registration certificate.

The lack of standardized procedures for determining the scope of copyright claims
further complicates the challenge of determining the freedom to operate surround-
ing copyrighted works. The long duration of copyright presents another difficulty
(Samuelson, 2016). The ownership of works will undoubtedly change over the life of
copyright protection, which extends 70 years beyond the life of the author or 95 years in
the case of works made for hire.
2. Notice externalities

Beyond the inherent difficulties of describing and disclosing intellectual property, resource
claimants can in several circumstances gain strategic advantages from obfuscating the
existence of property interests, the scope of their intellectual property protection, and the
identity of the resource owners. Menell and Meurer (2013) characterize the phenomena
as a form of externality in which the private and social interests diverge.9
Patent applicants are able to take advantage of the patent system by intentionally filing
ambiguous claims, updating claims during prosecution to target competitors’ products,
and leveraging the differential between ex ante negotiating (before investments have been
sunk) and ex post injunctive relief and monetary damages by obfuscating discovery of
their patents. They can also flood the Patent Office with patents on interrelated technolo-
gies, producing patent thickets (Shapiro, 2001; Heller and Eisenberg, 1998).
The first circumstance overlaps with the difficulty of fully specifying intangible
resource claims. The high cost of specifying patent boundaries inevitably results in
ambiguous claims. But it is the strategic benefits of stretching or contracting patent
claims after a patent issues that fuels an externality. Risch (2007, p. 188) explains that
‘[p]atent applicants have an incentive to allow claims to remain vague so that they can
mold the claims to fit the future product of a currently unknown, potential infringer
or to avoid invalidation if previously undiscovered prior art comes to light.’ Patent
prosecution guides preach these strategies (Sheldon, 2005, pp. 6–114 (including a section
entitled ‘Include Ambiguous Claims,’ which offers numerous ‘strategies’ for ‘intentionally
writ[ing] ambiguous claims’); Fish, 2007, pp. 7–35 (advising drafters to ‘[a]void . . . like the
plague’ claim language that clearly identifies the ‘gist of the invention’ or the ‘factor’ that
makes it ‘unique’); Wheeler, 2003, pp. 40–45 (advising prosecutors to draft broad claims,
minimally, and without disclaimers); Pressman, 2012, p. 204; McJohn, 2008, p. 971 (noting
that ‘experts in claim drafting advise patent drafters: “Do not define the terms used in
your claims; do not identify the category of invention in the preamble to the claims; do
not identify features of the invention as ‘important’ . . .”’; and noting that such claim
drafting has been described as a trend toward ‘intentional obscurity’ (footnote omitted))).
The ability to revise patent claims during prosecution provides a further opportunity to
9
Such divergences can also arise in the real estate context, but are less common, especially with
modern surveying methods and recordation systems. Libecap and Lueck (2011a) explain that under
metes and bounds land systems—which affords greater leeway in the specification of boundaries than
formal marked boundaries or rectangular survey systems—settlers had had at least two incentives
to leave boundaries vague and flexible: (1) ‘in a wilderness it was costly to locate precise boundaries
during the initial land claim, and hence difficult for the surveyor who followed to find those boundary
markers’; and (2) ‘given the lack of information about the location of the most desirable lands at the
time of the initial land entry, claimants did not want to be bound to absolute markers.’

game the patent system (Lemley and Moore, 2004; Chiang, 2010, p. 526). Applicants can
amend their claims as competitors introduce new products, thereby improving the likeli-
hood that the competitor falls within the scope of the claimed invention. Although the
courts seek to police such opportunism through the written description doctrine (Gentry
Gallery, Inc. v. Berkline Corp., 134 F.3d 1473 (Fed. Cir. 1998)), the incentive to update
claims as the market evolves remains strong.
Perhaps most significantly, patent applicants and patentees can gain tremendous
leverage over independent inventors through hiding the existence of patent claims. The
emergence of wireless email technology illustrates this type of notice externality, which
is commonly referred to as the patent troll problem. In the early 1990s, Research in
Motion (RIM) set out to develop mobile email devices. In mid-1995, the time that RIM
attracted substantial institutional and venture capital, there were no apparent ‘neighbors’
in this technological marketplace. Nor had any patents on wireless email communication
systems issued. Unfortunately for RIM, a non-practicing entity received a patent cover-
ing ‘Electronic Mail System with RF [Radio Frequency] Communications to Mobile
Processors and Method of Operation Thereof’ (U.S. Patent No. 5,436,960 (filed July 25,
1995)). Even though RIM’s business strategy was publicly known and successfully intro-
duced in its BlackBerry device in 1999, NTP lay low. NTP sent RIM a cease and desist
letter in early 2000. RIM conducted a review of the patents and reached the conclusion
that the NTP patents were invalid and not infringed by the BlackBerry.
In November 2001, NTP sued RIM for patent infringement in the Eastern District
of Virginia, a district known for its ‘rocket docket.’ RIM requested that the USPTO
re-examine the patents at issue. Following trial a year later, the jury found infringement
on all asserted claims and awarded damages of $33 million. The re-examination process
bogged down. Facing shutdown of its BlackBerry service, RIM settled the case for $612.5
million in March 2006 (Krazit and Broache, 2006).10 The problems reflected in this
case—identifying patent neighbors and ascertaining patent boundaries—have increas-
ingly plagued the information technology sector, adding substantially to the costs and
risks of technological resource development (Bessen and Meurer, 2008).
Copyright owners can also benefit from lying low. Whereas copyrighted works—such
as a photograph, a sound recording, or an audiovisual clip—might be readily available,
developers can encounter great cost and difficulty in identifying, locating, and contacting
the rights holders of such works. In the extreme, the work is an apparent ‘orphan’—
lacking an identifiable owner—thereby presenting the cumulative creator seeking to
incorporate the prior work with a stark choice: omit the props of uncertain provenance,
thereby resulting in a less satisfying project, or run the risk of an owner emerging and
threatening to block the entire integrated work (or extract a high price). Even though the
cost of licensing copyrighted works for use as film props or set design is often relatively
low before a film is shot because the director will have alternative props and designs
available, the threat of an injunction on the eve of theatrical release—after the film has
been produced and marketed—places the infringer over a barrel.
10
Most of NTP’s patent claims were ultimately invalidated in re-examination, but these deci-
sions came too late to benefit RIM (In re NTP, 654 F.3d 1279, 1289–90 (Fed. Cir. 2011); Bessen
and Meurer, 2014).

3. Interplay with remedies

Particular features of intellectual property remedies cause additional notice-related
pathologies. Patent law authorizes courts to enhance damages awards up to triple the
actual damages based significantly on whether the defendant willfully infringed the patent
(35 U.S.C. § 284; Halo Electronics, Inc. v. Pulse Electronics, Inc., 136 S. Ct. 1923 (2016)).
Hence, knowledge that a technology is patented can result in substantially greater patent
liability. Similarly, courts consider willfulness in exercising their discretion to award attor-
ney fees (35 U.S.C. § 285; Octane Fitness, LLC v. ICON Health & Fitness, Inc., 572 U.S.
545 (2014)). These ramifications discourage many technology companies, particularly
those in the digital technology fields, from searching for patents out of concern that such
knowledge can result in greater liability for infringement11 (Lemley, 2008; Note, 2007; Lee
and Cogswell, 2004; Lemley and Tangri, 2003).
Copyright remedies also have caused unintended consequences on cumulative
creativity and the deployment of advanced distribution platforms. US copyright
law has long used statutory damages to deter small-scale infringements that would
not be economical to enforce. This regime worked well to encourage restaurants,
bars, and other small establishments to obtain public performance licenses. With the
specter of large-scale infringement in the digital age, however, Congress substantially
ramped up the high end of statutory damages to $150,000 per work in cases of willful
infringement. This much higher level of damages exposes distribution platforms to
crushing liability. It can also expose cumulative creators who use multiple copyrighted
works in the mistaken belief that their uses qualify as fair use to potentially massive
liability.
4. Subversion of public policy

Trade secrets are the most pervasive form of intellectual property in the economy. Nearly
every enterprise seeks to protect information about its operations, strategy, technol-
ogy, funding, personnel, and customers. Employers of all types routinely require their
employees and contractors to sign restrictive nondisclosure agreements (NDAs) and
return confidential information upon their departure or completion of services. Without
such restrictions, these enterprises would jeopardize trade secret protection and risk
violating privacy and other laws. As noted earlier, trade secrecy protection functions as
an anti-notice regime.
While trade secret protection is generally viewed as a means for promoting techno-
logical innovation, overbroad NDAs barring employees and contractors from reporting
illegal activity to the government subvert public policy. Such restrictions on disclosure
outside of the commercial trade secrecy context undermine law enforcement, compliance
with civil rights, worker safety, environmental protections, and reporting of fraud and tax
evasion (Menell, 2017).
11
Companies in the pharmaceutical industry are less prone to this perverse effect because
the benefits from searching for patents and assessing patent validity and scope typically outweigh
the risk of enhanced exposure. Pharmaceutical patents are more easily found and typically cover
specific chemical structures.

III. POLICY ANALYSIS
Resource notice is a jointly produced public good. The government provides the foun-
dational legal infrastructure for establishing and enforcing resource rights. The market
operates on top of these rules and institutions. Private parties acquire, buy, sell, divide,
and develop such rights through private transactions. The interaction of the state and
private parties produces resource rights and notice of their existence, contours, and
extent. Thus, the government is uniquely situated to establish the authoritative registry
of resource claims. If infrastructure is adequate to prevent notice failure, as is typical in
the real property context, then notice externalities do not arise.
Many intellectual property claimants confront a difficult task of describing the
scope of their advance. Relatedly, developers of many intangible resource projects
face significant problems identifying potentially conflicting rights, ascertaining the
boundaries of those properties that they can find, locating the owners of potentially
conflicting properties, and assessing the scope of potentially conflicting rights. Unlike
county land recording offices, neither the USPTO nor the Copyright Office can look
up an intellectual property development tract and provide the names, addresses, and
contours of all ‘neighboring’ property owners. The courts cannot quiet title among
all intangible rights holders. The developer cannot typically gain pre-development
clearance of identified potentially conflicting rights short of negotiating a license.
Nor can developers easily obtain insurance against infringement risks. These problems
undermine the coordination that can be critical to resource planning and development.
They can hamper progress in technological innovation and expressive creativity by
encouraging some resource claimants to hide and obfuscate notice information. As a
result, the ratio of notice costs to overall development cost can be inefficiently high for
intangible resource projects.
The foregoing discussion highlights the importance of publicly accessible, easily
searchable registries. They provide the primary tool for tracing ownership and assessing
the scope of intangible resources. There remains, however, a difficult question regarding
the specificity of disclosure. Just as ‘[f]ully specifying a contract is not ordinarily pos-
sible’ and ‘[t]he benefits of nailing down a particular allocation of risk to cover most
extremely unlikely events will often not be worth the costs,’ (Coleman, 1992, p. 80) the
effort to specify ex ante all of the elements and exclusions of intangible resources would
be unduly costly. Thus, the optimal specificity of notice varies across the intellectual
property domain. Furthermore, the various ways of internalizing notice externalities
will depend on the contours of the intellectual property regime. The remainder of
this section traces the ramifications for the principal modes of intellectual property
protection.
A. Utility Patent
Disclosure of the content, boundaries, and ownership of utility patents can be improved
through prompt publication of patent applications, standardization and clearer explica-
tion of patent boundaries, and timely updating of ownership information. Patent notice
can be further enhanced by the provision of better search and mapping tools, calibration
of filing fees and substantive rules, and fostering of patent licensing.

1. Disclosure timing

Patent applications are generally not disclosed until 18 months after filing.12 Menell and
Meurer (2013) question the basis for such a rule, which prevents market entrants from
being able to assess the potential impediments to pursuing new products and services.
2. Claim clarity

There is critical trade-off between examination screening costs and the social costs of
poorly specified or examined patent claims. Menell (2013, 2015) advocates that the
Patent Office implement a standardized web-based claiming system aimed at clarify-
ing claim construction. As illustrated in Figure 16.1, such a system would streamline
patent examination, provide efficient tools for office actions, and generate a transparent
Preamble Limitation
comprising
Transitional
comprising consisting of
Clause
consisting essentially of
Claim
Limitation 1
novel
Means/Step + Function §112(f)
• hyperlink to corresponding structure, material, acts
Claim
Limitation 2
novel
Means/Step + Function §112(f)
• hyperlink to corresponding structure, material, acts
***
combination of limitations - novel
Glossary of Defined Terms:
Claim Term Claim Meaning
default general dictionary:

default technical dictionary:
Figure 16.1 Proposed claim filing template
12
US law allows applicants seeking only US protection to opt out of the 18-month publication
rule (35 U.S.C. § 122). Therefore, such applications remain secret for the full duration of patent
prosecution. If the patent is never issued, the application remains confidential.

rosecution history. Furthermore, a standardized presentation of patent claim elements

p
would promote the development of better search and mapping tools, reduce searching
and assessment costs, and facilitate compliance with early case management disclosures.
Such conventional uses of word processing and visualization technology could greatly
improve the accessibility of patent boundaries and knowledge at relatively low cost.
The claim filing template in Figure 16.1 would easily resolve various commonly
litigated claim scope issues: (1) whether the claim preamble is a limitation; (2) whether
a claim term is intended to be ‘means-plus-function’; (3) the precise ‘corresponding
structure, material, or acts’ associated with ‘means-plus-function’ claim limitations; (4)
whether embodiments in a claim are intended as illustrations or limitations; and (5) the
identification and delineation of claim restrictions. The designation of a default diction-
ary would reduce later disputes about which dictionary to consult (Miller, 2005; Miller
and Hilsenteger, 2005) A more costly, but potentially valuable requirement, would be to
compel applicants to disclose the relevance of prior art set forth in their Information
Disclosure Statement. The Patent Office can improve the transparency of examination as
well as the context for disclaimers by recording all interviews with applicants and placing
such recordings on the public record.
3. Ownership transparency

The obfuscation of patent ownership increases the risks and costs of developing new
products and services. Patent Office records provide poor notice of current patent
ownership, making it difficult for companies to trace the true owners of patents (FTC,
2011, pp. 130–31). Some non-practicing entities utilize shell companies and complex
assignments to avoid having the identities revealed. In 2014, the Patent Office proposed
the establishment of rules to better disclose and update patent ownership (Anderson,
2015). Although the USPTO dropped the proposal out of concern for the regula-
tory burden, there are valid concerns warranting consideration of less burdensome
alternatives.
4. Systematic judicial procedures for determining patent scope

The construction of patent claims plays a critical role in nearly every patent case. It is
central to the evaluation of infringement and validity, and can affect or determine the
outcome of other significant issues such as unenforceability, enablement, and remedies.
As Judge Giles Rich poignantly captured, ‘the name of the [patent] game is the claim’
(Rich, 1990, p. 499).
Claim construction is central to the evaluation of infringement and validity, and can
affect or determine the outcome of other significant issues such as unenforceability,
enablement, and remedies. The process by which courts interpret patent claims thus
represents one of the most important aspects of patent litigation. As the use of juries in
patent cases has grown since 1980 (Anderson and Menell, 2014), whether the judge or
the jury should construe the terms of patent claims has emerged as a pressing issue. Until
1996, it was common for courts to charge juries with claim construction. Resolving the
scope of patent claims in this manner, however, significantly increased the complexity
and uncertainty of trials. In Markman v. Westview Instruments (517 U.S. 370 (1996)), the
Supreme Court held that claim construction is a matter for the court and hence beyond
the province of the jury. The Court emphasized that judges are better equipped than

juries to construe the meaning of patent claim terms, given their training and experience
in interpreting written instruments (such as contracts and statutes).
Since 1996, district courts routinely conduct a pre-trial hearing (commonly referred
to as a ‘Markman hearing’) to resolve claim construction (Menell et al., 2016, ch. 5.1).
Timing is critical. A majority of courts have found that the most opportune time to hold
the Markman hearing is midway through, or before the close of, fact discovery, and prior
to expert discovery. This timing affords the parties sufficient discovery in advance of the
claim construction hearing to gain an understanding of the liability issues and accurately
identify the terms needing construction. It also leaves time for the parties to finish fact
discovery and to focus on expert discovery after the court has issued its claim construction
ruling. This timing also avoids requiring an expert to base his or her opinion on alternative
claim constructions or to do a new report if the court does not adopt either party’s con-
struction. Early Markman hearings can be appropriate where there is a well-crystallized
question of claim construction that can resolve liability without extensive discovery.
The primary goals of the procedures before a Markman hearing are to: (1) ensure that
the parties’ claim construction positions are squarely joined; and (2) resolve any disputes
about how the Markman hearing should be conducted so the hearing is efficient, helpful
to the court, and without procedural disarray. Courts typically request technical tutorials
in conjunction with the Markman hearing.
To structure and facilitate the claim construction process, more than 30 districts have
adopted patent local rules (PLRs) setting forth a standardized timeline and framework for
disclosures and submissions leading up to a Markman hearing. While the specific timing,
sequence, and content of disclosures and submissions vary among districts, PLRs share a
basic principle—they seek to present the court with a limited set of actual and meaningful
disputes (Menell et al., 2016, ch. 2.1; Ware and Davy, 2009).
PLRs require parties to disclose asserted claims, infringement contentions, and inva-
lidity contentions (and associated prior art references) through a structured process in
preparation for the Markman hearing. These disclosures force parties to crystallize their
theories early in the case and thereby identify matters that the Markman hearing must
resolve. They also help streamline discovery by mandating the disclosures that are core to
patent cases, thus reducing the need for interrogatories, document requests, and conten-
tion depositions. Some courts will consider summary judgment motions in conjunction
with the claim construction process so as to bring to the surface and address dispositive
issues in a timely manner (Menell et al., 2016, ch. 6.1). This integrated approach requires
the parties to state the reasons for seeking construction of any terms that are litigated
in the Markman process, regardless of whether they are asserted for summary judgment
purposes. It affords courts the context for making important rulings in the Markman
process and also provides a useful tool for reducing disputes. Figure 16.2 illustrates a
claim construction chart.
The development and implementation of standardized claim construction rules have
vastly improved the efficiency and quality of patent litigation. It has improved patent
notice and enhanced the transparency of judicial review.
5. Internalizing and counteracting adjustments

Due to inherent uncertainties surrounding patent resources, improvements to patent
disclosure and registries alone cannot fully address the incentives of resource owners to

Claim Plaintff’s Defendant’s

Term Construction Ramifications Construction Ramifications
“steering “any “a circular Noninfringement:
wheel” device for device for Accused device lacks
directing a directing a a circular steering
vehicle” vehicle” device.
Invalidity:
Reference XYZ anticipates
the claims as a
matter of law so the
patent [or specific
claim] is invalid.
Figure 16.2 Claim construction chart
overclaim, confusingly claim, obscure, and opportunistically assert patent rights. Drawing
upon the toolbox for addressing other forms of externalities, other adjustments to the
activities of the Patent Office, the patent fee structure, and patent liability and remedies
can internalize notice externalities and otherwise counteract or balance the adverse effects
of inadequate notice (Menell and Meurer, 2013).
i. Patent landscape mapping and search tools The availability of a transparent,
accessible, and easily searchable registry provides one of the best antidotes to notice
externalities. Intellectual property law should strive to empower developers to identify
all potential impediments to their projects and the counterparties with whom they can
resolve differences.
Thus, subsidizing the development, maintenance, and evolution of intellectual property
registry search tools and portals can play an important role in supporting accessible
notice information. The government further contributes to good notice by standardizing
the way that new property rights are claimed and by developing taxonomies and clas-
sifying property rights. The USPTO maintains the Acceptable Identification of Goods
and Services Manual to classify trademarks and the United States Patent Classification
Manual to classify patents.
The government could make investments in infrastructure to reduce the cost of search-
ing for patents and copyright. These investments would produce public goods that tend
to be underprovided by the private sector. One of the most pressing needs in the patent
field is the development of standard terminology for use in software inventions. Chemical
patent boundaries are more easily understood and searchable because patent attorneys and
inventors rely heavily on a system of chemical nomenclature developed and maintained by
the International Union for Pure and Applied Chemistry. There is no guarantee that an
equally successful system can be developed for software inventions but the idea has drawn
the attention of the Institute of Electrical and Electronics Engineers (IEEE) Standards
Association and other non-governmental organizations who are working on developing
standard software nomenclature. One bit of evidence suggesting progress could be made
on this front comes from semiconductor patents where the Texas Instruments TTL Data
Book has become a de facto standard for naming certain components of semiconductor
inventions.

Technology groups within the Patent Office could develop default glossaries for
commonly used terms of art that arise frequently in applications within their field.
Applications would be interpreted based on those definitions unless applicants set forth
their own definitions.
ii. Fees Likewise, notice costs in the patent system can be controlled through the use
of fees. Economists find that the volume of patent applications falls in response to an
increase in application fees, and the decision to maintain an issued patent in force is sensi-
tive to the magnitude of maintenance fees. Even the very low maintenance fees imposed
cause the majority of US patents to lapse before their full term has run (Moore, 2005).
Higher maintenance fees can be a desirable way to clear away patents on inventions that
have not been commercialized so as to provide greater freedom for developers who are
willing to take on the costs of commercializing products (Parchomovsky and Wagner,
2005).
Patent application and maintenance fees (as well as copyright and trademark registra-
tion fees) should be set so that applicants internalize not only the costs that they impose
on the government, but also the costs that they impose on third parties. Ideally, fees should
be relatively higher on inventions that are apt to generate higher notice costs. Software
patents, for example, are likely to impose greater notice costs than chemical composition
patents. The number of inventions per product tend to be much greater for software than
for chemical compositions (Bessen and Meurer, 2008). Furthermore, software patent
boundaries tend to be much harder to discern causing more claim construction disputes
and more inadvertent infringement. Thus, strictly on notice grounds, stiffer fees should
be imposed on software patents than on chemical composition patents. The larger point
is that fees are tools that can and should be used to influence the decision to apply for a
patent, the number of claims to include, and the length of time for which the patent is
maintained in force.
Internalization can also be implemented through a system of tradeable patent permits
(Ayres and Parchomovsky, 2007). In an ideal world, a policymaker with full information
could equally well regulate private decisionmakers by collecting patent fees or by specify-
ing the quantity of patents that can be granted. Permits alone or a permit-fee hybrid
system would alleviate notice externalities and in theory could be superior to regulation
by fees alone.
iii. Substantive doctrines A third approach to internalizing notice externalities is

to limit the harm to those resource developers who cannot reasonably determine the
landscape of encumbrances. Real property law has developed various doctrines for
decreasing the disruption caused by notice failure. The doctrine of adverse possession is
one tool the law uses to prevent stale claims from being asserted (Merrill, 1985). Courts
are also disinclined to read property interests broadly when they suspect claimants are
engaging in opportunistic behavior (Faus v. City of Los Angeles, 67 Cal. 2d 350, 431 P.2d
849, 62 Cal. Rptr. 193 (Cal. S. Ct. 1967)). In some states, good faith improvers of land can
take title to improved property subject to compensating the original owner for the value
of the land prior to the improvements (Good Faith Improver Act, Cal. C. Civ. Proc. §§
871.1–7; Raab v. Casper, 124 Cal. Rptr. 590 (Ct. App. 1975)). Limitations on the running
of non-possessory interests promote notice and constrain assertion of remote property

interests (French, 1982, 1998; Sterk, 1985; Tulk v. Moxhay (1848) 41 Eng. Rep. 1143 (Ch.),
1143–5; Ames, 1904; Van Houweling, 2008, pp. 891–9); Restatement (Third) of Property:
Servitudes (2000)).
Policies aimed at reducing the harms associated with notice failure already play a
significant role in intangible resource law. Liability standards cover a wide range of policy
levels: required liability elements, defenses, exemptions, compulsory licenses, and safe
harbors.
Patents can be infringed through entirely independent activities. In fact, relatively few
patent infringement cases outside of Hatch-Waxman actions (which inherently involve
copying to provoke a challenge) involve copying (Cotropia and Lemley, 2009; Bessen
and Meurer, 2008). Given the difficulty of determining the patent neighborhood due to
confidential patent applications, the proliferation of patents, and ambiguity of claims,
a strong case can be made for prior user rights and/or an independent creation defense.
At a minimum, such policy reforms appear justified in the software field where various
conditions—such as alternative means of appropriability (first-mover advantages, copy-
right protection, trade secret protection), low capital costs, and relatively low technological
risk (Menell, 1987)—obviate patent protection and opportunistic enforcement has been
rampant. The independent invention defense would go a long way toward limiting the risk
posed by opportunistic entities. It would also improve the conditions needed to support
a robust insurance market for spreading the risks of patent claims. By contrast, given the
high capital costs of pharmaceutical research, the risk of near-simultaneous invention in
conjunction with an independent invention defense could reduce invention incentives by
diminishing appropriability.
The costs of notice problems can also be alleviated by tightening the doctrine of equiva-
lents. The doctrine of equivalents allows for a finding of infringement where the accused
product or process is close to the patented invention, but does not literally infringe. The
doctrine of equivalents evolved in response to the concern that an ‘unscrupulous copyist’
could avoid literal infringement of a patented invention by making insubstantial changes
to the invention (Graver Tank & Mfg. Co. v. Linde Air Prods. Co., 339 U.S. 605, 607–08
(1950)). The doctrine is judge-made and has long served to provide courts some leeway
to ensure that insubstantial variations do not destroy the value of patents. Unfortunately,
the doctrine of equivalents undermines the notice function of patent claims. Cumulative
creators must assess not only the literal boundaries of patent claims but must also assess
whether a process or product ‘performs substantially the same function in substantially the
same way to obtain the same “result.”’ Although the Federal Circuit and the Supreme Court
have reduced the uncertainty surrounding this doctrine—Festo Corp. v. Shoketsu Kinzoku
Kabushiki Co., 535 U.S. 722 (2002) (restricting the scope of the doctrine of equivalents
in the context of amended claims to equivalents that would not have been foreseeable at
the time of amendment); Warner-Jenkinson Co. v. Hilton Davis Chem. Co., 520 U.S. 17
(1997) (requiring that the accused product or process contain each claim limitation or its
equivalent)—more can be done to better calibrate the reach of the doctrine of equivalents.
In particular, the doctrine should only be available to reach equivalents that were unforesee-
able as of the effective filing date of the application or at the time of amendment.
iv. Remedies The type and size of remedies greatly affect the chilling effects of notice
failure (Hovenkamp, 2011; FTC, 2011 (the Federal Trade Commission (FTC) report drew

heavily on a presentation of an early version of Menell and Meurer (2013)). The Supreme
Court’s eBay decision affords courts substantial discretion to consider opportunistic
enforcement of intellectual property rights (eBay v. MercExchange, 547 U.S. 388 (2006)).
Justice Kennedy’s concurring opinion expressed concern about patent owners using the
threat of an injunction ‘as a bargaining tool to charge exorbitant fees to companies that
seek to buy licenses to practice the patent’ (p. 396). He went on to note that injunctive
relief ‘may have different consequences for the burgeoning number of patents over busi-
ness methods, which were not of much economic and legal significance in earlier times.
The potential vagueness and suspect validity of some of these patents may affect the
calculus under the four-factor test’ (p. 397).
The awarding of injunctive relief and damages presents a useful place to integrate
concerns about notice externalities (Graham, 2017; Siebrasse and Cotter, 2016; Lemley
and Shapiro, 2007). There is also good reason to discount damage awards in those circum-
stances in which it is particularly difficult to identify and evaluate intellectual property
encumbrances, the boundaries surrounding the plaintiff’s rights are amorphous, or where
the law is ambiguous (e.g., doctrine of equivalents). In addition, courts should consider
notice concerns—such as opportunistic assertion—in determining the award of attorney
fees (Robinson, 2016).
6. Fostering licensing

Patent pools and other licensing institutions play a critical role in counteracting and
ameliorating patent notice and thicket issues (Rice, 2015; Mattioli, 2012; Barnett, 2009;
Merges, 1996, 2004). Coordination is critical in the growing array of network industries
(Menell, 2018). SSOs facilitate the development of FRAND licenses among competi-
tors to optimize and harmonize network product development and avoid costly patent
battles. In addition, the free and open source software (FOSS) movement promotes
ex ante dedication of technological advances to the public domain and commitment
to open development principles. A growing number of companies have joined these
efforts on an ex post basis by pledging their patents to the public. Most recently,
technology companies have developed the Defensive Patent License and License on
Transfer pledges as a means of reducing the costs and risks of opportunistic patent
assertions (Schultz and Urban, 2012). Chien (2016) advocates that patents disclose
contextual information, such as regulatory information, FRAND commitments, and
patent pledges.
7. Depropertization
The most extreme reform strategy calls for a shift away from proprietary rights systems
altogether. Demsetz (1967) theorized that private property rights emerge over time as the
value of the asset subject to the property rights increases, and as the costs of measuring,
monitoring, and enforcing private property rights decreases. By extension, as the costs of
identifying and avoiding property rights rise, then policy should move away from prop-
ertization (Smith, 2002). This line of analysis supports legislative abolition or significant
recalibrating—for example, much shorter duration—of patent protection for computer
software and business methods on notice and other grounds (Menell, 1987, 2007d; Bessen
and Meurer, 2008).

B. Copyright
Copyright faces two sets of notice challenges: tracing of copyright ownership and
assessing the scope of copyright protection (Menell, 2016b). Tracing issues—linking
copyrighted works to subsistence information about the works and contact informa-
tion about their owners—are largely solvable through implementation of existing and
developing technological means (such as digital content recognition), international
standardization, and reform of obsolete legal rules, most notably the Berne Convention
limits on formalities. The inherent uncertainty surrounding copyright scope, however,
stands in the way of copyright notice nirvana—a transparent database of fully specified
copyright resources and reliable tools for determining liability exposure ex ante. Unlike
tracing of subsistence and ownership information, current or foreseeable technology
alone cannot solve the problem of forewarning the public of the precise boundaries of
copyright interests. Nonetheless, other notice failure-based adjustments to the copyright
system can ameliorate scope clarity concerns.
1. Mandatory registration

In contrast to patents, copyright resources share the uniqueness attribute of real estate.
Their protection emanates from a particular work of authorship that can be inspected.
Unlike land resources, however, copyrighted works are neither geographically unique nor
are they limited to a singular instantiation. Copyright governs both the fixed work created
by the author as well as reproductions and performances. Unlike trespass, which can occur
only at the situs of the land resource, copyright infringement can manifest in distant locales
as well as across international borders. The practical aspects of protecting copyrights
across the globe led authors to eschew registration (recordation), marking, and other
formalities as requirements for copyright protection in the primitive technological age of
the early twentieth century (Menell, 2016b, pp. 985–6; van Gompel, 2011; Sprigman, 2004).
As reuse and adaptation of preexisting works became more feasible in the mid- to
late-twentieth century, the demand to trace and clear use of those works increased con-
comitantly. During this same time period, the duration of copyright protection expanded
significantly, further increasing the costs of tracing copyright ownership and the need for
better notice institutions. The concept of orphan works—copyrights that are very costly
or impossible to trace—emerged as a central challenge for promoting progress in the
expressive arts (Hargreaves, 2011).
The advent, rapid advance, and diffusion of digital technologies over the past quarter-
century have revolutionized the creation and dissemination of expressive creativity. The
shift from analog to digital storage media transformed the creative process. Authors and
artists can costlessly and instantaneously reproduce preexisting works and seamlessly
manipulate and edit digital sound tracks and video images. These features have generated
new art forms, such as mashup music, and have greatly expanded the creative community
in ways that vastly expand cumulative creativity.
The development of computer network technology, leading to the Internet, revo-
lutionized cumulative creativity in other ways. With the diffusion of powerful search
engines—beginning with text and leading to images, music, and video—authors and
artists obtained the ability to find all manner of preexisting works and combine them in
novel, creative ways.

The Internet revolutionized the creative industries in an even more profound way: by
enabling creators to reach vast audiences without working through traditional media
companies. Anyone with access to the Internet can now post works for the world to see,
hear, and copy. Authors can self-publish books and anyone can inexpensively develop
their own web portal. More significantly, a new industry of Internet service providers
(ISPs) and crowd-sourced web portals has emerged to regulate dissemination of all
manner of information goods.
This shift in content gatekeeping has fundamentally changed the need for and modali-
ties of copyright notice. Whereas traditional media companies relied upon tight-knit,
insular, professional networks to clear works prior to release and used a conservative
screen (‘if in doubt, leave it out’), the new breed of gatekeepers (ISPs) rely on different
methods (notice and takedown procedures and, in the case of large portals like YouTube,
content identification pre-screening technologies) for regulating what reaches the vast and
growing Internet community.
Advances in digital technology have not only expanded the need for better tracking of
and greater access to up-to-date copyright information, they have revolutionized tracing,
tracking, and updating such information. Had such tools been available at the turn of the
twentieth century, the Berne Union would likely not have prohibited formalities as pre-
requisites to the ‘enjoyment and the exercise’ of copyright (Paris Act relating to the Berne
Convention for the Protection of Literary and Artistic Works of September 9, 1886, as
amended on July 24, 1971, § 5(2)). To the contrary, such technology could have improved
international copyright enforcement and promoted progress in the expressive arts.
The arrival of the Internet initially created tremendous, unprecedented copyright
enforcement challenges. The first decade of the World Wide Web can be characterized as
a new Wild West, in which the rule of (copyright) law barely operated. Napster largely
ignored the rampant copyright infringement on its service. YouTube initially operated in
a cavalier manner. MegaUpload turned a blind eye to rampant infringement. The Pirate
Bay operated in open defiance of copyright protection. Grooveshark employees were
instructed to upload infringing material. Several waves of enforcement litigation brought
down the more egregious operators. More importantly for copyright notice, the chaos of
the Wild West inspired remarkable technological innovation.
Advances in digital identification technologies over the past decade have created
promising means for tracking, tracing, and updating copyright information. These
technologies have the ability to identify audio, textual, and visual works at low cost
and with high precision. Audible Magic Corporation was among the first to develop
sophisticated acoustic fingerprinting technologies. It now provides audio and content
identification tools to companies seeking to track digital media and identify and block
infringing content. Shazam offers an application that allows a mobile phone to identify
almost any sound recording. Gracenote developed innovative, versatile, and standardized
audio databases. YouTube’s Content ID (Audio ID and Video ID) system enables content
owners to block, monetize, and track usage of their works within YouTube’s expanding
online ecosystem.
These technologies provide the framework for an inexpensive, universal copyright
notification system. Furthermore, by tying notice information to a work’s unique digital
fingerprint rather than copies of the work, third parties could independently determine
whether a work that they have encountered is governed by copyright, the duration and

terms of its protection, who owns it, and how to contact the owner. With a modest legal
requirement, such information could be updated by the owner. If all copyrighted works
were digitized and registered, potential users of copyrighted works could employ rela-
tively inexpensive and now commonplace optical scanning and audio devices to identify
the copyright status of any registered work.
The technological advances of the past quarter-century provide the foundation for
instituting an effective and efficient system of notice and recordation for promoting
expressive creativity and free expression. We now have the tools for creating unique
digital fingerprints for nearly all types of copyrighted works using widely available and
reliable digital technologies. And unlike traditional marking, which can become obsolete
or disengaged from a copyrighted work, a digitized version of the work would serve as
the digital fingerprint and could be linked to an official, updateable, universally accessible
registry. Copyright owners would be able to embed standardized meta-tags or other digi-
tal dossiers containing subsistence information and ownership information and linking
to an official registry. User-based systems would allow copyright owners to update this
information easily through secure systems as ownership information changes.
A digital registration requirement could also serve to build a universal digital library.
Nearly all copyrighted works—from visual art to sound recordings and three-dimensional
objects—can be digitized at reasonable cost. Such digital works would serve as both the
basis for digital fingerprinting and digital deposits. A modern Copyright Office would
be built around server farms as opposed to dusty and costly physical libraries and
warehouses.
The Music Modernization Act of 2018 (MMA) powerfully moves the world toward
such a resource through an outsourced copyright database. The MMA directs the
Copyright Office to designate a collective institution to develop a comprehensive publicly
accessible database of copyrighted musical compositions and sound recordings along
with detailed ownership information. The MMA affords digital music providers, such as
Spotify, with access to a blanket license and limitations on liability. Hence, the legislation
motivates copyright owners to provide such information to the database as a means to be
paid for streaming of their copyrighted works.
Menell and Meurer (2013, pp. 50–51) propose a general digital registration system.
Once such a system was available, any person could scan a copy of a work, such as a pho-
tograph or sound recording file, into a standard computer system and conduct a search of
the global database. Just as YouTube’s Content ID system is able to run uploaded videos
against countless audiovisual works in its archive to find a match and determine whether
to allow the video to be publicly available and how to distribute advertising revenue, the
Copyright Office could, with sustainable funding, provide pertinent subsistence and
tracing information.
The next steps in developing a modern copyright notice system are technical, institu-
tional, and political. Many of the elements of an ideal system are already in use in the
private sector. Audible Magic and Google have implemented highly sophisticated systems
for digitizing works and searching for matches.
International SSOs have developed systems for recording various classes of copy-
rightable works. For example, the sound recording industry in conjunction with the
International Organization for Standardization has developed the International Standard
Recording Code for uniquely identifying sound recordings and music video recordings.

Other standardized unique identification number systems exist for books, audiovisual
works, and musical works. Such identifiers can be assigned to digitized works of author-
ship and associated digital dossiers in meta-data or other formats (Van Houweling, 2012).
The Copyright Office, or another designated entity, could host these materials on secure
servers. Copyright registrants could have supervised ability to maintain and update
identifying information.
The principal obstacles are political. There would be start-up costs to developing such
a system and inputting the data. Most major copyright industry participants in the US
already register their works, and a growing number of these companies are working with
Google, Audible Magic, and other services for identification, monetization, and policing
their works. The larger challenge would be at the international level, where formalities
have not been used for over a century (Sprigman, 2004). As online licensing grows, the
advantages of participation in a standardized, well-functioning global registration system
will increase.
A fully digitized registration and notice system can be implemented relatively easily for
new works of authorship. Ideally, however, the system would also capture prior works
that are still subject to copyright. Part of the challenge lies in motivating the copyright
owners to come forward and input their works into the universal database. The Copyright
Office could also work with Google, the Internet Archive, or other digitization projects
to pull in such works. Google’s Book Search project has already scanned a large portion
of library materials.
A mandatory copyright registration and digital deposit system would provide the
foundation for a robust digital clearance system for copyright owners and users.
Suppose, for example, that a documentary filmmaker was seeking to use photographic
works of unknown provenance. Under a decentralized safe harbor regime (and
assuming no actual knowledge of the photograph’s copyright status and ownership),
the filmmaker would scan the work using approved technology. If the scan did not
produce a match, then she would be able to use the work without fear of injunctive
relief. Furthermore, the scan would reduce costs in locating true owners if a universal
registration system was in place. As with other orphan work proposals (Pallante,
2012; U.S. Copyright Office, 2006), various forms of liability rules could be developed
(ranging from zero to fair market value) to address any claim by a legitimate copyright
holder who comes forward.
Aspects of this system are in limited use. Audible Magic’s Automatic Content
Recognition technology and Google’s Content ID screening and monetization system
show that effective digital fingerprints can be implemented and scaled. The Google Books
Library Project has successfully scanned a substantial portion of published books. Digital
tags are increasingly embedded into photographs and other works (Van Houweling,
2012).
2. Optimal specificity

The optimal specificity of copyright registration depends on the relative costs and
benefits. Copyright registration has never required anywhere near full specification of
what is excluded from the copyright claim—that is, the Swiss cheese holes—because the
costs of doing so fall well below any realizable benefit. While it would be useful to know
the freedom to operate surrounding a copyrighted work based on ex ante disclosure,

the ambiguity surrounding copyright doctrines and the myriad possible uses make such
specification highly speculative. Thus, requiring copyright claimants to specify all or
even many of the non-protected elements and compilations of elements would be both
excessively burdensome and of relatively little practical utility. Disclosure of the registered
work and the general limiting principles and doctrines of copyright law are close to the
optimal ex ante level. There may, however, be some net benefits of requiring explication
of the claimed protectable elements or disclaimers of unprotectable elements with the
registration of useful articles, computer software, architectural works, and other works
that garner relatively limited copyright protection (Denicola, 1983, p. 720 (observing that
‘the line Congress attempted to draw between copyrightable art and noncopyrightable
design “was neither clear nor new”’ (quoting H.R. Rep. No. 1476, 94th Cong. (2d Sess.,
54, 55, 1976))). With regard to other works, copyright law can better address the notice
scope challenges through other adjustments to the copyright system rather than ex ante
specification of the Swiss cheese holes.
3. Preclearance institutions

Even when copyright resources can be traced to particular owners, the inherent uncertain-
ties surrounding the scope of copyright protection can make it difficult for cumulative
creators to plan their investments. There can be significant risks in proceeding with
projects that the developer considers to be fair use. Insurance for creative projects, which
is often required by investors, can be especially difficult to obtain. Therefore, as with other
areas involving significant risk, there are advantages to being able to clear usage before
undertaking substantial investments. Land use zoning institutions make extensive use of
such permitting. Similarly, the Internal Revenue Service’s opinion letter and the Securities
and Exchange Commission’s ‘no-action’ letter processes provide mechanisms for testing
tax treatment prior to important planning decisions. Various scholars have proposed
comparable preclearance mechanisms for copyright resources (Carroll, 2007, pp. 1123–7;
Mazzone, 2009, pp. 415–21; Nimmer, 2006, p. 12; Simon 2010, pp. 527–50 (focusing on
teaching resources)).
A copyright preclearance institution could significantly reduce the risk posed by uncer-
tain copyright scope and potentially massive liability. Nonetheless, preclearance institu-
tions entail substantial administrative costs and would need to be aligned with copyright
protection, due process concerns, and cumulative creators’ temporal constraints.
4. Systematic judicial procedures for determining copyright scope

Due to the substantial limitations on copyright protection—such as the originality
doctrine, idea/expression dichotomy, and useful article doctrine—as well as the fair use
privilege, it is often difficult to assess the scope of copyright protection. As with patent
law, the only way to determine violation of copyright protection is through a judicial
determination. The process for making that determination, however, is far less predict-
able. There is no standardized process used in copyright cases, although some courts have
used informal versions of the patent process (see, e.g., Harbor Software v. Applied Sys.,
925 F. Supp. 1042, 1046 (S.D.N.Y. 1996) (observing that copyright ‘filtration analysis is a
matter of law for the Court, rather than for the jury’ based on an analogy to the Markman
case)). Such procedures should be followed in copyright litigations that turn on issues of
protectability, infringement, and fair use (Lemley and McKenna, 2016).

Table 16.1 Copyright infringement—protectable/unprotectable elements: doll drawings
Protectable Elements Include: Unprotectable Elements Include:

The precise shape, size, and placement The resemblance or similarity to human form (the
of the dolls’ anatomical features presence of hair, head, eyes, a nose, mouth, lips, ears, a
neck, legs, arms, and a torso)
Hair-dos Exaggerated features, including but not limited to an
oversized head, oversized eyes, large lips, oversized feet,
slim arms, a diminished nose, and a long torso
Face paint Idealized features, such as luscious lips, high cheekbones,
almond-shaped eyes, a slim waist, a small nose, and long
limbs
Fashion outfits and accessories The idea of a young, fashion-forward female with an
attitude, and the expression traditionally associated with
that idea, including but not limited to heavy make-up,
an ‘urban’ look, defiant poses, defiant gazes, angular
eyebrows, trendy clothing, shoes and accessories
The precise combination of features Features shared by a particular race or ethnicity (e.g.,
skin tone)
Source: Bryant v. Mattel, Inc., 2011 WL 13238416 (C.D. Cal. Apr. 21, 2011) (Instruction No. 47)).
The litigation concerning whether the Bratz line of fashion dolls infringe the drawings on
which the dolls were based (Mattel, Inc. v. MGA Entertainment, Inc., 616 F.3d 904 (9th
Cir. 2010)) illustrates how such copyright claim construction can usefully be conducted.13
The drawings in question were based significantly on human features and prior works.
After extensive pre-trial briefing, the court provided the jury instruction shown in
Table 16.1 to aid the jury in evaluating whether the Bratz dolls infringed the protectable
elements of Carter Bryant’s drawings.
5. Substantive reforms

The inherent subjectivity of copyright scope stands in the way of a transparent database
of fully specified copyright resources and reliable tools for determining liability exposure
ex ante. Unlike content recognition technologies for tracing of subsistence and ownership
information, current or foreseeable technology alone cannot solve the problem of fore-
warning the public of the precise internal boundaries of copyrighted works. Copyright
scope is far too multi-faceted. Artificial intelligence might one day assist in reducing the
uncertainty, but doctrines like the idea/expression dichotomy, substantial similarity, fair
use, and remedies are well beyond current (and foreseeable) capabilities. Other notice-
failure-based adjustments to the copyright system can, nonetheless, ameliorate scope
clarity concerns.
13
The author served as a consultant to MGA in this litigation.

i. Compulsory licensing regimes Compulsory licensing regimes provide a default that
avoids transaction costs of licensing. The Copyright Act provides several compulsory
licenses that have substantially relieved pressure on the uncertain scope of copyrighted
works. Over a century ago, Congress established the nation’s first compulsory license as
part of the 1909 Copyright Act (Copyright Act of 1909, § 1(e)). The provision authorized
anyone to sell piano rolls of musical compositions that had been released for a statutory
fee of two cents per copy. With the emergence of the sound recording industry over the
next several years, the compulsory mechanical license morphed into a mechanism for
recording artists to record their own versions of previously released musical compositions,
often referred to as a ‘cover’ license—for a prescribed statutory fee. As updated by the
omnibus Copyright Act of 1976, the ‘compulsory license includes the privilege of making
a musical arrangement of the work to the extent necessary to conform it to the style or
manner of interpretation of the performance involved, but the arrangement shall not
change the basic melody or fundamental character of the work’ (17 U.S.C. § 115(a)(2)).
The statutory rate for the ‘cover’ license has gradually risen over the past century. It now
stands at 9.1 cents or 1.75 cents per minute of playing time or fraction thereof, whichever
is greater (U.S. Copyright Office, 2010).
As a result of the ‘cover’ license, recording artists have enjoyed substantial freedom to
record and distribute their own versions of musical compositions, resulting in many of
the more memorable sound recordings. The cover license sidesteps complex scope issues
that might otherwise complicate infringement and fair use analysis. The statutory default
provides a convenient mechanism for facilitating cumulative creativity and free expression
while dividing the resulting revenue (Ginsburg, 2014; Kozinski and Newman, 1999).
Over the past century, the United States has enacted compulsory licenses for jukeboxes,
cable and satellite retransmission of publicly available broadcasts, ephemeral recordings
by broadcasters, and streaming of musical compositions and sound recordings. These
regimes economize on transaction costs and promote greater access to copyrighted works.
The rates are set through adjudicatory administrative processes. The MMA significantly
expands the use of compulsory licenses with regard to music streaming services.
The need for and appropriate scope of compulsory license regimes reemerged recently as a
result of the growth of mashup art in the digital age (McLeod and DiCola, 2011). A new gen-
eration of disc jockeys (or mashup artists) using digital sampling technology has developed
a vibrant new musical genre based on slicing, dicing, and mashing previously released sound
recordings into musical mosaics. Many of the samples are relatively short, and the resultant
works reflect innovative soundscapes. There is wide division in scholarship on whether these
types of works qualify as fair use. The current market equilibrium is far from ideal, with much
of this work circulating outside of authorized channels. This has alienated new generations
of artists and fans from the copyright system. I have proposed bringing order to this chaos
through a new compulsory license (Menell, 2016a). Such an approach directly responds to
the uncertain scope of copyright protection and promotes free expression by sidestepping
the inherent uncertainty of litigation through a balanced authorized distribution channel.
ii. Uncertain scope adjustments Where tracing the provenance of prior work is dif-
ficult or the scope of copyright protection for a prior work is murky, copyright law can
reduce the adverse effects of tracing costs and uncertain scope through adjustments to
substantive rights and remedies.

a. Fair Use The open-ended quality and subjective nature of the fair use doctrine
contributes to scope uncertainty. At the same time, the flexibility afforded by fair use
provides a means to harmonize copyright’s uncertain scope of protection with the goals
of promoting creativity and free expression (Gordon, 1982). In particular, courts should
promote ex ante bargaining over cumulative creativity by encouraging fair bargaining.
That appears to be a substantial factor in how the district court and the Second Circuit
in Bill Graham Archives v. Dorling Kindersley Ltd., 386 F.Supp.2d 3244 (S.D.N.Y. 2005),
aff’d, 48 F.3d 605 (2d Cir. 2006), resolved whether using entire concert posters and ticket
artwork as small illustrative components of a comprehensive anthology about the life
and times of the Grateful Dead constituted fair use. Both courts went out of their way to
discuss the copyright owner’s stinginess in the bargaining process, even though such facts
do not play an express role in the fair use statutory analysis (17 U.S.C. § 107; Campbell v.
Acuff-Rose Music, Inc., 510 U.S. 569, 592 (1994)). In essence, the courts were signaling
that fair bargaining is a part of fair use (Menell and Depoorter, 2014).
By making this consideration explicit, courts can effectively reduce copyright’s scope
uncertainty by promoting fair bargaining. Uses that do not directly compete or result in
lost sales ought to be resolved through ex ante accommodation rather than litigation.
Relatedly, outright refusals to bargain for non-copyright market purposes, such as politi-
cal commentary, censorship, and reputational harm, should be strongly disfavored. The
case law on this later point is mixed, as some judges have been loath to permit musical
works and sound recordings strongly associated with their authors to be used as theme
songs for politicians with opposing viewpoints (Moser, 2013; Gardner, 2011 (identifying
several copyright disputes in the context of political campaigns and noting instances of
capitulation); Henley v. DeVore, 733 F. Supp. 2d 1144 (C.D. Cal. 2010)).
b. Remedies Tailoring copyright remedies can significantly ameliorate the chilling
effects of uncertain copyright scope. Even a small risk of crushing liability can deter
cumulative creativity and free expression (Gibson, 2007). Equitable and monetary relief
can be better tailored to reduce the adverse effects of uncertain copyright scope.
The Supreme Court’s decision in eBay, Inc. v. MercExchange, 547 U.S. 388 (2006),
holding that the granting of injunctive relief in intellectual property cases is not automatic
and turns on a balancing of equitable factors, affords district courts greater discretion in
awarding injunctive relief. Courts can now split the baby, so to speak, where a cumulative
creator has significantly added to the expressive works of others but might not warrant
a fair use finding. By withholding injunctive relief in these situations, courts can enable
more works to propagate while still affording compensation to the owner of the infringed
work. The uncertainty created by the eBay framework, however, can be reduced by
expressly recognizing that injunctions should be rare outside of the piracy context (Sterk,
2005, pp. 455–62).
Liability rules are a particularly supple device for balancing competing considerations
and thus can be especially effective in dealing with the uncertainty surrounding copyright
scope. Current liability standards are too open-ended, thereby contributing to larger
potential exposure than is appropriate. The Copyright Act authorizes the copyright owner
to elect between the actual damages and disgorgement of any additional profits of the
infringer or statutory damages where the prerequisite for statutory damages (registration)
has been satisfied (17 U.S.C. § 504(a)). Copyright law could require copyright owners

to meet a working requirement, analogous to the patent working requirement used in

Europe (Wegner, 2006, pp. 158–61 (tracing the history of the working requirement)) in
order to pursue disgorgement or injunctive relief for non-piratical infringement. Limiting
disgorgement to piracy cases would remove some of the sting of the uncertain legal
standard. The infringing party has something valuable to offer the owner of the infringed
work and the consuming public.
Copyright law takes notice into consideration by barring statutory damages and attor-
neys’ fees where the copyright owner has not registered the work prior to the infringing
activity (17 U.S.C. § 504(c)). More should be done to prevent over-deterrence of activities
that fall into copyright law’s substantial scope uncertainty. Statutory damages ought not
be available outside of piratical activity (Menell, 2016b). Where an author undertakes
efforts to transform prior works by adding substantial independent creative elements,
ordinary compensatory damages should apply if the resulting work is found to be infring-
ing. Moreover, copyright law ought to cabin or eliminate statutory damages with respect
to noncommercial educational and experimental uses of copyrighted works. Often the
best way to learn a musical instrument or develop artistic or creative writing skill is to imi-
tate the works of others. Yet these acts, if publicly performed or recorded and uploaded
to a social media website, create risk of copyright liability. Copyright owners need not
worry about these uses. Fan fiction has often enriched their coffers. More importantly,
there is no better way to promote progress than to nurture artistic, musical, and literary
skills among the next generation of creators.
Furthermore, Congress should recalibrate statutory damages to avoid the risk of highly
disproportionate damage awards (Menell, 2014, pp. 313–17). Web 2.0 services can involve
thousands of works. Viacom’s lawsuit against YouTube, for example, alleged infringement
of 79,000 works (Viacom Int’l, Inc. v. YouTube, Inc., 676 F.3d 19, 26 (2d Cir. 2012)),
creating a liability range of $50 million to nearly $12 billion. Under the Supreme Court’s
decision in Feltner v. Columbia Pictures Television, Inc., 523 U.S. 340 (1998) the deter-
mination of statutory damages is a question for a jury, which potentially increases the
unpredictably of exposure. The theoretical availability of astronomical statutory damages
undermines the balance sought by copyright law. A more rational damages regime could
have produced more rational litigation, with parties identifying the gains from compro-
mise. Instead, the potential for vast, undeserved transfers of Internet-generated wealth
produced a scorched-earth court battle and deepened distrust between the content and
technology sectors. At a minimum, Congress should set a limit on the number of works
for which statutory damages are available and/or set an overall cap for statutory damages.
In a related vein, the procedural rules (including fee shifting rules) and remedies for
infringement should be re-equilibrated to make declaratory relief more prompt and effec-
tive. Given the attorneys’ fees, injury to reputation, and distraction of defending a copy-
right lawsuit as well as the risks to the project and financial exposure for liability, many
cumulative creators take a cautious approach to using other works without authorization
(Donaldson, 2008, pp. 363–7). Yet, cumulative creators often have little leverage to bring
copyright holdouts to the table. Witness the reluctance of filmmakers, Hollywood studios,
restaurants, and other businesses to challenge Warner/Chappell Music’s long-doubted
claim to copyright protection for the song ‘Happy Birthday’ (Gardner, 2015; Brauneis,
2009). Many filmmakers and recording artists over the years have agreed to significant
license fees rather than challenge the basis of the copyright claim. The availability of class

action status in such cases could potentially motivate greater efforts to invalidate dubious
copyright claims.
Menell and Depoorter (2014) propose a novel mechanism that would afford a limited,
cost-effective process for preclearing works, promote fair negotiation over cumulative uses
of copyrighted works, and reduce the exposure of cumulative creators to the inherent
risks of relying on copyright’s de minimis and/or fair use doctrines. Under this mechanism,
a cumulative creator would have the authority to make a formal offer of settlement to use
copyrighted material for a project. If the copyright owner does not respond to the offer,
the cumulative creator would be permitted to use the work provisionally by paying the
settlement amount into escrow. If the copyright owner rejects the proposed license fee and
sues for infringement, the copyright owner will bear the cumulative creator’s litigation
costs if (1) the court determines that the use of the material qualifies as fair use, or (2)
the court determines that the fair use doctrine does not excuse the use but the cumulative
creator’s offer of settlement (the proposed license fee) exceeded the amount of damages
that the court determined to be appropriate. In the former case, the escrow amount would
be returned to the cumulative creator. In the latter case, the copyright owner would receive
the infringement award from the escrow account, and the remainder would be returned
to the cumulative creator. This fee shifting proposal would encourage copyright owners
to take settlement offers seriously and negotiate around the fair use doctrine’s inherent
uncertainties. In so doing, this mechanism protects the reliance costs of cumulative
creators, reduces transaction costs, and discourages holdout behavior. Overall, this
mechanism would enrich cultural production by increasing the use of copyrighted content
in follow-on works while fostering markets for cumulative creativity and providing fair
compensation to copyright owners of underlying works.
6. Voluntary licensing regimes

As highlighted by the compulsory licensing discussion above, low transaction cost licens-
ing sidesteps the challenges of uncertain copyright scope. In addition to statutory licenses,
a variety of efficient voluntary licensing regimes has emerged to foster cumulative creativ-
ity and dissemination of copyrighted works. The organizations developing and evolving
these regimes play a vital role in promoting economic efficiency and expressive creativity
(Merges, 2004, 1996). Some of these licensing institutions are specific to particular artistic
fields; others function across fields.
Several music industry organizations have resolved important classes of bargaining
through collective licensing systems (Merges, 1996). The American Society of Composers,
Authors, and Publishers (ASCAP) pioneered the development of blanket licensing that
greatly facilitated authorized public performance of copyrighted works. The Harry Fox
Agency, established by the National Music Publishers’ Association, has long served as
an efficient clearinghouse for mechanical licenses of musical compositions for sound
recordings. These institutions have greatly facilitated commerce in, and dissemination
of, musical works and sound recordings. ASCAP and Broadcast Music, Inc. (BMI),
competing performance rights consortia, for example, operate within consent decrees
supervised by the United States District Court for the Southern District of New York
(U.S. Department of Justice, 2014).
These arrangements are coming under increasing strain as major record labels and
music publishers have sought to withdraw some of their works from these collective

licenses in an effort to extract greater value from distribution outlets (Pandora Media,
Inc. v. Am. Soc’y of Composers, Authors, and Publishers, 785 F.3d 73 (2d Cir. 2015)
(affirming decision barring ASCAP members from partially withdrawing rights as
violative of the ASCAP consent decree); U.S. Copyright Office, 2015, pp. 162–4; Internet
Policy Task Force, 2013, pp. 80–98). Such withdrawals fragment already tenuous music
platforms at a time when attracting more paid subscribers is vital to persuading music fans
to join authorized services as opposed to pirate channels (Menell, 2014, pp. 292–7; Van
Houweling, 2010, pp. 561–64; Heller, 1998, p. 624 (highlighting that ‘[w]hen there are too
many owners holding rights of exclusion, the resource is prone to underuse—a tragedy
of the anticommons’)).
The MMA addresses some of these concerns by bringing interactive streaming of
music compositions within a blanket compulsory licensing regime and shifting the rate-
setting standard from an amorphous ‘reasonableness’ standard to a market-based ‘willing
buyer/willing seller’ standard (see LaFrance, 2018, pp. 323–4). The MMA also improves
the framework for setting the compulsory license for non-interactive streaming of sound
recordings. The new regime eliminates arbitrary distinctions between older (SiriusXM,
Muzak) and newer services (Pandora) (LaFrance, 2018, p. 324). The MMA leaves the
setting of rates for interactive streaming of sound recordings to free market negotiation
between sound recording owners and music services.
In addition to collective licensing within creator communities, technology companies
have developed platforms that enable users to upload content for widespread distribution.
Google’s YouTube platform is a prime example. As discussed previously, YouTube has
developed a sophisticated pre-screening process to prevent unauthorized distribution of
copyrighted works. The Content ID system works most productively when it screens exact
copies of copyrighted works that merely substitute for the original. The copyright owner
has the choice of blocking the upload or permitting the work to be uploaded subject to a
claim to the advertising revenue.
The Content ID screening/monetization system is more controversial, however, when
the uploader is a true cumulative creator—someone who is reworking, augmenting, or
commenting on the underlying work. The Content ID screen is effectively conducting an
infringement and fair use analysis, which, for the reasons previously canvassed, is unlikely
to be reliable. When the owner of the sampled copyrighted work(s) agrees to the work
being uploaded and monetized, the cumulative work reaches a broad audience—but
potentially at the cost of losing the revenue that might rightfully be attributable to a
fair use. The cumulative creator would, of course, have the option of posting the work
elsewhere, but would not gain access to YouTube’s extensive network of users.
The most serious notice-related concern arises where the copyright owner of the
sampled work blocks the cumulative work (Sawyer, 2009). If the cumulative work is
a fair use, then the Content ID screening algorithm has interfered with cumulative
creativity and free expression. Although the Copyright Act provides a mechanism to
penalize copyright owners who ‘knowingly materially misrepresent[] . . . that material
or activity is infringing’ (17 U.S.C. § 512(f)), this deterrent has proven ineffective as a
practical matter. As illustrated by the leading case addressing remedies under § 512(f),
the litigation costs and modest remedies discourage those individuals unfairly accused
of infringing copyright law from pursuing a misrepresentation claim (Lenz v. Universal
Music Corp., 572 F. Supp. 2d 1150 (N.D. Cal. 2008), aff’d, 815 F.3d 1145 (9th Cir. 2016)).

In view of the uncertainties attendant to copyright scope and the ease of issuing take-
down requests, copyright law should afford enhanced damages for improper takedown
notices so as to encourage greater care and bargaining as opposed to costly litigation
(Menell, 2014, p. 358).
In addition to these industry-based institutions, open source projects provide an
alternative licensing model that has been especially important in the computer software
industry (Van Houweling, 2008; Merges, 2004). Open source software traces its origins to
the early 1970s and the culture of collaborative research on computer software (Weber,
2004). To perpetuate that model in the face of increasingly proprietary software, Richard
Stallman, a former researcher in MIT’s Artificial Intelligence Laboratory, established the
Free Software Foundation (FSF) to promote users’ rights to use, study, copy, modify, and
redistribute computer programs. Such rights obviously conflict with the default bundle of
rights of copyright law. For that reason, FSF developed the GNU General Public License
(GPL), a complex licensing agreement designed to prevent programmers from building
proprietary limitations into ‘free’ software. Stallman set forth a task list for the develop-
ment of a viable UNIX-compatible open source operating system. Many programmers
from throughout the world contributed to this effort on a voluntary basis, and by the
late 1980s most of the components had been assembled. The project gained substantial
momentum in 1991 when Linus Torvalds developed a UNIX-compatible kernel, which
he called ‘Linux.’ Torvalds structured the evolution of his component on the GNU GPL
‘open source’ model. The integration of the GNU and Linux components resulted in
a UNIX-compatible open source program (referred to as GNU/Linux) and has since
become widely used throughout the computing world. In the process, it has spawned a
large community of computer programmers and service organizations committed to the
principles of open source development. The growth and success of Linux has brought the
open source movement into the mainstream computer software industry. Today, a variety
of vendors, such as Red Hat, Caldera, and Ubuntu, distribute open source software, and
it has tens of millions of users worldwide.
Building on the open source model, Creative Commons provides tools to assist authors
in promoting the reuse and remixing of their works at the time of creation by opting into
a different set of defaults than those provided by the Copyright Act (Van Houweling,
2008). Creative Commons licenses distinguish among four categories of permissions: (1)
attribution—requires users to attribute the original author; (2) modification—the author
can designate whether the work can be modified; (3) share and share alike—anyone using
the work agrees to make the resulting work available on the same basis; and (4) noncom-
mercial use—the work may be used for noncommercial purposes. Creative Commons
provides tools to assist creators in finding licensed works that can be shared, remixed, or
reused.
7. Depropertization
As with patent law, there may be circumstances in which the notice problems associated
with a class of copyrighted works outweigh the social benefits of affording protection.
The most significant problems involve orphan works.
The MMA addresses the orphan works concern in several ways. First, it authorizes
noncommercial use of pre-1972 sound recordings if: (A) the person ‘makes a good faith,
reasonable search for, but does not find, the sound recording’ either in the records of the

Copyright Office, or ‘on services offering a comprehensive set of sound recordings for sale
or streaming’; (B) the person files a notice identifying the sound recordings and the nature
of the use in the Copyright Office; and (C) within 90 days after the notice ‘is indexed into
the public records of the Copyright Office,’ the rights owner does not file a notice opting
out of the noncommercial use (17 U.S.C. §1401(c)(1)). This provision provides a model
for other orphan works. Second, the MMA distributes royalties for online streamlining of
musical compositions attributable to orphan works that are unclaimed after three years
to other musical work copyright owners ‘in a transparent and equitable manner based on
data indicating the relative market shares of such copyright owners’ (17 U.S.C. §115(d)
(3)(J)(i)(II)).
A mandatory digital notice regime would effectively depropertize those works as
discussed above. Instituting a mandatory digital notice system poses a problem for
unpublished works—everything from personal letters or photographs to trade secret
manuals and object code. Demanding that such works be registered and included within
searchable digital databases for clearance matches would impose new costs and jeopardize
the privacy of these materials.
This issue highlights whether copyright protection is a public system which brings
responsibilities with its protections (Menell, 2016b, p. 1004). The real property system
requires owners to provide means for the public to know definitively whether land is
encumbered. Copyright protection should also entail such notice elements.
A personal letter or trade secret document could leak onto the Internet, and an unwit-
ting copier of such information might have no way to determine its provenance. While
current copyright law provides a cause of action and remedies for such acts, it is not at all
clear that the copyright system is well attuned to the real sources of harm, which sound
more in privacy violations and possibly hacking of computer systems. There is little
reason to believe that copyright’s market-based remedies are well suited to compensat-
ing authors who do not intend to publish their works. Moreover, exempting authors of
unpublished letters and trade secret documents from responsibility to inform the public
of their copyright claims undermines a public copyright system.
It is useful to distinguish between unpublished works that are intended for publication
(such as books, films, or sound recordings in production) and those that are not intended
for publication (such as personal letters and trade secrets). The copyright system is
designed for the former. And, in fact, Congress has established a preregistration process
to deal with the growing risk of works intended for publication being leaked to the public
(Family Entertainment and Copyright Act of 2005, Pub. L. No. 109-9, § 104(b), 119 Stat.
218, 222 (amending 17 U.S.C. § 411(a))). Such preregistration systems, as well as liability
for purloined materials that interfere with planned publication, fit well within a public-
regarding copyright system. Using copyright remedies for this narrow, but important, set
of works makes sense. We are typically not dealing with cumulative creators so much as
pirates—individuals or entities that seek to undermine the public release of costly and
highly anticipated works of authorship.
The harms associated with accessing (and copying) unpublished works that the author
intends to keep private fit better within privacy and anti-hacking regimes. Rather than
undermine the increasingly important notice function of copyright protection, Congress
could better address the problem of accessing and copying unpublished works through
suitable reforms of privacy, trade secrecy, and computer law protections and remedies.

Similarly, there are serious questions about whether photography should enjoy the full
duration and range of protections afforded other copyrighted works (Hughes, 2012).
C. Design Patent
Despite their name, design patents are much more similar to copyrights—in particular,
pictorial, graphic, and sculptural copyrighted works—than utility patents. Congress
established the design patent system in 1841 to provide protection for original ornamental
features of articles of manufacture (Hudson, 1948). The textile industry was developing
and concern arose about piracy of patterns depicted on rugs and fabrics. In addition,
metal industries sought protection for the surface ornamental and shape of stoves,
radiators, and other articles of manufacture. Copyright protection at that time did not
extend to useful articles. Although copyright protection was eventually expanded to cover
the expressive features of pictorial, graphic, and sculptural works (to the extent such
designs ‘can be identified separately from, and are capable of existing independently of,
the utilitarian aspects of the article’ (17 U.S.C. § 101 (definition of ‘[p]ictorial, graphic,
and sculptural works’))), Congress retained the largely overlapping design patent regime
(Mazer v. Stein, 347 U.S. 201, 218–19 (1954) (concluding that both regimes can cover
pictorial, graphic, and sculptural elements of useful articles)).
The principal design patent notice issue concerns the scope of protection. The scope
of design patents is based upon two-dimensional drawings. Courts will find infringe-
ment of a design patent ‘if, in the eye of an ordinary observer, giving such attention as
a purchaser usually gives, two designs are substantially the same, if the resemblance is
such as to deceive such an observer, inducing him to purchase one supposing it to be the
other’ (Egyptian Goddess, Inc. v. Swisa, Inc., 543 F.3d 665, 670 (Fed. Cir. 2008) (quoting
Gorham Co. v. White, 81 U.S. 511, 528 (1871))).
The application of this test is complicated by the need to protect only the ornamental,
and not functional, aspects of the design. The Federal Circuit interprets this to require
that a design patent be ‘primarily ornamental’ and not ‘dictated by the utilitarian purpose
of the article’ (L.A. Gear Inc. v. Thom McAn Shoe Co., 988 F.2d 1117, 1123 (Fed. Cir.
1993) (‘In determining whether a design is primarily functional or primarily ornamental
the claimed design is viewed in its entirety, for the ultimate question is not the functional
or decorative aspect of each separate feature, but the overall appearance of the article,
in determining whether the claimed design is dictated by the utilitarian purpose of the
article’); Avia Group International Inc. v. L.A. Gear California Inc., 853 F.2d 1557, 1563
(Fed. Cir. 1988) (‘[A] distinction exists between the functionality of an article or features
thereof and the functionality of the particular design of such article or features thereof
that perform a function’); Manual of Patent Examining Procedure, § 1504.01(c) (‘The
design for the article cannot be assumed to lack ornamentality merely because the article
of manufacture would seem to be primarily functional’)).14
14
Based on the inherent logic of the intellectual property system as reflected in the Supreme
Court’s Baker v. Selden decision (101 U.S. 99 (1879)) and other seminal cases (Bonito Boats, Inc.
v. Thunder Craft Boats, Inc., 489 U.S. 141 (1989) and TrafFix Devices, Inc. v. Marketing Displays,
Inc., 532 U.S. 23 (2001)) as well as legislative history of the intellectual property statutes, Menell
and Yablon (2017) contend that only the utility patent statute can protect functionality. Their

Figure 16.3 Design patent drawing
Determining whether a design is functional is especially difficult where form and function
are intertwined (Richardson v. Stanley Works, 597 F.3d 1288 (Fed. Cir. 2010) (concluding
that the ‘overall effect’ of allegedly infringing design was very different after ‘discounting’
the functional aspects of the claimed design); OddzOn Products, Inc. v. Just Toys, Inc.,
122 F.3d 1396, 1405 (Fed. Cir. 1997) (‘[w]here a design contains both functional and
non-functional elements, the scope of the claim must be construed in order to identify the
non-functional aspects of the design as shown in the patent’); compare Amini Innovation
Corp. v. Anthony Cal., Inc., 439 F.3d 1365, 1372 (Fed. Cir. 2006) (holding that while it is
proper to factor out the functional aspects of various design elements, that discounting
of functional elements must not convert the overall infringement test to an element-by-
element comparison)).
Design patent applicants limit the scope of their design patent drawings through the
use of broken or phantom lines to illustrate the environment, but not claimed aspects,
of the design (Manual of Patent Examining Procedure, § 1504.04). Only the solid lines
constitute the claimed design. As Figure 16.3 illustrates, the applicant sought protection
only for the design of the shank portion of the drill bit (Application of Hajo Zahn, 617
F.2d 261 (C.C.P.A. 1980)).
Design patent applicants can strengthen the reach of their claim by focusing on unique
portions of the article and omitting features or by disclaiming them by the use of broken
lines (Katz, 2002). This allows them to assert infringement actions against products
containing only a portion of the claimant’s entire article of manufacture.
Burstein (2016) explains how applicants can expand the scope of protection through
strategic use of claiming techniques in conjunction with continuation practice:
The PTO also allows design patent applicants to broaden their claims—in amendments or later
applications—by changing solid lines to broken lines. Sophisticated applicants exploit this rule,
using a ‘keep [one] “in the oven”’ strategy. First, the applicant files an application that claims the
entire design of a new product. ‘Then, while that application is pending, the company files one
or more continuation [or divisional applications] that claim [a smaller] portion[] of the design.’
Assuming the new application could claim priority to the original application, this strategy
allows a design patent applicant to go back to the PTO and capture competing products that
were introduced after the first design patent application was filed—even if those competing
products did not infringe the original patent claim. Importantly, there is no requirement that
the smaller portion or portions claimed in a continuation (or divisional) represent an important,
distinctive or otherwise salient design feature. (Burstein, 2016, pp. 115–16 (footnotes omitted))
analysis questions the Federal Circuit’s apparent view that a design patent can protect functional
elements of a design if those features are embodied in a ‘primarily ornamental’ overall design.

D504,889 D618,677
Figure 16.4 Apple design patents
The murkiness of the Federal Circuit’s functionality jurisprudence as well as the Patent
Office’s flexible claiming and continuation rules create tremendous uncertainty regarding
the scope of design patents. As with copyright protection, design patents are like Swiss
cheese. Although the solid lines of the design patent claim are easily perceived, the unpro-
tected elements—unoriginal, obvious, and functional aspects—are not readily apparent.
Even though the drill bit depicted above contained broken lines, it is unclear what aspects
of the shank are protected.
Burstein (2016, pp. 125–8) highlights numerous ‘trite, uncreative, or obviously unpa-
tentable designs’ issued by the Patent Office. Menell and Yablon (2019) present empirical
evidence showing a gradual, but substantial, shift from ornamental surface ornamenta-
tion toward more functional, shape-based designs over the century and a half of the
design patent regime. The risks associated with the uncertain scope of design patents are
exacerbated by the potential for outsize damages created by design patent law’s disgorge-
ment remedy (35 U.S.C. § 289 (liability for ‘total profit’)).
Apple’s smartphone and tablet design patents (shown in Figure 16.4) illustrate the
design patent notice problems. It is difficult to see how rounded rectangle designs were
considered original and nonobvious in light of prior tablets and conventional geometric
shapes (Monroe, 2017; Compaq TC1000, Wikipedia, https://en.wikipedia.org/wiki/
Compaq_TC1000). It is also difficult to see how these minimalist, shock-resistant designs
were found to be ‘primarily ornamental.’ They lack surface ornamentation and the shapes
are closely tied to their function, conventional screen, paper, and pocket sizes, and engi-
neering considerations. (The shapes of these products were ruled functional under trade
dress law. (Apple Inc. v. Samsung Electronics Co., Ltd., 786 F.3d 983 (Fed. Cir. 2015).)
Yet the courts upheld the validity of these design patents, found that Samsung infringed
both, and ordered damages of more than half a billion dollars.
Thus, trade dress protection creates severe notice challenges. The costs of these prob-
lems have become more salient as a result of the protection afforded to Apple’s minimalist
consumer devices and growing recognition of the large potential recovery.
These problems can be addressed at multiple levels. The Patent Office should imple-
ment more robust design patent examination and require applicants seeking protection
for product shape to expressly disclaim functional elements. The Federal Circuit should
re-examine its functionality jurisprudence. The ‘primarily ornamental’ and ‘dictated by

functionality’ standards misapprehend the preemptive supremacy of the utility patent

system in protection of functional advances. Product designers should not be able to cam-
ouflage functional elements with ornamental flourishes. Where the form and function are
inextricably intertwined, then design patent protection should not be available. Barring
such scrutiny, the Supreme Court should look more carefully at the Federal Circuit’s
lax standard for determining functionality. It is out of step with the standards that the
Supreme Court has applied to copyright and trade dress protection, and conflicts with the
logic of the overall intellectual property system (Menell and Yablon, 2017).
D. Trademark
The notice issues surrounding trademarks are far less significant than in utility patent,
design patent, and copyright fields. Nonetheless, the geographic and product/service
scope of trademarks can be ambiguous. Trademark notice could be significantly enhanced
through a mandatory public registration system and the implementation of digital tools
for scanning/searching graphic and trade dress marks.
E. Trade Secret
Trade secrets function as an anti-notice regime. Disclosing the information destroys

trade secrecy. Hence, the notice concerns raised with the other intellectual property
regimes have no analog in the trade secret system. Nonetheless, as highlighted above in
section II(C)(4), the overbroad scope of conventional NDAs can subvert public policy by
discouraging employees and consultants from reporting illegal activity within corpora-
tions and other organizations. The Defend Trade Secrets Act of 2016 (DTSA) addresses
this problem by immunizing whistleblowers who disclose trade secrets to the government
confidentially (Menell, 2017). One of the challenges in achieving the worthy goal of the
whistleblower immunity provision is in providing the target audience with information
about their immunity. DTSA addresses this problem by requiring that employers ‘provide
notice of the immunity . . . in any contract or agreement with an employee that governs
the use of a trade secret or other confidential information’ (18 U.S.C. § 1833(b)(3); Menell,
2017, p. 54).
REFERENCES
Ames, J.B. 1904. ‘Specific Performance for and against Strangers to the Contract,’ 17 Harvard Law Review
174–85.
Anderson, J. Jonas, and Peter S. Menell. 2014. ‘Informal Deference: A Historical, Empirical, and Normative
Analysis of Patent Claim Construction,’ 108 Northwestern University Law Review 1–83.
Anderson, Nathan P. 2015. ‘Striking a Balance: The Pursuit of Transparent Patent Ownership,’ 30 Berkeley
Anton, James J., and Dennis A. Yao. 2002. ‘The Sale of Ideas: Strategic Disclosure, Property Rights, and
Contracting,’ 69 The Review of Economic Studies 513–31.
Arrow, Kenneth. 1962. ‘Economic Welfare and the Allocation of Resources for Inventions,’ in R.R. Nelson, ed.,
The Rate and Direction of Inventive Activity. Princeton, NJ: Princeton University Press.
Ayres, Ian, and Gideon Parchomovsky. 2007. ‘Tradable Patent Rights,’ 60 Stanford Law Review 863–94.

Bessen, James, and Michael J. Meurer. 2008. Patent Failure: How Judges, Bureaucrats and Lawyers Put
Innovation at Risk. Princeton, NJ: Princeton University Press.
Bessen, James, and Michael J. Meurer. 2014. ‘The Direct Costs from NPE Disputes,’ 99 Cornell Law Review
387–424.
Brauneis, Robert. 2009. ‘Copyright and the World’s Most Popular Song,’ 56 Journal of the Copyright Society
of the U.S.A. 335–426.
Burk, Dan. L. 2008. ‘The Role of Patent Law in Knowledge Codification,’ 23 Berkeley Technology Law Journal
1009–34.
Burk, Dan L., and Mark A. Lemley. 2009. ‘Fence Posts or Sign Posts? Rethinking Patent Claim Construction,’
157 University of Pennsylvania Law Review 1743–99.
Burstein, Michael J. 2012. ‘Exchanging Information without Intellectual Property,’ 91 Texas Law Review 227–82.
Burstein, Sarah. 2016. ‘Costly Designs,’ 77 Ohio State Law Journal 107–57.
Carroll, Michael W. 2007. ‘Fixing Fair Use,’ 85 North Carolina Law Review 1087–154.
Chiang, Tun-Jen. 2010. ‘Fixing Patent Boundaries,’ 108 Michigan Law Review 523–76.
Chien, Colleen V. 2016. ‘Contextualizing Patent Disclosure,’ 69 Vanderbilt Law Review 1849–90.
Coleman, Jules L. 1992. ‘Risks and Wrongs,’ 15 Harvard Journal of Law & Public Policy 637–48.
Conti, Annamaria, Jerry Thursby, and Marie Thursby. 2013. ‘Patents as Signals for Startup Financing,’ 22
Journal of Industrial Economics 592–622.
Cotropia, Christopher A., and Mark A. Lemley. 2009. ‘Copying in Patent Law,’ 87 North Carolina Law Review
1421–66.
Denicola, Robert C. 1983. ‘Applied Art and Industrial Design: A Suggested Approach to Copyright in Useful
Articles,’ 67 Minnesota Law Review 707–48.
Devlin, Alan. 2010. ‘The Misunderstood Function of Disclosure in Patent Law,’ 23 Harvard Journal of Law &
Donaldson, Michael C. 2008. Clearance and Copyright: Everything You Need to Know for Film and Television.
Beverly Hills, CA: Silman-James Press. 3rd edn.
Dwyer, John P., and Peter S. Menell. 1998. Property Law and Policy: A Comparative Institutional Perspective.
Westbury, NY: The Foundation Press, Inc.
Fish, Robert D. 2007. Strategic Patenting. Bloomington, IN: Trafford Publishing.
French, Susan F. 1982. ‘Toward a Modern Law of Servitudes: Reweaving the Ancient Strands,’ 55 Southern
French, Susan F. 1998. ‘The Touch and Concern Doctrine and the Restatement (Third) of Servitudes: A Tribute
to Lawrence E. Berger,’ 77 Nebraska Law Review 653–66.
Fromer, Jeanne C. 2009. ‘Patent Disclosure,’ 94 Iowa Law Review 539–606.
FTC (Federal Trade Commission). 2011. The Evolving IP Marketplace: Aligning Patent Notice and Remedies with
Competition, accessed March 21, 2019 at https://www.ftc.gov/reports/evolving-ip-marketplace-aligning-patent-
notice-remedies-competition.
Gardner, Eriq. 2011. ‘Michele Bahmann in Legal Spat for Using Tom Petty’s “American Girl” at Rally,’ The
Hollywood Reporter: Hollywood, Esq. (June 28, 2011), accessed March 21, 2019 at http://www.hollywoo-
dreporter.com/thr-esq/michele-bachmann-legal-spat-using-206257.
Gardner, Eriq. 2015. ‘“Happy Birthday” Copyright Ruled to Be Invalid,’ The Hollywood Reporter (September
22, 2015), accessed March 21, 2019 at http://www.hollywoodreporter.com/thr-esq/happy-birthday-copy
right-ruled-be-826528.
Gibson, James. 2007. ‘Risk Aversion and Rights Accretion in Intellectual Property Law,’ 116 Yale Law Journal
882–951.
Ginsburg, Jane C. 2014. ‘Fair Use for Free, or Permitted-but-Paid?,’ 29 Berkeley Technology Law Journal
1383–446.
and its Predecessors,’ 82 Columbia Law Review 1600–657.
Law and Technology Patent Damages Workshop,’ 25 Texas Intellectual Property Law Journal 115–42.
Hargreaves, Ian. 2011. ‘Digital Opportunity: A Review of Intellectual Property and Growth,’ accessed March
21, 2019 at https://www.gov.uk/government/uploads/system/uploads/attachment_data/file/32563/ipreview-
finalreport.pdf.
Heller, Michael A. 1998. ‘The Tragedy of the Anticommons: Property in the Transition from Marx to Markets,’

Heller, Michael. 2008. The Gridlock Economy: How Too Much Ownership Wrecks Markets, Stops Innovation, and
Costs Lives. New York, NY: Basic Books.
Heller, Michael A. and Rebecca S. Eisenberg. 1998. ‘Can Patents Deter Innovation? The Anticommons in
Hovenkamp, Herbert. 2011. ‘Notice and Patent Remedies,’ 88 Texas Law Review 221–34.
Hudson, Thomas B. 1948. ‘A Brief History of the Development of Design Patent Protection in the United
States,’ 30 Journal of the Patent Office Society 380–99.
Hughes, Justin. 2012. ‘The Photographer’s Copyright: Photograph as Art, Photograph as Database,’ 25 Harvard
Journal of Law & Technology 327–428.
Internet Policy Task Force, Department of Commerce. 2013. Copyright Policy, Creativity, and Innovation in the
Digital Economy, accessed March 21, 2019 at https://www.uspto.gov/sites/default/files/news/publications/
copyrightgreenpaper.pdf.
Katz, Robert S. 2002. ‘Examination of Design Patents in the United States,’ 10 University of Baltimore
Intellectual Property Law Journal 109–15.
Kozinski, Alex, and Christopher Newman. 1999. ‘What’s So Fair about Fair Use?,’ 46 Journal of the Copyright
Society of the U.S.A. 513–30.
Krazit, Tom, and Anne Broache. 2006. ‘BlackBerry Saved,’ CNET News (March 3, 2006).
Lee, William P., and Lawrence P. Cogswell. 2004. ‘Understanding and Addressing the Unfair Dilemma Created
by the Doctrine of Willful Patent Infringement,’ 41 Houston Law Review 393–458.
Lemley, Mark A. 2000. ‘Reconceiving Patents in the Age of Venture Capital,’ 4 Journal of Small and Emerging
Business Law 137–48.
Lemley, Mark A. 2008. ‘Ignoring Patents,’ 1 Michigan State Law Review 19–34.
Lemley, Mark A., and Colleen V. Chien. 2003. ‘Are the U.S. Patent Priority Rules Really Necessary?,’ 54 Hastings
Lemley, Mark A., and Mark P. McKenna. 2016. ‘Scope,’ 57 William & Mary Law Review 2197–286.
Lemley, Mark A., and Kimberly A. Moore. 2004. ‘Ending Abuse of Patent Continuations,’ 84 Boston University
Lemley, Mark. A., and Carl Shapiro. 2007. ‘Patent Holdup and Royalty Stacking,’ 85 Texas Law Review 2163–74.
Lemley, Mark A., and Ragesh K. Tangri. 2003. ‘Ending Patent Law’s Willfulness Game,’ 18 Berkeley Technology
Libecap, Gary D., and Dean Lueck. 2011a. ‘Land Demarcation Systems,’ in Kenneth Ayotte and Henry E.
Smith, eds., Research Handbook on the Economics of Property Law. Cheltenham, UK and Northampton,
Libecap, Gary D., and Dean Lueck. 2011b. ‘The Demarcation of Land and the Role of Coordinating Property
Institutions,’ 119 Journal of Political Economy 426–67.
Lincoln, Abraham. 1953. ‘Second Lecture on Discoveries and Inventions,’ in, Roy P. Basler, ed., The Collected
Works of Abraham Lincoln, vol. III. New Brunswick, NJ: Rutgers University Press.
Mattioli, Michael. 2012. ‘Communities of Innovation,’ 106 Northwestern University Law Review 103–56.
Mazzone, Jason. 2009. ‘Administering Fair Use,’ 51 William and Mary Law Review 395–438.
McJohn, Stephen M. 2008. ‘Patents: Hiding from History,’ 24 Santa Clara Computer & High Technology Law
Journal 961–80.
Durham, NC: Duke University Press.
Menell, Peter S. 2007a. ‘Knowledge Accessibility and Preservation Policy for the Digital Age,’ 44 Houston Law
Review 1013–72.
Menell, Peter S. 2007b. ‘Bankruptcy Treatment of Intellectual Property Assets: An Economic Analysis,’ 22
Menell, Peter S. 2007c. ‘The Property Rights Movement’s Embrace of Intellectual Property: True Love or
Menell, Peter S. 2007d. ‘A Method for Reforming the Patent System,’ 13 Michigan Telecommunications and
Menell, Peter S. 2011. ‘Governance of Intellectual Resources and the Disintegration of Intellectual Property in
the Digital Age,’ 26 Berkeley Technology Law Journal 1523–60.
Menell, Peter S. 2013. ‘It’s Time to Make Vague Software Patents More Clear,’ Wired (February 7,
2013), accessed March 21, 2019 at http://www.wired.com/opinion/2013/02/its-time-to-make-vague-software
-patents-more-clear.
Menell, Peter S. 2014. ‘This American Copyright Life: Reflections on Re-Equilibrating Copyright for the
Internet Age,’ 61 Journal of the Copyright Society of the U.S.A. 235–371.
Menell, Peter S. 2015. ‘Promoting Patent Claim Clarity, Comments Submitted in Response to the PTO’s Request

for Comments on Enhancing Patent Quality,’ 80 Fed. Reg. 6475 (February 5, 2015), accessed March 21, 2019
at http://btlj.org/2016/05/promoting-patent-claim-clarity/.
Menell, Peter S. 2016a. ‘Adapting Copyright for the Mashup Generation,’ 164 University of Pennsylvania Law
Review 441–512.
Menell, Peter S. 2016b. ‘Economic Analysis of Copyright Notice: Tracing and Scope in the Digital Age,’ 96
Boston University Law Review 967–1024.
Menell, Peter S. 2017. ‘Tailoring a Public Policy Exception to Trade Secret Protection,’ 105 California Law
Review 1–64.
Menell, Peter S. 2019. ‘Economic Analysis of Network Effects and Intellectual Property,’ in Ben Depoorter and
Peter S. Menell, eds., Research Handbook on the Economics of Intellectual Property, vol. I. Cheltenham, UK
Menell, Peter S., and Ben Depoorter. 2014. ‘Using Fee Shifting to Promote Fair Use and Fair Licensing,’ 102
Analysis 1–59.
Shavell, eds., Handbook of Law and Economics, vol. II. Amsterdam: North-Holland.
Menell, Peter S., and Daniel Yablon, 2017. ‘Star Athletica’s Fissure in the Intellectual Property Functionality
Landscape,’ 166 University of Pennsylvania Law Review Online 137–47.
Menell, Peter S., and Daniel Yablon. 2019. ‘Useful Article Protection: Tracing and Disentangling the Intellectual
Property Landscape’ (manuscript).
Menell, Peter S., Mark A. Lemley, and Robert P. Merges, 2018. Intellectual Property in the New Technological
Age, vol. II – Copyrights, Trademarks, and State IP Protections. Berkeley, CA: Clause 8 Publishing.
Menell, Peter S., Matthew D. Powers, and Steven C. Carlson. 2010. ‘Patent Claim Construction: A Modern
Synthesis and Structured Framework,’ 25 Berkeley Technology Law Journal 711–829.
Menell, Peter S., Matthew Powers, Lynn Pasahow, James Pooley, Steven Carlson, Jeffrey Homrig, George
Pappas, Carolyn Chang, Colette Reiner Mayer, and Marc David Peters. 2016. Patent Case Management
Judicial Guide. Washington, D.C.: Federal Judicial Center. 3rd edn.
Merges, Robert P. 2004. ‘A New Dynamism in the Public Domain,’ 71 University of Chicago Law Review 183–204.
Merrill, Thomas W. 1985. ‘Property Rules, Liability Rules, and Adverse Possession,’ 79 Northwestern University
Merrill, Thomas W., and Henry E. Smith. 2000. ‘Optimal Standardization in the Law of Property: The Numerus
Clausus Principle, 110 Yale Law Journal 1–70.
Miller, Joseph Scott. 2005. ‘Enhancing Patent Disclosure for Faithful Claim Construction,’ 9 Lewis & Clark
Miller, Joseph Scott, and James A. Hilsenteger. 2005. ‘The Proven Key: Roles and Rules for Dictionaries at the
Patent Office and the Courts,’ 54 American University Law Review 829–940.
Monroe, Bryan. 2017. ‘Apple’s New Tablet? Been There, Done That,’ Huff Post (December 6, 2017), accessed
March 21, 2019 at https://www.huffingtonpost.com/bryan-monroe/apples-new-tablet-been-th_b_416960.html.
Moore, Kimberly A. 2005. ‘Worthless Patents,’ 20 Berkeley Technology Law Journal 1521–52.
Moser, Jana. 2013. ‘Songs in Contention: Copyright Holders Have Begun to Challenge the Customary
Appropriation of Songs for Political Campaigns,’ 36 Los Angeles Lawyer 28–33.
Nimmer, David. 2006. ‘A Modest Proposal to Streamline Fair Use Determinations,’ 24 Cardozo Arts &
Note. 2007. ‘The Disclosure Function of the Patent System (or Lack Thereof),’ 118 Harvard Law Review
2007–28.
Ouellette, Lisa Larrimore. 2012. ‘Do Patents Disclose Useful Information?,’ 25 Harvard Journal of Law &
Pallante, Maria A. 2012. ‘Orphan Works and Mass Digitization: Obstacles and Opportunities,’ 27 Berkeley
Review 1–78.
Pressman, David. 2012. Patent it Yourself: Your Step-by-Step Guide to Filing at the U.S. Patent Office. Berkeley,
CA: Nolo. 16th edn.
Rice, James M. 2015. ‘The Defensive Patent Playbook,’ 30 Berkeley Technology Law Journal 725–76.
Rich, Giles S. 1990. ‘The Extent of the Protection and Interpretation of Claims: American Perspectives,’ 21
International Review of Industrial Property and Copyright Law 497–510.
Risch, Michael. 2007. ‘The Failure of Public Notice in Patent Prosecution,’ 21 Harvard Journal of Law &

Robinson, W. Keith. 2016. ‘Awarding Attorney Fees and Deterring “Patent Trolls,”’ 20 Lewis & Clark Law
Review 281–302.
Samuelson, Pamela. 2016. ‘Notice Failures Arising from Copyright Duration Rules,’ 96 Boston University Law
Review 667–90.
Sawyer, Michael S. 2009. ‘Filters, Fair Use and Feedback: User-Generated Content Principles and the DMCA,’
& Technology 1–68.
Schwartz, David L., and Ted Sichelman. 2019. ‘Data Sources on Patents, Copyrights, Trademarks, and Other
Intellectual Property,’ in Peter S. Menell and David L. Schwartz, eds., Research Handbook on the Economics
of Intellectual Property, vol. II. Cheltenham, UK and Northampton, MA, USA: Edward Elgar Publishing.
Seymore, Sean B. 2010. ‘The Teaching Function of Patents,’ 85 Notre Dame Law Review 621–70.
Shapiro, Carl. 2001. ‘Navigating the Patent Thicket: Cross Licensing, Patent Pools, and Standard Setting,’ in, Adam
B. Jaffe, Josh Lerner, and Scott Stern, eds., Innovation Policy and the Economy. Cambridge, MA: MIT Press.
Sheldon, Jeffrey G. 2005. How to Write a Patent Application. New York, NY: Practicing Law Institute.
Siebrasse, Norman V., and Thomas F. Cotter. 2016. ‘A New Framework/or Determining Reasonable Royalties
in Patent Litigation,’ 68 Florida Law Review 929–99.
Simon, David A. 2010. ‘Teaching without Infringement: A New Model for Educational Fair Use,’ 20 Fordham
Intellectual Property, Media & Entertainment Law Journal 453–562.
Smith, Henry E. 2002. ‘Exclusion Versus Governance: Two Strategies for Delineating Property Rights,’ 31
Journal of Legal Studies S453–88.
Smith, Henry E. 2003. ‘The Language of Property: Form, Context, and Audience,’ 55 Stanford Law Review
1105–92.
Smith, Henry E. 2011. ‘Standardization in Property Law,’ in Kenneth Ayotte and Henry E. Smith, eds., Research
Handbook on the Economics of Property Law. Cheltenham, UK and Northampton, MA, USA: Edward Elgar
Publishing.
Sprigman, Christopher. 2004. ‘Reform(aliz)ing Copyright,’ 57 Stanford Law Review 485–568.
Sterk, Stewart E. 1985. ‘Freedom from Freedom of Contract: The Enduring Value of Servitude Restrictions,’
70 Iowa Law Review 615–62.
Sterk, Stewart E. 2005. ‘Intellectualizing Property: The Tenuous Connections between Land and Copyright,’ 83
Washington University Law Quarterly 417–70.
van Gompel, Stef. 2011. Formalities in Copyright Law: An Analysis of their History, Rationales and Possible
Futures. Alphen aan den Rijn: Kluwer Law International.
Review 549–642.
Van Houweling, Molly Shaffer. 2012. ‘Atomism and Automation,’ 27 Berkeley Technology Law Journal
1471–502.
Van Houweling, Molly Shaffer. 2019. ‘Intellectual Property as Property,’ in Peter S. Menell and Ben Depoorter,
eds., Research Handbook on the Economics of Intellectual Property, vol. I. Cheltenham, UK and Northampton,
U.S. Copyright Office. 2006. Report on Orphan Works, accessed March 21, 2019 at http://www.copyright.gov/
orphan/orphan-report.pdf.
U.S. Copyright Office. 2010. Mechanical License Royalty Rates, accessed March 21, 2019 at http://www.
copyright.gov/licensing/m200a.pdf.
U.S. Copyright Office. 2015. Copyright and the Music Marketplace, accessed March 21, 2019 at https://copy
right.gov/docs/musiclicensingstudy/copyright-and-the-music-marketplace.pdf.
U.S. Department of Justice. 2014. Antitrust Consent Decree Review: ASCAP and BMI, accessed March 21, 2019
at https://www.justice.gov/atr/ascap-bmi-decree-review.
U.S. Patent and Trademark Office. 2019. How to Satisfy a Disclaimer Requirement, accessed March 21, 2019 at
https://www.uspto.gov/trademark/laws-regulations/how-satisfy-disclaimer-requirement.
U.S. Patent No. 5,205,473 col. 4-5 (filed April 27, 1993).
U.S. Patent No. 5,436,960 (filed July 25, 1995).
Ware, James, and Brian Davy. 2009. ‘The History, Content, Application and Influence of the Northern District
of California’s Patent Local Rules,’ 25 Santa Clara Computer & High Technology Law Journal 965–1032.
Wegner, Harold C. 2006. ‘Injunctive Relief: A Charming Betsy Boomerang,’ 4 Northwestern Journal of
Technology and Intellectual Property 156–170.
Wheeler, George F. 2003. ‘Creative Claim Drafting: Claim Drafting Strategies, Specification Preparation, and
Prosecution Tactics,’ 3 John Marshall Review of Intellectual Property Law 34–60.

Wikipedia, ‘Compaq TC1000,’ accessed March 21, 2019 at https://en.wikipedia.org/wiki/Compaq_TC1000.

Wikipedia, ‘Metes and Bounds,’ accessed March 21, 2019 at https://en.wikipedia.org/wiki/Metes_and_bounds.
Cases
Amini Innovation Corp. v. Anthony Cal., Inc., 439 F.3d 1365 (Fed. Cir. 2006).
Apple Inc. v. Samsung Electronics Co., Ltd., 786 F.3d 983 (Fed. Cir. 2015).
Application of Hajo Zahn, 617 F.2d 261 (C.C.P.A. 1980).
Ariad Pharmaceuticals v. Eli Lilly and Co., 598 F.3d 1336 (Fed. Cir. 2010).
Avia Group International Inc. v. L.A. Gear California Inc., 853 F.2d 1557 (Fed. Cir. 1988).
Baker v. Selden, 101 U.S. 99 (1879).
Bill Graham Archives v. Dorling Kindersley Ltd., 386 F.Supp.2d 3244 (S.D.N.Y. 2005), aff’d, 48 F.3d 605 (2d
Cir. 2006).
Blaustein v. Burton, 88 Cal. Rptr. 319 (Cal. Ct. App. 1970).
Bonito Boats, Inc. v. Thunder Craft Boats, Inc., 489 U.S. 141 (1989).
Bright Tunes Music Corp. v. Harrisongs Music, Ltd., 420 F. Supp. 177 (S.D.N.Y. 1976).
Bryant v. Mattel, Inc., 2011 WL 13238416 (C.D. Cal. Apr. 21, 2011).
Campbell v. Acuff-Rose Music, Inc., 510 U.S. 569 (1994).
Desny v. Wilder, 46 Cal. 2d 715 (Cal. S. Ct. 1956).
eBay v. MercExchange, 547 U.S. 388 (2006).
Egyptian Goddess, Inc. v. Swisa, Inc., 543 F.3d 665 (Fed. Cir. 2008).
Faus v. City of Los Angeles, 67 Cal. 2d 350, 431 P.2d 849, 62 Cal. Rptr. 193 (Cal. S. Ct. 1967).
Feltner v. Columbia Pictures Television, Inc., 523 U.S. 340 (1998).
Festo Corp. v. Shoketsu Kinzoku Kabushiki Co., 535 U.S. 722 (2002).
Gentry Gallery, Inc. v. Berkline Corp., 134 F.3d 1473 (Fed. Cir. 1998).
Gorham Co. v. White, 81 U.S. 511 (1871).
Graver Tank & Mfg. Co. v. Linde Air Prods. Co., 339 U.S. 605 (1950).
Halo Electronics, Inc. v. Pulse Electronics, Inc., 136 S. Ct. 1923 (2016).
Harbor Software v. Applied Sys., 925 F. Supp. 1042 (S.D.N.Y. 1996).
Henley v. DeVore, 733 F. Supp. 2d 1144 (C.D. Cal. 2010).
In re NTP, 654 F.3d 1279 (Fed. Cir. 2011).
L.A. Gear Inc. v. Thom McAn Shoe Co., 988 F.2d 1117, 1123 (Fed. Cir. 1993).
Lenz v. Universal Music Corp., 572 F. Supp. 2d 1150 (N.D. Cal. 2008), aff’d, 815 F.3d 1145 (9th Cir. 2016).
Markman v. Westview Instruments, Inc., 517 U.S. 370 (1996).
Mattel, Inc. v. MGA Entertainment, Inc., 616 F.3d 904 (9th Cir. 2010).
Mazer v. Stein, 347 U.S. 201 (1954).
MedImmune, Inc. v. Genentech, Inc., 549 U.S. 118 (2007).
Octane Fitness, LLC v. ICON Health & Fitness, Inc., 572 U.S. 545 (2014).
OddzOn Products, Inc. v. Just Toys, Inc., 122 F.3d 1396 (Fed. Cir. 1997).
Pandora Media, Inc. v. Am. Soc’y of Composers, Authors, and Publishers, 785 F.3d 73 (2d Cir. 2015).
Raab v. Casper, 124 Cal. Rptr. 590 (Ct. App. 1975).
Richardson v. Stanley Works, 597 F.3d 1288 (Fed. Cir. 2010).
Sheldon v. Metro-Goldwyn Pictures Corp., 81 F.2d 49 (2d Cir. 1936).
Smith v. Dravo, 203 F.2d 369 (7th Cir. 1953).
TrafFix Devices, Inc. v. Marketing Displays, Inc., 532 U.S. 23 (2001).
Tulk v. Moxhay (1848) 41 Eng. Rep. 1143 (Ch.).
Viacom Int’l, Inc. v. YouTube, Inc., 676 F.3d 19 (2d Cir. 2012).
Warner-Jenkinson Co. v. Hilton Davis Chem. Co., 520 U.S. 17 (1997).
17 U.S.C. § 101, § 102(a), § 102(b), § 107, § 115(a)(2), § 115(d)(3)(E)(i), § 115(d)(3)(E)(ii), § 504(a), § 504(c), §
512(f).
18 U.S.C. § 1833(b)(3).
35 U.S.C. § 112(a), § 112(b), § 112(f), § 122, § 284, § 285, § 289.
Copyright Act of 1909, § 1(e).
Copyright Office, Form VA, line 6.
Defend Trade Secrets Act of 2016, Pub. L. No. 114-153, 130 Stat. 376 (codified in scattered Sections of 18 and
34).

Family Entertainment and Copyright Act of 2005, Pub. L. No. 109-9, § 104(b), 119 Stat. 218, 222 (amending
17 U.S.C. § 411(a)).
Good Faith Improver Act, Cal. C. Civ. Proc. §§ 871.1-7.
H.R. Rep. No. 1476, 94th Cong. (2d Sess. 1976).
Manual of Patent Examining Procedure, § 1504.01(c).
Manual of Patent Examining Procedure, § 1504.04.
Music Modernization Act of 2018, Pub. L. 115–264, 132 Stat. 3676, codified at §§ 114, 115, 801, 803, 1401.
Northern District of California Patent Local Rules.
Paris Act relating to the Berne Convention for the Protection of Literary and Artistic Works of September 9,
1886, as amended on July 24, 1971, § 5(2).
Restatement (Third) of Property: Servitudes (2000).
Restatement (Third) of Unfair Competition § 41.
U.S. Const. art I, § 8, cl. 8.

PART IV
IP AND INSTITUTIONS

17. Patent institutions: shifting interactions between
legal actors
Arti K. Rai* 15
Contents
I. Introduction
II. Institutional Actors
A. The Era of Federal Circuit Supremacy, 1982–2002
B. The Supreme Court and Executive Branch Strike Back
C. The Role of Congress
D. The Role of Inferior Tribunals: Article III Trial Courts, the International
Trade Commission (ITC), and the PTAB
III. Normative Analysis
A. The Federal Circuit and Inferior Tribunals
B. The Supreme Court
IV. Future Research Questions
A. The Supreme Court
B. PTAB
C. Comparison with Other Technically Complex Areas
References
I. INTRODUCTION
As demonstrated by multiple relevant contributions in Volumes I and II, the literature on

patent institutions is voluminous. To keep the inquiry tractable, this contribution covers
only research on actors responsible for the development and implementation of US patent
law. Moreover, because Part II of this Volume addresses in detail the empirical literature
on several very important legal actors – most notably the Court of Appeals for the Federal
Circuit (Federal Circuit) and the US Patent and Trademark Office (USPTO) – I focus on
interactions between legal actors. I consider both legalist and strategic perspectives on
these interactions.
I turn to this topic about a decade after having written several articles (Rai, 2000, 2003)
making a strong normative claim that Congress missed the mark when it decided in 1982
to enhance expertise solely at the appellate court level. I argued that many of the most
* Elvin R. Latty Professor Law, Duke Law School; Faculty Director, Duke Law Center for
Innovation Policy; Duke Innovation and Entrepreneurship Initiative Research Fellow. I thank
participants at the ‘Research Handbook: Economics of Intellectual Property Rights – Volume 1
Theory’ conference for extremely helpful comments.
473

important inquiries associated with determinations of patent validity and infringement

involve relatively technical adjudicative facts. Moreover, under conventional legal process
principles (Hart and Sacks, 1994), this reality counseled in favor of deploying greater
expertise and resources at the administrative and trial court levels. With deployment of
such expertise and resources, the Federal Circuit would have no cause to act as a fact-finder
in its review of the USPTO and trial courts. Further, the Federal Circuit’s predisposition
towards a rule-formalism that was largely insensitive to policy goals – previously justified by
the limited competence of inferior actors in the patent system – would lose its justification.
Over the intervening years, some of this ‘devolution’ has arguably occurred.
Congressional intervention, intervention by the Supreme Court allocating power to lower
tribunals, and even aggressive ‘competition’ for patent cases by certain district courts
have taken power away from the Federal Circuit. But devolution has not necessarily been
accompanied by fortification of expertise and resources. And even where devolution has
been accompanied by such fortification, such as through the creation of the USPTO
Patent Trial and Appeals Board (PTAB), the results have been controversial. Meanwhile,
many questions persist regarding whether current interactions between patent institutions
are creating either appropriate results in individual cases or appropriate policy for the
system as a whole. Perhaps as a consequence, the Supreme Court continues its active
supervision of institutional interactions. A review of interactions between legal actors in
the system is thus quite timely.
Although my prior work and this contribution focus on the period after the creation
of the Federal Circuit, the Constitution of course mentions patents and Congress saw fit
to create a patent system as early as 1790. Through the centuries, however, Congressional
involvement on central substantive matters has often been limited to articulation of some
very basic language (Nard, 2010). Given the many gaps in statutory law, Congress has
arguably delegated authority to other institutions (Rai, 2003; Burk and Lemley, 2003;
Nard, 2010). In any event, these other institutions have either implicitly or explicitly
developed patent law beyond the statutory text.
Particularly since the creation of the Federal Circuit in 1982, the courts have been the
primary institution (Nard, 2010). Although the USPTO has been conducting examina-
tion since 1836, its very longevity as an institution means that the agency substantially
predates the rise of the modern administrative state (Duffy, 2000). As discussed below,
path-dependence from the USPTO’s original limited role; a suite of concerns surrounding
administrative regulation of ‘property rights’; aggressive review by the Federal Circuit;
and persistent problems in managing workflow that undermine the agency’s reputation
have all conspired to keep the USPTO from exercising significant control over questions
of law and policy.
The USPTO’s limited power vis-à-vis other institutional actors has had a number of
follow-on effects. In general, because the courts rather than the USPTO have maintained
the standards by which applications are judged, concerns about USPTO workload have
not entered the calculus for these standards. The resulting influx of new, and repeat, appli-
cations has further taxed the USPTO’s already limited and uncertain financial resources.
That said, the USPTO, together with other executive agencies, has in recent years had
some influence over the powerful Office of the Solicitor General. The net result has been
some recent USPTO influence over the Supreme Court. The USPTO also has a first
mover advantage that it is sometimes able to exploit. Finally, the powers, and additional

Patent institutions: shifting interactions between legal actors 475
resources, conferred on the agency by the America Invents Act of 2011 (AIA) may give
it broader authority going forward. This broader authority has already raised, and will
continue to raise, a plethora of interesting research questions.
Within the court system, the Federal Circuit was the predominant player from 1982
through the early 21st century. The elimination of appeals to regional circuits did not,
however, eliminate litigants’ desire to seek out particular district courts (Moore, 2001). To
the contrary, litigants’ predilection for forum shopping accelerated (Lemley, 2010), with
the evidence indicating that non-practicing entities in particular sought out the Eastern
District of Texas (Anderson, 2015; Klerman and Reilly, 2016). With district courts of
course come the institution of juries, and many have questioned the considerable power
given to juries (Lemley, 2013).
In the last decade, the Federal Circuit has also shared the spotlight with the Supreme
Court. The emergence of the Supreme Court as a relevant actor has begun to generate
research on its role and also on strategic interactions between the Federal Circuit and the
Supreme Court.
As the preceding discussion may suggest, the relevant descriptive and normative
literatures are quite rich. To keep the normative discussion tractable, I adopt the standard,
if narrow, economic account that patent law exists to promote innovation. From an
institutional and procedural standpoint, achieving innovation requires that patent law
institutions promote efficiency and accuracy in rights allocation. Efficiency and accuracy
of course map directly on to the familiar goal of minimizing the sum of administrative
costs and error costs emphasized by standard economic analyses of legal procedure. Many
patent scholars also note that institutional predictability and consistency are particularly
important for reducing the sum of administrative and error costs.
Finally, I address what I view as the most pressing open research questions. Some of
these questions have been addressed in part, but more research needs to be done. In other
cases, the events in question are of recent vintage, and a full institutional account must
await further developments.
Section II reviews descriptive accounts of patent law’s institutional actors. Section III
reviews normative accounts. Section IV presents open research questions.
II. INSTITUTIONAL ACTORS
In 1982, Congress set up the Court of Appeals for the Federal Circuit to hear all appeals
in patent cases. The current institutional debate continues to focus heavily on the Federal
Circuit. Discussions of other important patent institutions, such as the Supreme Court,
the USPTO, other agencies, and trial courts often refer to the Federal Circuit. For
purposes of the historical and descriptive account, the creation of the Federal Circuit in
1982 thus serves as a useful point of demarcation.
A. The Era of Federal Circuit Supremacy, 1982–2002
As Ryan Vacca’s contribution to Volume II discusses in detail, momentum to create the

Federal Circuit emerged in part from a more general project on reducing appellate court
caseload and improving legal consistency undertaken by the Hruska Commission (see

references cited in Vacca, pp. 733–61). Although the Commission’s main recommendation

of a new appellate court to handle cases referred by the Supreme Court was rejected, and
the Commission itself rejected creating a specialized court for patent appeals, the policy
community paid substantial attention to the Commission’s determination that existing
regional appellate courts had particular difficulty with patent law.
The Congressional decision to create the Federal Circuit was driven by arguments
that regional appellate courts had created highly undesirable levels of unpredictability
and inconsistency in patent law. Congress also found persuasive claims that ‘strong’
patents – that is, patents that were easy to acquire and enforce – were critical for national
competitiveness and economic growth (references in Vacca, pp. 733–61).
The Federal Circuit quickly made a number of changes to strengthen patents and unify
patent law. Within the first five years of its formation, the court adopted a strong pre-
sumption of validity for issued patents. The Court also made clear its view that, contrary
to certain Supreme Court suggestions, inventions that represented combinations of prior
art did not create any need for a special non-obviousness standard. Further, the Court
adopted a strong presumption of validity for issued patents (Dreyfuss, 1989).
For its part, the Supreme Court largely withdrew from the patent field for almost
two decades. To the extent the Court intervened, it did so primarily not on issues of
substantive patent law but on allocation of power questions (Janis, 2001). Perhaps not
surprisingly, the Congressional attempt to concentrate all expertise – on questions of fact
as well as law and policy – at the appellate level produced some challenges for conventional
legal process principles regarding allocation of power.
Indeed, the first Federal Circuit decision reviewed by the Supreme Court, Dennison
Manufacturing v. Panduit, 475 U.S. 809 (1986), involved a complaint that the Federal
Circuit was exercising plenary power over the non-obviousness determination. The
Supreme Court vacated the Federal Circuit’s decision and remanded the case to the
Federal Circuit with instructions to the court to explain why Federal Rule of Civil
Procedure 52(a) did not mandate deference to the trial court’s factual determinations
regarding non-obviousness. However, this rebuke by the Supreme Court did not appear
dramatically to change Federal Circuit practice.
Although commentators discussed the Federal Circuit’s aggressive review of trial court
fact-finding in many areas of patent law (Rooklidge and Weil, 2000; Rai, 2003), the focal
point for scholarly criticism was the Federal Circuit’s aggressive review of district court
claim construction. Because claim construction is critical for purposes of determining
both patent validity and infringement, aggressive review can have many follow-on effects.
For example, if appellate reversal on claim construction occurs after a full district court
determination of validity and infringement, the district court must once again determine
validity and infringement (Rai, 2003).
In fairness to the Federal Circuit, most district courts were not necessarily constituted
to handle patent cases. District courts were particularly problematic because judges
frequently turned complex and important exercises like claim construction over to lay
juries. Notably, rates of jury trial use, which had begun to rise in the 1970s, had reached
over 70 percent by the early 1990s (Anderson and Menell, 2013).
In Markman v. Westview Instruments, 52 F.3d 967 (Fed. Cir. 1995) the Federal Circuit
determined that claim construction was a question of law for the judge and that such
constructions should therefore be reviewed de novo. The next year, in another one of the

‘allocation-of-power’ interventions that the Supreme Court conducted during this period,
the Court agreed with the Federal Circuit’s conclusion that claim construction was a ques-
tion for the judge (Markman v. Westview Instruments, 517 U.S. 370 (1996)). However, the
Court explicitly stated that claim construction represented a mixed question of law and
fact and relied on functional considerations rather than the fact/law distinction to assign
the inquiry to judges.
Despite the Supreme Court’s statements, the Federal Circuit’s en banc 1998 decision
in Cybor Corp. v. FAS Technologies, 138 F.3d 1448 (Fed. Cir. 1998) once again declared
claim construction a pure question of law to be reviewed de novo. Invoking the Supreme
Court among other authorities, scholars vigorously challenged this view, arguing that
patent law’s reference to the ‘person having ordinary skill in the art’ required attention
not only to text but also to context and industry custom (Nard, 2000).
Controversies over claim construction leading to Cybor (1998), as well as the Cybor
decision itself, generated several empirical studies of the Federal Circuit’s reversal rate on
claim construction (Moore, 2001; Chu, 2001). These studies found that for the subset of
cases in which parties elected to go through the full appeal process (and thus subject to
any selection bias for this subset of cases), de novo review appeared to operate not only in
theory but also in practice – about one-third of all claim decisions appealed to the Federal
Circuit were reversed.
Henry and Turner (2006) provided a perspective on Federal Circuit interaction with
trial courts that focused not on particular doctrines but on whether the court was more
‘pro-patent’ than its predecessors. As it happened, the question of pro-patent bias had
implications not only for assessment of the Federal Circuit but also for interactions
between the district courts and the Federal Circuit. Based on a large-scale empirical
analysis of district and appellate court opinions from 1953–2002, Henry and Turner
found that the Federal Circuit was significantly more reluctant than its predecessors to
affirm district court findings of invalidity. However, it was not more reluctant to affirm
‘not infringed’ decisions. Henry and Turner argued that because of the Federal Circuit’s
tendencies, district courts decided patents to be invalid significantly less often, patentees
appealed decisions of invalidity significantly more often, and infringement frequently
became the dispositive inquiry at the trial court.
Scholars have not conducted large-scale studies comparing Federal Circuit review
of the USPTO with that of its predecessor, the Court of Customs and Patent Appeals
(CCPA). In any event, because the Federal Circuit adopted the decisions of the CCPA
and included many of its judges (Lefstin, 2010), the Federal Circuit’s creation did not
represent a structural break vis-à-vis the USPTO. Scholars have noted, however, that
during the era of Federal Circuit supremacy, the court’s direct review of USPTO patent
decision making generally took place in the context of patent denials. Thus, Federal
Circuit reversal of the USPTO necessarily expanded the bounds of patentability (Rai,
2000; Masur, 2011; Wasserman, 2011).
Through the 1990s, this dynamic played out in both biotechnology and software.
Specifically, as detailed in Rai (1999), in the area of gene sequence patents, the Federal
Circuit repeatedly overruled the USPTO’s view that, for the average scientist working
in the area, knowing a general method for selecting genes through the use of nucleotide
probes, as well as the complete or partial amino acid of the protein for which a gene of
interest coded, would render the DNA base sequence for the gene obvious. The Federal

Circuit also repeatedly overturned the USPTO’s position that an algorithm run on a
general-purpose computer was not patent-eligible subject matter. The latter set of deci-
sions laid the foundation for State Street Bank v. Signature Financial Group, Inc., 149
F.3d 1368 (Fed. Cir. 1998), in which the Federal Circuit essentially equated the patentable
subject matter requirement of patentability with the utility requirement.
While Rai (1999) focused on repeated reversals of USPTO patent denials by the Federal
Circuit, some commentators deployed strategic models to argue that patent expan-
sion could happen even without actual reversal of the USPTO by the Federal Circuit.
Specifically, such expansion could occur if the USPTO’s incentives led it primarily to fear
reversal by the Federal Circuit, with the consequence that the USPTO denied patents only
in cases where the most ‘pro-patent’ Federal Circuit judge would agree with the denial
(Masur, 2011; Wasserman, 2011).
As these analyses would suggest, Federal Circuit supremacy had particularly concen-
trated effects on the USPTO. Hall (2005) deployed rigorous empirical analysis to argue
that the creation of the Federal Circuit was in part responsible for the surge in patent
applications that occurred in the 1980s. Federal Circuit decisions through the 1990s
lowering the non-obviousness standard and essentially eliminating patentable subject
matter as a separate requirement of patentability may have prompted yet more patent
applications. From 1990 to 2000, the number of applications increased by almost 80
percent (as contrasted with about 58 percent from 1980 to 1990). At the same time, about
$750 million in fee-based revenue that could have been used to hire additional examiners
was diverted by Congress from USPTO coffers. The application backlog thus increased
through the 1990s (Long, 2009).
Of course, as discussed in Volume II, backlog and other challenges were also exacer-
bated by various longstanding structural features of the USPTO system for managing
workflow. One of these structural features – the ability of applicants before the USPTO
(uniquely among patent offices around the globe) to file an unlimited number of repeat,
or continuation, applications (Moore and Lemley, 2004) – directly implicates the institu-
tional interaction and power questions on which this contribution focuses. Although the
USPTO attempted in 2007 to place limits on continuations, the agency ultimately backed
down in the face of a vigorous litigation challenge asserting that the limits represented
improper expansion of the USPTO’s relatively narrow rulemaking authority (Rai, 2009).
Another axis of ‘court-watching’ assessed the court’s jurisprudential approach. A
number of scholars noted the court’s predilection for rule-formalism (Rai, 2003; Thomas,
2003). While some of these bright-line rules were specific to particular technologies,
Thomas (2003) emphasized a trans-technological bright-line rule on non-obviousness
imposed by many three-judge panels of the Federal Circuit. This requirement moved well
beyond the modest statement made in the early years of the Federal Circuit that combina-
tion inventions did not have to comply with a higher standard of non-obviousness. Rather,
under this requirement, the USPTO examiner (and trial courts) had to identify a specific
documentary ‘teaching, suggestion, or motivation’ to combine prior art references when
using these references for purposes of demonstrating obviousness.
While noting that Federal Circuit language often abjured policy considerations, Burk
and Lemley (2003) argued that some of its decisions were implicitly driven by policy
concerns relevant to particular areas of technology. For example, the rule lowering the
non-obviousness standard in biotechnology essentially to a test of novelty emerged

from the court’s view that patents needed to be available for gene sequences that encode
therapeutic proteins. Without these patents, the biopharmaceutical sector would not be
able to secure the patents necessary to drive investment through expensive regulatory
approval processes. Burk and Lemley (2003) did note, however, that the Federal Circuit
was not transparent about its policy-based reasoning.
Yet another axis of Federal Circuit ‘court-watching’ has assessed uniformity and clarity
in its jurisprudence. For example, based on extensive coding of opinions, Wagner and
Petherbridge (2004) argued that the court’s claim construction jurisprudence depended
heavily on the panel that rendered the claim construction.
In contrast, Nard and Duffy (2007) argued that the court had shown very significant
(and, in their view, undesirable) uniformity. In response to the descriptive claim made
by Nard and Duffy, other scholars used frequency of dissents and frequency of en banc
opinions to argue that diversity of opinion was apparent within the Federal Circuit, both
on particular issues such as the doctrine of equivalents (Petherbridge, 2009) and also as
compared to other appellate courts (Vacca, 2011).
Notably, the 2007 publication of the Nard and Duffy article happened just as the
Federal Circuit was beginning to take more cases en banc. For example, as a comprehen-
sive listing by Vacca (2011) shows, the court took 28 cases en banc during the 21 years
after its creation – an average of about 1.3 cases a year. During the five-year period from
2007 to 2011, by contrast, it took 13 cases en banc – an average of 2.6 a year. This ramp-up
may have been a response to emerging interest in the patent system by other legal actors,
including the Supreme Court. Indeed, Golden (2009) explicitly nominated the Supreme
Court as the ‘prime percolator’ – an institution well suited for the purpose of dislodging
any stagnation on the part of the Federal Circuit.
I turn next to the prominent role that the Supreme Court and the executive branch
(including, but not limited to, the USPTO) have played over the last 10–15 years.
B. The Supreme Court and Executive Branch Strike Back
Although Federal Circuit supremacy in patent law obviously did not end on a date certain,
by the 2001 term, an increase in Supreme Court interest in patent law was clear (Duffy,
2010). Additionally, by this time, the emergence of large numbers of patent applications
and grants on inventions such as business methods (software-enabled and otherwise) and
scientific research tools had attracted policy and scholarly criticism. The Federal Trade
Commission (FTC) conducted an inquiry focused in significant part on institutional
challenges, and on interactions between institutions (FTC, 2003). It criticized the Federal
Circuit’s requirement that patent challengers marshall clear and convincing evidence
to overturn granted patents as well as its formalist directive that inferior institutions
that examined patents for non-obviousness – whether USPTO examiners or district
courts – show a specific teaching, suggesting, and motivation (TSM) to combine prior art
references. The FTC also recommended the creation of a robust administrative system to
review granted patents. A prominent report by the National Academy of Sciences (NAS,
2004) also criticized the TSM test and discussed in detail the rationale for administrative
review as a robust alternative to patent litigation.
By the end of the first decade of the 21st century, the Supreme Court had issued opinions
rejecting a bright-line TSM approach (KSR International Co. v. Teleflex, 550 U.S. 398

(2007)), holding that injunctive relief was not mandatory upon a finding of validity and
infringement (eBay Inc. v. MercExchange, LLC, 547 U.S. 338 (2006)), and stating that
patentable subject matter doctrine in fact excluded certain categories of invention (Bilski v.
Kappos, 561 U.S. 593 (2010)). The Supreme Court’s decisions to take certiorari, and its sub-
stantive positions, usually reflected intervention by the Solicitor General (Duffy, 2010). The
Solicitor General, in turn, was influenced by a number of agencies in the executive branch,
including but not limited to, the USPTO (Rai, 2012). In general, the Supreme Court moved
the case law away from bright line rules and towards more flexible standards (Lee, 2010).
One of the most important Supreme Court decisions, not only for purposes of substan-
tive patent law, but also for purposes of institutional interaction, was KSR International
Co. v. Teleflex Inc. (2007). According to Rantanen (2013), in the 10 years prior to the
grant of certiorari in KSR, the Federal Circuit found 54 percent of patent claims on
appeal non-obvious. Since then, it has found 43 percent non-obvious. Notably, the shift
was driven by greater reluctance on the part of the Federal Circuit to reverse district court
findings of obviousness. Indeed, according to Nock and Gadde (2011), during the 2.5-
year period immediately after KSR that they studied, the Federal Circuit did not reverse
a single lower tribunal determination that a patent was obvious.
As the Nock and Gadde (2011) study suggests, KSR was quite significant not only for
Federal Circuit review of trial courts but also of the USPTO. In KSR, the government
had filed an amicus brief detailing the ways in which a rigid requirement that examiners
show ‘teaching, suggestion, or motivation’ – particularly in documentary form – posed an
unnecessary burden and was contrary to accepted administrative law principles of official
notice. In its unanimous opinion, the Supreme Court agreed.
KSR’s embrace of conventional principles of administrative review was consistent with
its decision in Dickinson v. Zurko, 527 U.S. 150 (1999). In that decision, the Court held
that the USPTO was an agency under the Administrative Procedure Act and thus the
Federal Circuit’s review of USPTO fact-finding was subject to a standard of review even
more deferential than appellate review of trial court fact-finding. The Supreme Court’s
approach stood in significant contrast to the approach to administrative law taken by the
Federal Circuit (Nard, 1995; Benjamin and Rai, 2007).
Similarly, in contrast to the Federal Circuit, the Supreme Court’s decision in eBay Inc. v.
MercExchange (2006) purported to embrace ordinary principles of remedies law. And in
two recent cases respectively involving attorneys’ fees and claim construction, Highmark
Inc. v. Allcare Health Mgmt. Sys. Inc., 572 U.S. __ (2014) and Teva Pharm. USA, Inc.
v. Sandoz Inc., 135 S. Ct. 831 (2015), the Court emphasized conventional principles of
deferential appellate review with respect to discretionary findings and findings of fact by
lower tribunals. These cases indicate that the Supreme Court not only prefers standards
to rules but also prefers standards that are not specific to patent law.
C. The Role of Congress
As Supreme Court interest in the patent system grew, so did Congressional interest.
Starting in 2005, Congress began holding a series of hearings on legislative patent
reform. As the literature on Congressional veto gates would have predicted, the process
of legislative reform was blocked at many turns by confrontation between rent-seeking
interest groups large and small. By the time the AIA was finally passed, many of the

substantive patent law recommendations made by bodies like the FTC and the National
Academies, ended up being addressed through Supreme Court opinions and also by en
banc determinations by the Federal Circuit (Burk and Lemley, 2009).
Congress did, however, ultimately pass legislation – the AIA – that had the effect of
changing institutional structure significantly. Congress did not follow the conventional
modern model of conferring on the administrative agency in question (i.e. the USPTO)
broad rulemaking authority. But it did create a relatively robust system of post-grant
review within the USPTO (Matal, 2012). As I discuss further in the normative analysis, to
the extent the PTAB can correct the USPTO’s initial errors in a manner that is both accu-
rate and relatively inexpensive, it should fulfill the standard economic goal of minimizing
the sum of administrative costs and error costs.
D. The Role of Inferior Tribunals: Article III Trial Courts, the International Trade
Commission (ITC), and the PTAB
As scholars have noted (Anderson, 2015), even when substantive law is uniform, differ-
ences in norms and procedures among district courts as well as local differences in the
composition of prospective jury pools can motivate litigants to shop for favorable district
court fora. Indeed, a study conducted during the period of Federal Circuit supremacy
(Moore, 2001) found significant clustering of cases outside traditional technology centers.
This clustering has increased substantially in recent years (Lemley, 2010), and several
scholars have argued that certain district courts, particularly the Eastern District of Texas,
actively compete for patent plaintiffs (Anderson, 2015; Klerman and Reilly, 2016).
How these trial courts, which have presumably watched closely Supreme Court deci-
sions that confer more power upon trial courts, will respond to these decisions remains
to be seen. While some of these decisions apply only to Article III trial courts, the Teva
v. Sandoz (2015) decision will also confer power on the International Trade Commission,
which has recently become a popular venue for seeking exclusion orders of imported
goods based on alleged patent infringement (Chien, 2008; Schwartz, 2009; Kumar, 2011).
As for the USPTO’s PTAB, it has been, and will be, affected not only by Teva v. Sandoz,
but also multiple Supreme Court decisions that involve direct challenges to its procedures.
The most dramatic of these cases, Oil States Energy Services v. Greene’s Energy Grp.
(2018), rejected an argument that the PTAB was unconstitutional because only Article
III institutions could properly decide the validity of granted patents. But given shifting
currents in administrative law, how much power the PTAB will be able to exercise over
the long term is unclear.
III. NORMATIVE ANALYSIS
The literature addressed in Section II generally includes normative conclusions and

recommendations. As noted in the Introduction, patent scholars typically view accuracy
and efficiency as the normative criteria against which the institutional and procedural
features of the system should be evaluated. The patent scholars’ frame of accuracy and
efficiency maps onto the frame of minimizing the sum of error costs and administrative
costs emphasized by economic analysts of legal procedure. Although patent scholars

generally agree on these standard economic goals, they often differ substantially on how
interactions between patent institutions should be structured to best achieve accuracy and
efficiency. This section discusses the diverse perspectives in the literature.
This section surveys normative recommendations, stressing those recommendations
that focus on interactions between legal actors. I first address interactions between the
Federal Circuit with lower tribunals. I then address the Supreme Court.
A. The Federal Circuit and Inferior Tribunals
In an early assessment, Dreyfuss (1989) evaluated Federal Circuit decisions on the metrics
of accuracy and predictability. While she acknowledged potential tensions between these
goals, she used examples like the Federal Circuit’s elevation of secondary considerations
in the area of non-obviousness to argue that, at least on questions of patent validity, the
court had largely achieved both goals.
Within a decade or so of this early assessment, the court’s success in achieving either goal
was questioned. As with the descriptive analysis, claim construction served as a focal point
for the normative discussion. Scholars questioned whether the Federal Circuit’s often high
reversal rates on claim construction (and attendant consequences for vertical predictabil-
ity) were consistent with the Congressional desire to create predictability and efficiency in
patent law (Moore, 2001). Empirical studies noting inconsistencies among Federal Circuit
panel decisions on claim construction methodology (Wagner and Petherbridge, 2004) also
called into question whether the Federal Circuit was creating predictable law.
In contrast to the standard view favoring uniformity and perceiving insufficient levels
of uniformity, other scholars equated uniformity not with desirable predictability and
certainty but with undesirable stagnation (Nard and Duffy, 2007). The critique alleging
excess Federal Circuit rule-formalism (Rai, 2003; Thomas, 2003) also viewed the Federal
Circuit as sacrificing accuracy on the altar of predictability.
Although the rule-formalism and stagnation critiques agreed that the system had failed
to achieve accuracy, their prescriptions pointed in different directions. After conceding
that rule-formalism might be justified by the Federal Circuit’s desire to give clear instruc-
tions to inferior actors of questionable competence, Rai (2003) called for a ‘first-best’ (if
politically challenging) solution in which inferior actors were endowed with considerable
competence in technical fact-finding, including through limits on juries. At that point, the
Federal Circuit’s rule-formalism would no longer be justified, and other actors, such as the
Supreme Court, could intervene to dislodge the Federal Circuit’s formalism.
In contrast, Nard and Duffy (2007) prescribed a regime in which the Federal Circuit
would be forced to compete with a few additional appellate courts. On this view, while
the earlier system in which all regional appellate courts could hear patent cases was too
unpredictable, and a single court was too uniform, a few courts would be just right. The
availability of a few appellate courts would not only introduce appropriate levels of
competition in post-issuance litigation, but it would also remove the Federal Circuit’s
monopoly control over USPTO determinations.
In Rai (2003), I also suggested a regime in which an expert, specialized trial court – and
appropriate deference thereto on questions of fact-finding – be used to generate vertical
predictability. Empirical work by Schwartz (2008, 2011) strikes a cautionary note about
the extent to which specialization in the form of repeated exposure to patent cases at the

trial level will improve such vertical predictability. Schwartz’s data on claim construction
indicate that, at least when the standard of review is de novo (as it was during the time
period studied by Schwartz), there is little correlation between a trial judge’s prior experi-
ence with claim construction and the judge’s reversal rate at the Federal Circuit. Kesan
and Ball (2011) report similar results on claim construction but do find that a judge’s
increased patent-specific experience lowers reversal rates on a number of other issues, such
as preliminary injunctions, judgments as a matter of law, and infringement. The Kesan and
Ball data suggest that for purposes of understanding institutional interactions in the patent
arena, claim construction, though very important, may not necessarily be representative.
In the wake of recent Supreme Court decision making, the patent system has moved
towards employing fact-intensive standards and to greater deference to lower tribunals.
As Holbrook (2011) observes, however, because fact-intensive standards decrease notice
and predictability, they might best be viewed as a check on the presumptions created by
legalistic rules. Also worrisome is the reality that devolution to lower tribunals has not
necessarily been accompanied by fortification of the actual capacity of these tribunals. To
the contrary, in the case of patent-heavy tribunals such as the Eastern District of Texas
and perhaps the District of Delaware, any expertise acquired by attracting very large
numbers of patent cases may be overwhelmed by the judicial bias that arguably caused
plaintiffs to file in that tribunal (Anderson, 2015).
Some scholars have suggested restricting venue to districts where the defendant has
its primary place of business, arguing that such venue restriction would not only reduce
forum shopping but would also create technology-specific expertise within particular trial
courts (Fromer, 2010). As noted, others have discussed the creation of a single specialized
trial court. Still others have suggested that more incremental measures like randomization
of judge selection within a district or the existing Patent Pilot Program, under which
judges can volunteer for patent cases, may have salutary effects (Anderson, 2015). Going
forward, empirical data on both the Patent Pilot Program and the impact of the Supreme
Court’s decision in TC Heartland LLC v. Kraft Food Brands LLC, 137 S. Ct. 1514 (2017)
restricting forum shopping by patentees will provide scholars with fodder for continued
research on improving district court adjudication.
Scholars have also discussed whether Article III litigation represents the best mecha-
nism for addressing errors in individual cases, particularly with respect to erroneously
granted patents. While some have argued that most erroneous grants are not commercially
significant, and have endorsed litigation as a good mechanism for targeting the small
percentage of erroneous grants that are commercially significant (Lemley, 2001), others
have emphasized collective action and other barriers to socially optimal use of Article III
litigation by defendants (Merges and Farrell, 2004; Benjamin and Rai, 2007). On the latter
view, even if the USPTO does not, or cannot, do more to ensure that erroneous grants
do not occur, these errors should be corrected through an ex post administrative process.
As discussed in Section IV, the emergence of a robust ex post administrative process as a
major competitor to Article III litigation is a development that needs close monitoring.
Scholars have also discussed how to evaluate accuracy in patent law decision making.
Beyond the difficulty of ensuring that the system reaches the correct result in a particular
case lies the larger question of whether the system as a whole is set up to develop law
and policy accurately. While Burk and Lemley (2009) are optimistic about the capacity
of courts, the conventional argument that courts have limited resources to make policy,

and can do so only in the context of highly selected cases that arise ex post, has led vari-
ous scholars to advocate that Congress consider a standard administrative approach to
patents – delegating rulemaking authority to an agency, presumably but not necessarily
the USPTO (Burstein, 2011; Masur, 2010).
The Federal Circuit’s recent interest in en banc decision making, and the policy-
oriented questions it has asked in some of its recent en banc orders, leads Vacca (2011)
to argue that policymaking by the Federal Circuit is not only normatively desirable but
actually happening. Nard (2010) and Burk and Lemley (2009), both writing before the
passage of the AIA, stress the USPTO’s alleged capture by patent applicants as a reason
to favor courts. Vertinsky (2010) focuses on transition costs associated with patent reform,
arguing that policymaking by courts is desirable because changes implemented by courts
are likely to be smaller and more specific than changes implemented by other institutions.
Congressional creation of a strong system of administrative review for granted patents
– the PTAB – has also begun to generate normative discussion. The creation of the
PTAB, which can be reversed not only when it rejects patents but also when it upholds
them, responds in part to the concerns of Masur (2011) and Wasserman (2011) regarding
asymmetric review.
Wasserman (2013) further argues that the PTAB’s procedures are best seen as formal
adjudications. Thus, under Chevron v. Natural Resources Defense Council, 467 U.S. 837
(1984) and its progeny, the Federal Circuit should defer not only to PTAB fact-finding (as
it is already required to do) but also to the PTAB’s findings on ultimate legal questions.
On an optimistic read, the result would be a significant PTAB role in improving accuracy
not only in individual cases but also system wide.
Golden (2014) argues that, even without formal deference, the USPTO has a first mover
advantage that it has sometimes been able to exploit to shape law and policy, particularly
in the context of patent denials. Rai (2013) argues that the AIA gives the PTAB the power
to exercise this first mover advantage in the context of both patent grants and denials.
Finally, prior to recent interventions by the Supreme Court, some commentators on
the Federal Circuit stressed that its review of inferior actors, as well as its behavior more
generally, was motivated by an undesirable ‘pro-patent’ bias (Jaffe and Lerner, 2004). On
this view, the Federal Circuit listened primarily to patent lawyers, and a pervasive role for
patents in the national economy enhanced the court’s power and prestige. However, the
Federal Circuit’s tendency to read patent scope narrowly – indeed, even more narrowly
than the Supreme Court (Rai, 2003) – does call into question a simple bias narrative.
Allegations of bias also pose familiar baseline problems. In the case of the Federal Circuit,
the baseline problem may be particularly severe given that the legislation establishing the
court was arguably motivated by a ‘pro-patent’ view.
B. The Supreme Court
Scholars have delivered mixed assessments of the Supreme Court’s recent activism in
the patent arena. Some early assessments tended to be relatively favorable. Duffy (2010)
argued that the Court is highly influenced by the Solicitor General and that this influence
could be beneficial so long as executive branch views were not subject to significant
election-related shifts. Dreyfuss (2010) concluded that the interaction between the two
tribunals (Supreme Court and Federal Circuit) was ‘highly salutary.’

Lee (2010) cautioned, however, that the Supreme Court’s embrace of fact-intensive
standards over bright-line rules might impose significant cognitive burden on inferior
actors. Rai (2012) noted that the highly ex post nature of Supreme Court decision making
had the potential to impose significant retroactive changes. Eisenberg (2013) observed
that the Supreme Court’s extended foray into questions of patentable subject matter
appears likely to have a particularly dramatic impact.
IV. FUTURE RESEARCH QUESTIONS
The highly dynamic institutional ecosystem that has emerged in the last decade raises numer-
ous questions for further research. As noted in Section III, the shifting geographic dynamics
of district court power over patent cases merit study. Here I suggest three additional areas
ripe for inquiry. One involves the institutional dynamics that shape patent law decision
making at the Supreme Court. Second, the PTAB’s decision making provides rich territory
to mine. A third area of research might compare institutional interactions in patent law with
institutional interactions and power relationships in other technically complex areas.
A. The Supreme Court
As scholars have discussed, Supreme Court review of the Federal Circuit has often
followed the recommendations of the Solicitor General and has generally emphasized
standards and ‘standard law.’ Future work on the Supreme Court, and on the relationship
between the Supreme Court and the Federal Circuit, might engage further the enormous
legal and political science literature on Supreme Court decision making.
One potential axis of engagement, ideology, may not be immediately fruitful. Although
political scientists and legal scholars agree that ideology often plays a significant role at
the Supreme Court (Epstein et al., 2013), the Court’s patent decisions have not generally
split along obvious ideological lines. Indeed, many of the decisions have been unanimous,
or nearly unanimous. The exception to this general statement may lie in cases that operate
at the patent-antitrust intersection, such as FTC v. Actavis, 133 S. Ct. 2223 (2013) and
Kimble v. Marvel Entertainment, LLC, 576 U.S. __ (2015).
Unanimity or near unanimity is coupled with a heavy emphasis on showing the
internal coherence and external dominance of the Court’s own prior case law on patents.
In the Court’s view, its case law (including case law from the 19th century) has either not
been affected by, or has in fact been adopted by, legislative enactments. Scholars might
explore whether these phenomena arise because the Court wishes its message of power
over the Federal Circuit to be clear; because individual members of the Court do not have
particularly differentiated views on patents and are therefore not interested in expending
the effort and political capital necessary to write dissents; or for other reasons.
B. PTAB
The first decisions from the PTAB post-grant proceedings set up by Congress in the AIA
have now made their way to the Federal Circuit. The PTAB’s popularity also ensures that
the PTAB will be a major source of appeals to the Federal Circuit. Thus far, the USPTO

has not argued for Chevron (1984) deference to the substantive results of its post-grant
validity determinations. However, it has successfully argued for Chevron deference for
procedural rules, such as its controversial rule requiring PTAB judges to conduct claim
construction under a ‘broadest reasonable interpretation’ standard that is different from
that used by the district courts. USPTO assertions that certain decisions of the PTAB are
not subject to judicial review have also been controversial. Questions about allocation of
power between the PTAB, Article III district courts, and the Federal Circuit are likely to
make their way to the Supreme Court and represent a fruitful area of study.
Another area ripe for study is the empirical question of whether PTAB proceedings are
in fact improving the system by serving as an efficient and accurate substitute for Article
III litigation over validity and/or by allowing appropriate challenges to validity that would
simply not occur in Article III courts. In Volume II, my co-authors and I present empirical
data that may help inform such analysis.
C. Comparison with Other Technically Complex Areas
A third area of research might compare institutional interactions in patent law with such
interactions in other technically complex areas. To the extent scholars (myself included)
have engaged this question, we have generally argued that the administrative model of
decision making should apply. Because a full-fledged administrative model is likely to
be a political non-starter, however, scholars should consider whether other technically
complex areas successfully employ administrative procedure only in part.
Finding the appropriate comparison area may also require further thinking about
why patent law represents such a profound institutional challenge. To what extent does
complexity emerge because of law, because of scientific or technological facts, or because
findings of law and fact need to serve as imperfect proxies for promoting the economic
goal of innovation? Even more fundamentally, to the extent the proxies we choose can
involve different combinations of law and facts, can other complex areas give us insight
into how to create proxies that our imperfect institutions can implement?
REFERENCES
Anderson, Jonas. 2015. ‘Court Competition for Patent Cases,’ 163 University of Pennsylvania Law Review 631–98.
Anderson, Jonas, and Peter S. Menell. 2013. ‘Informal Deference: A Historical, Empirical, and Normative
Analysis of Patent Claim Construction,’ 108 Northwestern Law Review 1–84.
Benjamin, Stuart, and Arti Rai. 2007. ‘Who’s Afraid of the APA: What the Patent System Can Learn from
Administrative Law,’ 95 Georgetown Law Journal 269‒336.
Burk, Dan, and Mark Lemley. 2003. ‘Policy Levers in Patent Law,’ 89 Virginia Law Review 1575–696.
Burk, Dan, and Mark Lemley. 2009. The Patent Crisis and How Courts Can Solve It. Chicago: University of
Chicago Press.
Burstein, Michael. 2011. ‘Rules for Patents,’ 52 William & Mary Law Review 1747–806.
Chien, Colleen. 2008. ‘Patently Protectionist? An Empirical Analysis of Patent Cases at the International Trade
Commission,’ 50 William & Mary Law Review 63‒114.
Chu, Christian. 2001. ‘Empirical Analysis of the Federal Circuit’s Claim Construction Trends,’ 16 Berkeley
Dreyfuss, Rochelle. 1989. ‘The Federal Circuit: A Case Study in Specialized Courts,’ 64 New York University
Law Review 1–74.
Dreyfuss, Rochelle. 2010. ‘What the Federal Circuit can Learn from the Supreme Court ‒ and Vice Versa.’ 59
American University Law Review 787‒807.

Duffy, John F. 2000. ‘The FCC and the Patent System: Progressive Ideals, Jacksonian Realism, and the
Technology of Regulation,’ 71 University of Colorado Law Review 1071–151.
Duffy, John F. 2010. ‘The Federal Circuit in the Shadow of the Solicitor General,’ 78 George Washington Law
Review 518–52.
Eisenberg, Rebecca. 2013. ‘Prometheus Rebound: Diagnostics, Nature, and Mathematical Algorithms,’ 122 Yale
Law Journal Online 341–49.
Epstein, Lee, William M. Landes, and Richard A. Posner. 2013. The Behavior of Federal Judges: A Theoretical
and Empirical Study of Rational Choice. Cambridge: Harvard University Press.
Farrell, Joseph, and Robert P. Merges. 2004. ‘Incentives to Challenge and Defend Patents: Why Litigation Won’t
Reliably Fix Patent Office Errors and Why Administrative Patent Review Might Help,’ 19 Berkeley Tech. L.J.
943–70.
Fromer, Jeanne. 2010. ‘Patentography,’ 85 New York University Law Review 1444–520.
FTC (Federal Trade Commission). 2003. To Promote Innovation: The Proper Balance of Competition and Patent
Law and Policy. Washington D.C.: U.S. Federal Trade Commission.
Golden, John. 2009. ‘The Supreme Court as “Prime Percolator”: A Prescription for Appellate Review of
Questions in Patent Law,’ 56 UCLA Law Review 657–724.
Golden, John. 2014. ‘The USPTO’s Software Power: Who Needs Chevron Deference?,’ 66 Southern Methodist
Hall, Bronwyn. 2005. ‘Exploring the Patent Explosion,’ 30 Journal of Technology Transfer 35–48.
Hart, Henry M. and Albert M. Sacks. 1994. The Legal Process: Basic Problems in the Making and Application
of Law. New York: Foundation Press. (William N. Eskridge, Jr. and Philip P. Frickey, eds.) (prepared for
publication from 1958 Tentative Edition).
Henry, Matthew D., and John Turner. 2006. ‘The Court of Appeals for the Federal Circuit’s Impact on Patent
Litigation,’ 35 Journal of Legal Studies 85‒117.
Holbrook, Tim. 2011. ‘Patents, Presumptions, and Public Notice,’ 86 Indiana Law Journal 779‒826.
Jaffe, Adam, and Josh Lerner. 2004. Innovation and Its Discontents. Princeton: Princeton University Press.
Janis, Mark. 2001. ‘Patent Law in the Era of the Invisible Supreme Court,’ 2001 University of Illinois Law
Review 387–420.
Kesan, Jay P., and Gwendolyn G. Ball. 2011. ‘Judicial Experience and the Efficiency and Accuracy of Patent
Adjudication: An Empirical Analysis of the Case for a Specialized Patent Trial Court,’ 24 Harvard J. of Law
& Tech. 394‒468.
Klerman, Daniel M., and Greg Reilly. 2016. ‘Forum Selling,’ 89 Southern California Law Review 241‒315.
Kumar, Sapna. 2011. ‘The Other Patent Agency: Congressional Regulation of the ITC,’ 61 Fla. L.Rev. 529‒80.
Lee, Peter. 2010. ‘Patent Law and the Two Cultures,’ 120 Yale Law Journal 2–82.
Lefstin, Jeffrey. 2010. ‘The Constitution of Patent Law: The Court of Customs and Patent Appeals and the
Shape of the Federal Circuit’s Jurisprudence,’ 43 Loyola of Los Angeles Law Review 843–92.
Lemley, Mark A. 2001. ‘Rational Ignorance at the Patent Office,’ 95 Northwestern University Law Review 1–34.
Lemley, Mark A. 2010. ‘Where to File Your Patent Case,’ 38 AIPLA Quarterly Journal 1–37.
Lemley, Mark. 2013. ‘Why Do Juries Decide if Patents Are Valid?,’ 99 Virginia Law Review 1673‒735.
Long, Clarisa. 2009. ‘The PTO and the Market for Influence in Patent Law,’ 157 University of Pennsylvania
Masur, Jonathan. 2010. ‘Regulating Patents,’ 2010 The Supreme Court Review 275–326.
Masur, Jonathan. 2011. ‘Patent Inflation,’ 121 Yale Law Journal 470–532.
Matal, Joseph. 2012. ‘A Guide to the Legislative History of the America Invents Act,’ 21 The Federal Circuit
Bar Journal 435–513, 539–653.
Moore, Kimberly A. 2001. ‘Are District Court Judges Equipped to Resolve Patent Cases?,’ 15 Harvard Journal
of Law & Technology 1–39.
Moore, Kimberly A., and Mark A. Lemley. 2004. ‘Ending Abuse of Patent Continuations,’ 84 Boston University
Nard, Craig Allen. 1995. ‘Deference, Defiance, and the Useful Arts,’ 56 Ohio State Law Journal 1415–509.
Nard, Craig Allen. 2000. ‘A Theory of Claim Interpretation,’ 14 Harvard Journal of Law & Technology 1–82.
Nard, Craig Allen. 2010. ‘Legal Forms and the Common Law of Patents,’ 90 Boston University Law Review 51–108.
Nard, Craig Allen, and John F. Duffy. 2007. ‘Rethinking Patent Law’s Uniformity Principle,’ 101 Northwestern
NAS (National Academy of Sciences). 2004. A Patent System for the 21st Century. Washington, DC: The
National Academies Press. (Stephen A. Merrill, Richard C. Levin, and Mark B. Myers, eds.)
Nock, Jennifer P., and Sreekar Gadde. 2011. ‘Raising the Bar for Nonobviousness: An Empirical Study of
Federal Circuit Case Law Following KSR,’ 20 Federal Circuit Bar Journal 369–408.
Petherbridge, Lee. 2009. ‘Patent Law Uniformity?,’ 22 Harvard Journal Law & Technology 421–73.
Rai, Arti K. 2000. ‘Addressing the Patent Gold Rush: The Role of Deference to PTO Patent Denials,’ 2
Washington University Journal of Law and Public Policy 199–227.

Rai, Arti K. 2003. ‘Engaging Facts and Policy: A Multi-Institutional Approach to Patent System Reform,’
Rai, Arti K. 2009. ‘Growing Pains in the Administrative State: The Patent Office’s Troubled Quest for
Managerial Control,’ 157 University of Pennsylvania Law Review 2051–81.
Rai, Arti K. 2012. ‘Patent Validity Across the Executive Branch: Ex Ante Foundations for Policy Development,’
61 Duke Law Journal 1237–81.
Rai, Arti K. 2013. ‘Improving (Software) Patent Quality through the Administrative Process,’ 51 Houston Law
Review 503‒43.
Rantanen, Jason. 2013. ‘The Federal Circuit’s New Obviousness Jurisprudence: An Empirical Study,’ 16
Rooklidge, William C., and Matthew F. Weil. 2000. ‘Judicial Hyperactivity: The Federal Circuit’s Discomfort
with Its Appellate Role,’ 15 Berkeley Technology Law Journal 725–52.
Schwartz, David L. 2008. ‘Practice Makes Perfect: An Empirical Study of Claim Construction Reversal Rates
in Patent Cases,’ 107 Michigan Law Review 223–84.
Schwartz, David L. 2009. ‘Courting Specialization: An Empirical Study of Claim Construction Comparing
Patent Litigation before Federal District Courts and the International Trade Commission,’ 50 William &
Mary Law Review 1699–1737.
Thomas, John R. 2003. ‘Formalism at the Federal Circuit,’ 52 American University Law Review 771–810.
Vacca, Ryan. 2011. ‘Acting Like an Administrative Agency: The Federal Circuit En Banc,’ 76 Missouri Law
Review 733–61.
Vertinsky, Liza. 2010. ‘Comparing Alternative Institutional Paths to Patent Reform,’ 61 Alabama Law Review
501‒52.
Wagner, R. Polk, and Lee Petherbridge. 2004. ‘Is the Federal Circuit Succeeding? An Empirical Assessment of
Judicial Performance,’ 152 University of Pennsylvania Law Review 1105–80.
Wasserman, Melissa F. 2011. ‘The USPTO’s Asymmetric Incentives: Pressure to Expand Substantive Patent
Law,’ 72 Ohio State Law Journal 379–435.
Wasserman, Melissa F. 2013. ‘The Changing Guard of Patent Law: Chevron Deference for the PTO.’ 54 William
& Mary Law Review 1959‒2019.
Cases

Chevron v. Natural Resources Defense Council, 467 U.S. 837 (1984).
Cybor Corp. v. FAS Technologies, 138 F.3d 1448 (1998).
Dennison Manufacturing v. Panduit, 475 U.S. 809 (1986).
Dickinson v. Zurko, 527 U.S. 150 (1999).
eBay Inc. v. MercExchange, LLC, 547 U.S. 338 (2006).
FTC v. Actavis, 133 S. Ct. 2223 (2013).
Highmark Inc. v. Allcare Health Mgmt. Sys. Inc., 572 U.S. __ (2014).
Kimble v. Marvel Entertainment, LLC., 576 U.S. __ (2015).
KSR International Co. v. Teleflex, 550 U.S. 398 (2007).
Markman v. Westview Instruments, 52 F.3d 967 (1995).
Markman v. Westview Instruments, 517 U.S. 370 (1996).
Oil States Energy Services LLC v. Green’s Energy Grp. LLC, 584 U.S. __ (2018).
State Street Bank v. Signature Financial Group, Inc., 149 F.3d 1368 (1998).
TC Heartland LLC v. Kraft Food Brands LLC, 137 S. Ct. 1514 (2017).
Teva Pharm. USA, Inc. v. Sandoz Inc., 135 S. Ct. 831 (2015).
Administrative Procedure Act, 5 U.S.C.A. §§ 500 et seq.

Federal Rule of Civil Procedure 52(a).
Leahy-Smith America Invents Act of 2011, Pub. L. No 112-29, 125 Stat. 284 (codified in scattered sections of
35 U.S.C.).

18. The economics of collective management
Daniel Gervais*
Contents
I. Defining Collective Management

A. Collective Management Organizations
B. Collective Management Functions
1. Core functions
2. Ancillary functions
II. Economic Analysis of Collective Management
A. The Economic Justification for Collective Management
B. Valuing a Repertory of Works in a Collective Management Context
C. Economic Models
D. Valuing Individual Works in a Collective Management Context
E. Measuring the Efficiency of CMOs
III. Collective Management of Online Uses
References
I. DEFINING COLLECTIVE MANAGEMENT
A. Collective Management Organizations
One should begin by defining the object of the chapter. What is a copyright ‘ collective’—
or what I prefer to refer to as a Collective Management Organization (CMO)?1 The
World Intellectual Property Organization (WIPO) Glossary mentions that there are
various models of collective management, but that at its most generic meaning a CMO
is a form of exercise of rights ‘where licenses are available from a single source’ (WIPO,
* Milton R. Underwood Chair in Law and Director of the Vanderbilt Intellectual Property
Program, Vanderbilt University Law School. The author is grateful to Jean Y. Xiao, PhD candidate
(Law and Economics) at Vanderbilt Law School for her most useful research assistance. All errors
are the author’s sole responsibility.
1
National laws using this more modern expression are now fairly common, especially in
countries that amended their copyright legislation recently. The term ‘Collective Management
Organization’ is used (as of June 2015) in at least the laws of Andorra, Azerbaijan, Bosnia and
Herzegovina, Burkina Faso, Cambodia, Cameroon, Georgia, Mongolia, Niger, Nigeria, Poland,
Romania, Slovakia and Zimbabwe. It is also the term used in Directive 2014/26/EU of the European
Parliament and of the Council of 26 February 2014 on collective management of copyright and
related rights and multi-territorial licensing of rights in musical works for online use in the internal
market, 84/72, 2014 O.J. (Eur.) (2014 EU Directive).
489

2003, p. 275). There are also a number of national statutes that define the notion directly
or indirectly.
An example of a direct definition is contained in the Canadian Copyright Act, which
defines ‘collective society’ as a ‘society, association or corporation that carries on the
business of collective administration of copyright [. . .] for the benefit of those who, by
assignment, grant of license, appointment of it as their agent or otherwise, authorize it to
act on their behalf in relation to that collective administration.’2 It adds that a collective
society operates a licensing scheme and/or ‘carries on the business of collecting and distrib-
uting royalties or levies payable pursuant to this Act.’ The US Copyright Act by contrast
only defines ‘performing rights society’ as an ‘association, corporation, or other entity
that licenses the public performance of nondramatic musical works on behalf of copyright
owners of such works’ (17 U.S.C. §101). A number of national laws define CMOs indirectly
by requiring that an entity be approved before operating as a CMO, which then requires
an administrative decision that the entity applying for approval is in fact a CMO—even
absent a formal statutory definition.3 If one were to limit the analysis to these definitions,
many types of entities could qualify as CMOs. Book and music publishers come to mind,
for example: They obtain transfers or exclusive control of and manage the rights of book
and music authors. Yet they are not considered as CMOs. The question is, why?
The answer lies in defining CMOs using elements of the statutory definitions above as
describing only one side, which one could call the positive side, of the ‘CMO coin.’ CMOs are
indeed in the business of licensing a repertoire of copyright rights (on works or other objects
created by several authors or other right holders), as other types of entities do as well.4 There
is, however, a negative side to the definitional coin, which is that a CMO is not in the primary
business of commercially exploiting individual works or objects of related rights. It licenses
others (users) who then commercially exploit the works. This explains why a book publisher,
for example, is not a CMO even though it does manage a repertoire of rights, because the
publisher also (and primarily) exploits the copyright works in the marketplace.
While this negative aspect (i.e., non-exploitation) is not present in the two statutory
definitions above, it is reflected in the definition contained in the 2014 EU Directive on
collective management (2014 EU Directive, art. 3). The Directive defines a CMO as an
organization that manages ‘copyright or rights related to copyright on behalf of more than
one right holder, for the collective benefit of those right holders, as its sole or main purpose.’
2
Copyright Act, R.S.C., c. C-42, s. 2. (1985) (Can.).
3
This is the case in several national laws. Some provide detailed criteria for approval; some do
not. Compare Australia, section 135ZZI of the Copyright Act, which defines a collecting society
(for purposes of retransmission) as ‘a body that is, for the time being, declared to be a collecting
society under section 135ZZT,’ see also sections 135ZZZF and 182B (Copyright Act (Act No.
63/1968) (Austl.)); in Kenya, Part VII of the Copyright Regulations, 2004 (Copyright Regulations
(Act No. 9/2004) (Kenya)); in Malta, Control of the Establishment and Operation of Societies
for the Collective Administration of Copyright Regulations (Copyright Regulations (Act No.
239/2016) (Malta)); in Saint Lucia, section 109A of the Copyright Act (Copyright Act (Act No.
7/2000) (St. Lucia)).
4
Here, to simplify, this includes related rights. That is, the rights of performers and sound record-
ing producers, where they exist as distinct rights. It might also include the rights of ‘producers of the
first fixation of films’ and also broadcasters in certain jurisdictions (WIPO, 2003, p. 307). The expres-
sion may also include the rights of makers of databases protected by a sui generis right in Europe.

The economics of collective management 491
These two (positive/negative) sides of the coin are functional in nature. Are there
structural components to the definition as well? The 2014 EU Directive suggests that
there are—at least in the EU context. The definition contained in the Directive also
requires that a CMO be (a) owned or controlled by its members and (b) organized on a
not-for-profit basis (WIPO, 2003, p. 307). This rule is not observed uniformly worldwide,
however. Hence, it seems better for our purposes to focus primarily on functions rather
than structure to define collective management (WIPO, 2003, p. 307).
B. Collective Management Functions
CMOs perform several functions. Those functions can be separated into core and ancil-
lary functions. ‘Ancillary’ should not be understood in this context to mean secondary or
unimportant; it only means that the functions on this list are not essential to the operation
of all CMOs.
1. Core functions

Let us begin with the core functions. They are:
(i) Obtaining the authority to license and/or collect remuneration;

(ii) Setting prices;
(iii) Licensing users;5
(iv) Obtaining relevant usage data, if necessary by developing and applying survey or
similar statistical methodologies;
(v) Developing and applying a method to distribute funds to represented right holders
and distributing payments with appropriate data reporting.
Each of these functions require some explanation.
(i) Most CMOs operate by delegation of authority to license (including by assignment in

certain cases) from those who own the rights. For example, music authors and publish-
ers authorize CMOs that manage music performing rights (ASCAP, BMI and SESAC
in the United States, PRS in the UK, GEMA in Germany, JASRAC in Japan, and so
on) to license on their CMO’s repertoire on their behalf. The terms are generally set in
the contract signed when an author or publisher ‘joins’ the organization (as a member,
affiliate, or the like). However, some rights are merely rights to obtain remuneration,
not exclusive rights to authorize or prohibit a use. For example, in several countries
record producers and/or performers only receive an equitable remuneration for the
broadcasting of their sound recordings, as provided for in the Rome Convention.6 In
such cases, a CMO will typically be appointed by law or the holders of the equitable
remuneration right to receive and distribute the payment. This is also typical in the
5
As discussed later, this includes enforcement of licensing contracts with the CMO (e.g., if
the user refuses to pay) but does not necessarily include broader enforcement efforts.
6
International Convention for the Protection of Performers, Producers of Phonograms
and Broadcasting Organizations, done at Rome on October 26, 1961, art. 12. As of June 2015, 92
countries were party to this Convention, but not the United States.

case of a compulsory license, such as the license for certain non-interactive digital
transmissions of sound recordings under US law, which led the Recording Industry
Association of America, Inc. (RIAA) to set up SoundExchange, Inc., to collect and
distribute the payments received from Pandora and others.7 However, CMOs often
operate only within the borders of their country. They obtain foreign rights by signing
contracts with similar organizations located in other jurisdictions. These agreements,
known as reciprocal representation agreements, provide for an exchange of rights and
most often also for the exchange of payments, that is, fees for the use of works from
country A used in country B will be paid by the CMO in country B to the CMO in
country A which will then pay the right holder(s) in country A. In some cases, CMOs
especially in very small markets may decide to lower costs by exchange only the author-
ity to license without reporting back, under so-called ‘B’ agreements (Pavel, 2006,
p. 226). This type of agreement is best seen as a temporary crutch for new CMOs in
smaller markets and, as normative matter should be discontinued as soon as possible,
a view which the 2014 EU Directive seems to support (WIPO, 2003, p. 281).
(ii) The second function is to set repertory prices. In the case of equitable remuneration
and compulsory licenses, a judicial or administrative mechanism usually sets the
price. The CMO may well be party to the relevant proceedings. In other cases (i.e.,
where the CMO manages ‘exclusive’ rights), users and the CMO may agree on a
price, or they may not. If they do not, the CMO may need to initiate a process to
get a governmental authority to set the price and perhaps other conditions of the
applicable rate or tariff. In the United States, ASCAP and BMI are governed by
different antitrust consent decrees, each of which appoints a federal judge in the
Southern District of New York to set rates (tariffs) for the use of musical works,
while SoundExchange operates under an explicitly separate system, outside of
antitrust scrutiny, to set prices for use of the sound recording on which the music is
used (Gervais, 2011, pp. 423–9). It is obvious to several observers that this discom-
bobulated system makes little economic sense.8
(iii) The third function is identifying and licensing users. If no rate or tariff exists (e.g.,
for a transactional use), this will include pricing, which may be set by the right
holder.9 This function includes decisions about filing lawsuits against users, by
which I mean both test cases that may be necessary if genuine debates exist about
the scope of exceptions and limitations,10 and simpler collection cases against
7
See 37 C.F.R. § 260 (2003); Determination of Reasonable Rates and Terms for the Digital
Performance of Songs, 68 Fed. Reg. 117, 36469-70 (June 18, 2003) (to be codified at 37 C.F.R. §
260) (noting that the proposed terms shall govern SoundExchange, the collecting rights entity that
was formed from the designated RIAA collective, in its capacity as the sole agent designated to
receive royalty payments from the three subscription services that were parties to the proceeding).
8
See U.S. Copyright Office (2015). One of the many recommendations in the report is to
migrate all rate-setting to the Copyright Royalty Board; 37 C.F.R. § 260, p. 4.
9
Indeed, even if there is a tariff, the law may allow a mutual agreement to override it (17
U.S.C. §112(e)(5)).
10
This happened in the United States in cases concerning photocopying (Am. Geophysical
Union v. Texaco Inc., 60 F.3d 913 (2d Cir. 1994); Patry, 1995). It was also the case more recently in
Canada, where the Supreme Court issued five opinions on the same day (July 12, 2012) all dealing
with collective management and the scope of exceptions and limitations (Gervais, 2013).

recalcitrant users who refuse to pay. The latter set of decisions (about filing lawsuits
against individual users who refuse to pay) forms part of a broader equation. CMOs
can spend more money to license and obtain usage data from smaller users. This
increases collections and costs. It may also create a backlash. Indeed, this type of
reaction is arguably what led the US Congress to exempt home-style equipment
used in bars, hotels, restaurants and supermarkets from paying music performing
rights in the United States, a decision which was later found to be in violation of
the United States’ obligations as a member of the World Trade Organization (17
U.S.C. § 115(b); Yu, 2003, p. 238). There is a flip side to this coin, however. If a
CMO reduces costs and political tension by limiting its licensing efforts to major
users, then right holders, especially larger ones, may see little value in working with
the CMO and may prefer direct licensing because they too can reach major users.
Reducing transaction costs in licensing multiple users and operating as a part of a
global network of CMOs is arguably the raison d’être of CMOs.11
It is also worth noting in connection with the licensing function that CMOs rarely
operate under a legal regime that actually obliges users to pay them. Some systems
offer a liability regime that limits the users’ obligation to pay the tariff and take up a
license, if the user so chooses. Nothing prevents a user from not paying if she thinks
that she can validly assert an exception or limitation.
(iv) The fourth function concerns usage data. Those data will be collected in one of
two ways: capture of all or almost all usage data, on the one hand, or surveys, on
the other hand. This means that some users will report all or almost all usage, for
example when a transactional license is obtained. For example, a university produc-
ing paper or electronic ‘course packs’ will license from a reprographic organization
(in the United States, Copyright Clearance Center, Inc.) and pay the per page fee
set by the right holder in each publication used.12 Even a large user such as a radio
broadcaster is often able to electronically report material used automatically because
codes and/or detection software are used to identify what was used. In such cases,
the CMO can match payments to material used and pay the relevant right holder,
making it easier to ‘follow the dollar.’13 If actual usage data are not reported, surveys
and other statistical methodologies will be used. This is often necessary when a
repertoire or blanket license is used.14 For example, for use of music in bars, hotels
and restaurants, this approach may be used.
It seems fair to posit that the more data are obtained, the higher the quality of the
distribution to right holders of material used, but the higher the cost. Comparing
11
The WIPO Glossary (2003, pp. 274–5) notes that collective management is used where ‘the
exercise of rights is impossible or highly impracticable on an individual basis.’
12
See Copyright Clearance Center, Pay-Per-Use Permissions (May 15, 2018), accessed March
27, 2019 at http://www.copyright.com/academia/pay-per-use/.
13
The expression often used to match every dollar to the use of a work and pay its right holder
(Massarsky, 2013).
14
The term repertoire license to designate the right to sue all rights to works or other objects
in the repertoire of the CMO is correct. The term ‘blanket’ suggests that all material in the relevant
category is covered. This is usually incorrect, unless users are protected by law for use of mate-
rial not formally part of the CMO’s repertoire but of the same nature, such as in the Extended
Collective License (or Extended Repertoire) scenario.

ASCAP and BMI, two of the US CMOs in the field of music performing rights, a
commentator noted that: ‘While ASCAP is dedicated to great precision generated
from small samples, BMI looks to include a greater number of performances absent
the rigorous precision [. . .] BMI’s recognized performances is over eight times
greater than ASCAP’s’ (Massarsky, 2013).
(v) Finally, distribution must be made to represented parties. This includes methods
of aggregating and reporting usage data to right holders, making decisions about
payments of small sums, how much and for how long to look for right holders
who have moved and failed to notify the CMO or cannot otherwise be located, etc.
CMOs also may decide to use the funds for purposes other than paying individual
right holders, typical for social or cultural purposes. Because this function is common
but not essential, it is considered an ancillary function, as discussed below.
2. Ancillary functions

As noted above, the ancillary functions are those that are not essential per se to operate as
a CMO, but they are common and, in some cases, are almost necessary unless performed
by a third party.
Many CMOs devote part of the funds collected to social and cultural purposes (Gervais,
2018). These vary greatly in scope, from award ceremonies and stipends to group medical
insurance and even a pension scheme (WIPO, 2003, p. 281). The conventional view is that
these functions are acceptable if desired by the right holders and decided as such by a
democratic process (WIPO, 2003, p. 281).
Another key ancillary function is education. When a new CMO is set up, particularly
in a new market such as a developing country, users who never had to pay for a license to
use copyright material may need to understand why they need to pay. This includes telling
them fairly about the law but also explaining the relevance of copyright, for example to
the country’s cultural production.
II. ECONOMIC ANALYSIS OF COLLECTIVE MANAGEMENT
From an economic viewpoint, copyright exists to solve the problem of underproduction

of creative works. Underproduction results from copyright owners not being able to
appropriate value from their works. If a CMO operates as an unregulated cartel (i.e., an
effective monopoly), then there is distortion in the form of a loss to welfare (deadweight
loss). Whether it is efficient to have collective management (assuming there is a right
that needs to be licensed for a user to operate legally) depends on whether the gain from
capturing the secondary demand from unauthorized uses outweighs the distortion. In
other words, competition in the primary market of traditional, authorized uses may still
be beneficial and welfare-enhancing (in part due to lower transaction costs of individual
enforcement), and the efficiency lies in having a collective to exploit secondary demand
where transaction costs make individual enforcement almost impossible (Smith, 1986).
That leaves a number of questions to be examined: How does one define efficiency in
this context? How does it apply to the value of individual works managed by a CMO and
to the value of a repertory of works licensed (as a bundle), and, last but not least, how
does one measure efficiency?

A. The Economic Justification for Collective Management
The traditional economic justification for collective management is the reduction in

transaction costs and related economies of scale in administering copyrights when a
single source holds the information needed to administer and enforce copyrights. In
simple terms, if there are n right holders, then n −1 rounds of license negotiation costs
can be saved (Snow and Watt, 2005). Costs of licensing typically arise from locating
copyright owners and obtaining the information needed to negotiate a price. Copyrights
are typically negotiated before works are used or performed, so most licenses are based
on information such as a licensee’s past use of music, potential signals of the work’s
popularity and the parties’ subjective predictions. CMOs typically reduce transaction
costs by issuing repertory licenses. This decreases the time and effort involved in licensing
negotiations. Due to the CMOs’ seriousness about enforcing licenses, they deter users
from infringement and, in turn, lower enforcement costs (Thorpe, 1998). However, aggres-
sive enforcement, coupled with the monopolistic behavior of CMOs, can lead to increased
prices and lower consumption.
Then, because CMOs are natural monopolies, inefficiencies may occur due to pricing
incentives. There is an overall welfare gain if total savings in transaction costs, which
should translate to lower licensing prices, offset higher prices due to the monopoly power
of collecting societies. Other factors that can mitigate the monopolistic inefficiencies
include regulation, bilateral monopoly (i.e., where users aggregate into a negotiating
entity or form their own CMO15) and price discrimination (with a view to extracting as
much surplus as possible from each user willing to license) (Handke and Towse, 2007).
Analyses of economic efficiency should not overshadow other aspects of collective
management. For individual authors, a CMO may provide direct access to licensing
markets in which they have (individually) little clout and expertise. The CMO allows them
to join forces and have increased clout, which some large users and scholars lament, argu-
ing that copyright owners are those who are too powerful (Lunney, 2001). In such cases,
a CMO is best seen as a trade organization or a union (Kretschmer, 2002). Indeed, that
was the rationale initially used by the Court of Justice of the European Union to accept
that CMOs’ contracts with users were not anticompetitive, based on the rationale that
collectives needed to protect their members from powerful users (e.g., major broadcasting
companies) (Cohen Jehoram, 2001).
The next step in the analysis is to see how CMOs set prices for the repertory of works
they license, the economic efficiency of this valuation process and how individual works
within that repertory can be valued. The final step will be to see whether, and if so how,
the efficiency of CMOs can be measured.
B. Valuing a Repertory of Works in a Collective Management Context
How does one value the use of a repertoire (say, the world’s repertoire of music)?
Background music adds to a user’s experience in a store, for example, but how much is it
15
As was the case in the United States with BMI (BMI’s Timeline through History, accessed
March 27, 2019 at http://www.bmi.com/about/75_years).

worth? Empirical studies showing an increase in consumption of n percent when music of

a particular genre is played can be used to establish a reference point.
For a user whose business involves dealing in protected works (e.g., a music station on
broadcast radio), the user has an incentive to lower the cost of the input. In a normal
market situation, a user would look at alternatives. In the first case above (background
music), the user may decide to opt not for a CMO license but rather for alternative ser-
vices offering ‘pre-cleared’ (pre-licensed or originally composed) music services such as
Muzak. In contrast, for a broadcaster of music, the road to the CMO is a passage obligé.
The alternative of letting each songwriter negotiate with each radio station (suggested
by a few academics) hardly seems realistic both on feasibility and normative grounds.
The former because of transaction costs; the latter because of the relative clout and
copyright expertise of the parties. Intermediation is essential in that context and CMOs
can provide it.
Should CMOs be the answer? One way to look at this question from an economic
perspective is to ask whether repertory licensing and pricing are structurally inefficient.
Even beyond transaction costs reduction for both right holders and users, the answer
seems negative. As Landes and Posner (2003) note, if there are declining average costs and
marginal cost pricing is not a feasible option, setting a price in a way that is discriminatory
(extracting consumer surplus) is not less efficient than setting a price to average total cost.
The only situation in which antitrust authorities would then need to step in is when the
price discrimination has an additional exclusionary effect.
Efficiency considerations aside, setting the value and/or prices for a repertory is often
difficult, owing mostly to the absence of comparable markets. What compares to the value
of music for a radio broadcaster, for example? The market for downloads or streams
is related to the use of music in radio broadcasts, but it is not fully comparable. Thus,
mispricing is arguably more likely and/or more severe (Liebowitz, 2005).
There are of course some applicable parameters. First, the prices in the market for
music performing rights are tied to the overall size of the market in which they are
licensed. Second, operators of streaming services or broadcasters will normally attempt
to maximize revenue, but how much of that is music? Should, say, music be a fixed priced
input independent of the licensee’s income (like power or rent)? Or is it better and fairer
to ask the licensee to pay a portion of its income? One might intuit that users would prefer
a fixed price. Indeed, in the Pandora-related rebates in the United States, the licensee
opted for the latter, for a variety of reasons from which it is difficult to draw hard and fast
conclusions (Marshall, 2013). One way or the other, however, a decision must be made on
sharing the surplus generated by the exploitation of the rights.
C. Economic Models
Organizations whose business is to determine the value of repertories of rights have cre-
ated useful models. For example, an interesting approach was used by the New Zealand
Copyright Tribunal in a 2010 case. It discussed the game theory model suggested by Lloyd
Shapley (as presented to the tribunal by Richard Watts), noting the following:
The Shapley model is based on game theory and is named after the economist who created
it, Lloyd Shapley. The model measures the contribution of each participant to a licensing

a rrangement and then calculates a split of the surplus. The outcome of this split is what is
known as the Shapley value. The willing buyer and willing seller premise means that the model
operates on the basis of co-operation between the parties. [. . .] Dr. Watt summarized the steps
to calculating the Shapley value as follows:
derive Radio’s maximum willingness to pay (WTP)
derive PPNZ and APRA’s [the music PRO’s] minimum willingness to accept (WTA)
derive the sharable surplus
share the surplus and
derive a revenue tariff [. . .]
The surplus is WTP − WTA. Sharing the surplus is based on the premise that none of the
parties can achieve any surplus at all unless the others are present. [. . .] We are persuaded by
the Respondents’ experts’ comments that the Shapley model was the better and more applicable
of the two models, although as noted above they doubted that sufficient data was available to
justify the use of the model at present. There are some aspects of the model suggesting that it
is not the only way that we should calculate what a reasonable royalty in the circumstances is.
Other circumstances are relevant. In any event, the model is just that, a model. Its utility is to
assist in arriving at the reasonable rate. (Phonographic Performances (NZ) Ltd v. Radioworks
Limited, NZCOP 1 (2010))
At bottom, this model does what any such model should do, namely identify the shareable
surplus and acknowledge that factors must be found to share it appropriately. Even if the
CMO asks for all of the shareable surplus and the user/exploiter says it wants to keep all
of it, they both know an equilibrium must be found, hence the potential role of economic
models. The CMO and the user are fighting to move the needle on the percentage of the
surplus that will be paid to the CMO and, via the CMO, to the creators of the licensed
content. Clearly, they also fight over WTP and WTA.
The Canadian Copyright Board has also discussed applicable models in a number of
decisions. In a 2009 decision, it considered and rejected the willingness to pay model
because ‘the willingness-to-pay for a particular input does not appear to be a reliable
proxy for the willingness-to-pay of another’ (Copyright Board of Canada, 2009, para.
138). In the same decision, it refused to apply the Shapley approach to value of the music
because, while interesting, it ‘relies heavily on data from a survey where respondents
were questioned on hypothetical scenarios,’ adding that ‘[u]nfortunately, we do not have
enough information to be able to test the variations and the stability of this model’ (para.
146).
The Board then considered what it termed an ‘efficiency’ approach presented by two
expert witnesses, Drs. Agrawal and McHale. They determined in their report to the Board
the size of the ‘trade-off between the loss of subscribers arising from higher subscription
costs and the increase of music production resulting from increased revenues for the
industry’ (para. 152). The Board did not reject their model. It found the approach useful
and interesting but, in the case at hand, found that the dataset was insufficiently convinc-
ing to justify using the proposed methodology. The Board ended up applying digital pay
audio as a proxy benchmark (para. 185).
In a 2012 decision, the Board discussed whether it should use a Nash equilibrium to
split the surplus equally among parties to a rate-setting case concerning digital audio
services (Copyright Board of Canada, 2012, para. 70). The Board provided six reasons
not to use this model:

– F
irst, the record companies were not part of the bargaining mode presented by the
experts (paras. 72, 74);
– Second, in some cases the model generated a negative surplus and the ‘Nash bargain-
ing model cannot yield negative benefits for any of the players’ (para. 76);
– Third, the model presented did not factor correctly the relationship between mechani-
cal and performance royalties, an issue identified as problematic also in the United
States in a Copyright Office report;16
– Fourth, the model did not properly account for costs relating to collection of both
types of royalties (para. 76);
– Fifth, the model as presented did not offer a convincing value or ‘normal profit’ to
deduct from the shareable surplus (para. 77); and
– Finally, and relatedly, ‘profit’ data are inherently suspect, since online music services
operate as business lines within larger corporations. In addition, data from business
lines are never audited; they can thus be subject to manipulation for the purpose of
minimizing the liability under the tariff (para. 78).
The model was not rejected per se, but the Board did not find the application suggested
in the case convincing enough. This strongly suggests that the application of economic
models to the valuation of repertoires of licensed content is an area in need of further
research and refinement.
D. Valuing Individual Works in a Collective Management Context
In the case of many CMOs that operate repertoire licenses, the value of the license from
the user’s perspective is the value of the input in aggregate (i.e., the repertoire). It may be
hard to measure individual works that form part of the repertoire.
From the right holder’s standpoint, aggregating her works in a repertoire is far from
economically neutral. In the case of a music CMO licensing on behalf of songwriters
and publishers, the total surplus (i.e., total value minus the license fees and other costs) is
usually shared equally in the sense that each song has the same value, under a principle of
‘nondiscrimination.’ It has been argued that, in a situation where membership is closed
and the CMO maximizes surplus per member, it produces fewer than the socially efficient
number of songs (Besen et al., 1992). In other words, if membership is open or unlimited,
then the membership will grow until the surplus is dissipated; at this point, the collective
produces more than the socially efficient number of songs. From the licensee’s (user)
perspective, a repertory license is second best because cooperative pricing is not perfectly
competitive pricing (i.e., marginal cost equals marginal benefit). Licensees do not benefit
because the fee does not always reflect variations in the repertory. Even repertory licenses
are imperfect, because new members typically do not do promotional pricing, and these
licenses may favor certain technologies (Handke and Towse, 2007).
An increase in membership increases administrative costs even if authors or right
holders whose works are not used will not get paid. Moreover, if the user uses the same
amount of music, the interchangeability of the songs or works from her perspective
16
See U.S. Copyright Office, 2015, p. 5.

means that having more songs or works may not add value. Yet there will be less surplus
to distribute to right holders. This suggests that offering options to licensees (how many
songs will be used, and so on) may produce more efficient outcomes. But is it efficient for
the individual author or right holder?
If one were to see the world as meeting the following three conditions, (a) imperfect
product differentiation (many differentiated works but often with the same value), (b) the
number of market participants/creators is endogenous (i.e., market participants can enter
and leave) and (c) joint production (i.e., works are used jointly—seemingly in a comple-
mentary way to produce other works), then the circulation of works for which rights are
offered individually would decrease, while the circulation of works for which rights are
offered collectively would increase—that is, if the demand in the collective market is not
affected by royalty rates of other markets. This admittedly may not be how the actual
world functions—as with any theoretical model. Still, it is useful to see that in a world in
which these three conditions are met, a welfare analysis here leads to ambiguous results.
The result depends on how much consumers marginally value product variety versus
monetary incentive to create another product (Hollander, 1984). This strengthens the call
for additional research mentioned at the end of the previous section.
One can build on this analysis and argue that copyright’s purpose, and, therefore, a
function of licensing, is to promote the creation of a particular type of works. Professor
Lunney (2001), for example, has argued that marginal works (as opposed to non-marginal
popular works) are the ones that the US Constitution aims to incentivize. He also suggests
that more copyright revenue may in some cases mean excess incentives to create, leading to
overproduction, which in turn may lead to a lack of allocative efficiency—that is, scarce
resources not being used in places that society values more. Excess incentives may increase
popular works at the cost of ‘great’ works and result in popular creators having more
access to leisure goods, which may distract them from working more.
If that is seen as a problem, it is so for all of copyright not just those works that are
managed collectively. Was Dan Brown paid too much for the Da Vinci Code? By far the
bulk of his income came from his contract with his publisher and from sales managed
directly by the publisher, not via a CMO. In fact, a CMO may rebalance this inequality
if it so wishes by using some of its funds for social and cultural purposes or by adjusting
its distribution key. For example, the Norwegian music CMO TONO adds a multiplier
to so-called serious music and also to works first performed in Norway, thus ‘overpaying’
composers of serious music and, more often than not, Norwegian ones, as a form of
cultural subsidy.17
In a variation on this theme, Snow and Watt (2005) proposed a model that suggests that
proportional distribution (i.e., where each right holder represented by the CMO is paid
in direct proportion to the intensity of the use of her work minus the member’s share of
the transaction costs) does not benefit collective members from a risk-sharing perspective
(Bouchard, 2010). A distribution rule where the collective pays each member an equal
share of the final net income Pareto dominates the follow-the-dollar approach (which
they term ‘learn-and-distribute’). The equal-sharing rule can be modified to be more risk
17
Naturally, this must be done in accordance with national treatment obligations (Melichar,
1991).

effective if each member receives a portion of the net income that is determined by the
member’s risk tolerance level divided by the sum of risk tolerance levels of all members.
They also suggest that the reason that many collectives still use the learn-then-distribute
model and not an equal-sharing rule is because of moral hazard (i.e., no incentive to
produce a successful work if success is not tied to earnings) and/or adverse selection (i.e.,
when the composers are better informed of the probability of success of the work than
the collective, the ones who will be successful are not likely to join the collective if success
is not tied to earnings). Further, copyright holders may be overly optimistic and overvalue
their works. The authors suggest that a solution to adverse selection and/or moral hazard
would be to tie part of the royalties to song performance (to solve the adverse selection/
moral hazard problem) and part of the royalties to the success of the collective’s repertory
as a whole (to provide a risk-related benefit).
Other interesting approaches have been proposed to refine valuation equations. In their
discussion of the value of individual works, Parisi and Depoorter (2003) suggest that a
major difference exists between what they termed complementary works and substitutive
works. If two or more works (A and B) are needed to produce a third, the lack of price
coordination between owners of rights in A and B will lead to inefficiencies. As they note:
If the copyrights are in a relationship of complementarity in the production of a derivative work,

the competitive Nash equilibrium would generate an anticommons pricing problem. [. . .] The
anticommons equilibrium pricing is simply the outcome of a prisoner’s dilemma that individual
copyright sellers face when pricing their copyrights independently. As in a traditional prisoner’s
dilemma, the inability of copyright holders to co-ordinate prices produces both private and
social inefficiencies. Quite strikingly, in this case the competitive outcome is socially inefficient,
even if compared to the alternative monopoly equilibrium. Competitive pricing of complemen-
tary goods generates a substantially larger social loss than the monopolistic equilibrium. (Parisi
and Depoorter, 2003, p. 168)
Their model shows that, in the case of complementary products, competition actually
increases prices, decreases consumer surplus and decreases overall welfare. By contrast,
in the case of substitutes, competition decreases prices, increases consumer surplus and
increases overall welfare. This model can then be used to explore effects of collective
management on pricing. To quote Parisi and Depoorter again:
[C]opyright collectives’ [. . .] independent authority to fix the price of licences has an obvious
effect on the two equilibria considered above. In the complements scenario, the intermediary
would choose prices that are lower than the prices copyright holders would have chosen if
pricing independently from one another. The salient point is that the lower price charged by the
intermediary is beneficial to all individual copyright sellers, since it allows them to maximise the
total profit from the sale of their licences, improving upon the alternative anticommons result
reached in the absence of price co-ordination. The paradox—that the intermediaries’ price is
lower than one that would have been chosen by the owners and yet it increases their total profits
from the sale—can be understood by recalling that the anticommons equilibrium pricing is the
direct outcome of a ‘prisoner’s dilemma’ that individual copyright holders face when pricing
copyrights independently. (Parisi and Depoorter, 2003, p. 170)
In the case of substitutes, however,
price-fixing authority renders monopolistic pricing sustainable in a Nash equilibrium. The

resulting equilibrium favours copyright owners, who are able to maximise total profit from the

sale of their licenses, as would happen in a cartel. But such co-ordination is socially inefficient
compared to the alternative competitive (or oligopolistic) equilibrium, since it prevents beneficial
competition with the creation of a social deadweight loss. (Parisi and Depoorter, 2003, p. 171)
This leads one to the conclusion that competition between CMOs and right holders will
lead to lower repertory prices in the case of substitutive works, which seems verified in
that US prices (where music CMOs are nonexclusive licensees of right holders) tend to be
lower than in other jurisdictions (caeteris paribus) though other factors are also at play.
This does not, however, necessarily lead to the conclusion that CMOs should offer limited
licenses (sub-repertory or per work). There are transaction costs (which may need to be
absorbed by all right holders represented by the CMO) issues. It is, once again, an issue of
pricing: ‘As long as consumers may acquire cheaper bundled licenses, the availability of a
per-use license does not constitute an impediment to the solution of the complementary
oligopoly problem’ (Parisi and Depoorter, 2003, p. 173).
Should pricing models be regulated? Regulation is inextricably linked to values (cultural,
economic or otherwise) that the regulator may wish to enforce. If the regulator wishes
to infuse more funds into the creation of new works by domestic authors, then above-
competitive pricing or distribution rules that favor, say, new works first disseminated in
the country in question may be considered desirable. This form of regulation may be seen
as a form of subsidy. The subsidization argument is not sufficient to negate the benefits
of collective management. Unless one wishes to argue that creators should work for free,
the actual issue is valuing and monetizing their input, which in turn means valuing the
surplus they helped create and parameters for sharing it. Clearly, users value music, films
and books. Non-right holder intermediaries (Facebook, Google, and the like) monetize
this content to the tune of billions of dollars each year, and one rarely hears opposition
in academic circles or elsewhere to that monetization.
If the creator chooses to disseminate without payment because she is ‘just trying to be
heard,’ as a blog poster might for example, or because she is otherwise paid independently
of the exploitation of her work (a university researcher or professor) then monetization
of her work to her benefit may not occur at all. Authors are and have always been free
not to exercise, or to assign, their right, in whole or in part (Gervais, 2015). If, however,
one believes, as a normative matter, that high quality content requires an investment of
time and resources—including the years necessary to master one’s craft—and that on the
whole this is welfare-enhancing, then societally we need to develop ways to nurture this,
and a market-driven solution which monetizes at least part of the value enjoyed by users
and exploited by intermediaries for the benefit of the creators of that content must be
found.18 The interests of those creators is to maximize use and, consequently, to minimize
unnecessary restrictions on use. Licensing content through CMOs that represent them
(exclusively or jointly with other right holders) is often a good means to that end.
18
While this is admittedly a normative argument, it is not less normative to say, even as a form
of technological determinism, that professional creators should not be paid and that only major
Internet intermediaries are allowed to monetize their inputs because they hold the keys to the
Internet kingdom.

E. Measuring the Efficiency of CMOs
A key aspect of the economic analysis of collective management is to devise adequate

metrics to gauge the efficiency of CMOs. Inefficient CMOs cost more, and they are likely
to offer poorer services to both represented right holders and users. Hence, even if the
collective management model makes sense in a given context—especially if repertory
pricing is more efficient than individual transactions with individual right holders—an
inefficient CMO may negate any such benefits. This leads one, quite naturally, to ask how
to measure the efficiency of a CMO.
Rochelandet’s work has shown that determining the measure of the efficiency of a
CMO, even if one excludes soft criteria such as educational inputs and social and cultural
effects, is a multifaceted analysis (Rochelandet, 2003). Variables that can be considered
include management ratio, annual variation of collected sums, annual variation of
distributable sums, elasticity of distributions compared to collections, annual variation
of administration expenses, gross distribution ratio, net distribution ratio, collected
sums per employee (COPE), annual variation of COPE, distributed sums per employee,
average cost of an employee and collected sums per member (Rochelandet, 2003, p. 187).
Which measure is considered most appropriate is partly driven by normative concerns
(e.g., how much should the CMO spend on non-purely financial operations, if any)
and partly by context. For example, an open CMO that any right holder can join will
not learn much by using the revenue per member ratio. Within a similar field, however,
there may be useful comparables. CMOs in comparable markets administering the same
right may be comparable, up to a point, and excluding unusual short-term variations
caused by, for example, adaptation to and involvement in major changes in case law or
legislation.
Typical efficiency analyses use distributions (D) and total collections (C), as in R1 =
D/C. This basically measures overall administrative expenses. Because represented right
holders can agree to use some of the receipts for social and cultural purposes (SCP), a
more complete picture may be provided by R2 = D + SCP/C. In the absence of recognized
standards determining how to calculate SCP, however, comparisons between two CMOs
using R2 are potentially problematic.
While the ratio of collections to expenses seems a solid starting point, it also involves
analytical shortcomings. A CMO can go for the low-hanging fruit at lower cost but, as
already noted, it then reduces the incentive of larger right holders to work within its col-
lective scheme because they too can license those uses directly without paying the CMO
fee. It may be precisely in going after harder dollars that overall efficiency gains can be
made. For example, if CMOs A and B operate in similar markets, if total collections
of CMO A are $10 million and its costs $1 million (leaving $9 million for distribution),
while CMO B has collections of $12 million and costs of $2 million (leaving $10 million
for distribution), which one is more ‘efficient’ is not necessarily obvious. The 10 percent
expense ratio of CMO A seems better, but CMO B extracted (as a high cost but still at
a net gain for its represented right holders) an extra $1 million from a similar market by
working harder.
A CMO could also decide to use less data in preparing distributions, or could decide to
report less thoroughly and pay less frequently, thus improving its ratio but perhaps not in
a way that optimally serves right holders. They may prefer monthly or quarterly payments

W (T, c) 5 3
# `
[ W (s,T ) 2c ] f (s 0 c) ds
s (T )
to annual payments, for example, but this increases costs. Along similar lines, the speed
0 # around into distributions may also affect
`
with which receipts are processed and turned
expenses and the quality of services in an 0T inverse relationship.
s (T )
One of Rochelandet’s most interesting ratios tracks variations in distribution over
collections. Gains in performance and efficiency should be reflected using this ratio. If D
is distribution and C is collections, is can be expressed as
DD
R3 5
DC
A CMO might usefully measure variations in the ratio of its collections per employee over
time to measure improvements of efficiency within that CMO, and compare the ratio to
those of CMOs operating in similar markets.
Rochelandet acknowledges the limits of all his proposed ratios, in part because they
tend to aggregate and compare a number of non-commensurate performance indicators
(e.g., collections and membership). He suggests an alternative, namely, a data envelop-
ment analysis (DEA), which he defines as a ‘non-parametric non-stochastic approach
frequently applied in the field of non-profit organizations such as hospitals and schools’
(Rochelandet, 2003, pp. 187–8). This approach identifies collections, costs, employees and
the number of members as inputs, and distribution as the output. The approach can be
used to measure either output maximization or input minimization.
Rochelandet applied his analysis to a large dataset of major European CMOs. Not
surprisingly, it identified governance as a key factor affecting efficiency. According to
him, if a number of major right holders are involved in governance, greater efficiency
(measured according to the ratio above or the DEA method) will result. This, as already
noted, is likely a multifactorial effect and thus outcomes may vary significantly. It may
be because ‘professional’ right holders will pick better management, or because the
management wants to keep those major right holders ‘happy’ (in turn, because they may
have other options), and/or because those major right holders will provide useful inputs
into the operations of the CMO as an efficient ‘business.’ There is a significant downside
to giving major right holders such a degree of control, however. Namely, this is the risk
that the interests of other, non-major right holders will be neglected.
A policy proposal can be derived from this analysis. An optimal governance policy
(including one imposed by regulation) should provide represented right holders, especially
major or more ‘professional’ ones, with direct—though not overbearing because that
may be counterproductive—oversight of the CMO’s operations, but with appropriate
regulation to ensure that all represented right holders are treated fairly. It also requires
transparency and assurances that management is independent. Interestingly, this is not
far from what is contained in the 2014 EU Directive.
III. COLLECTIVE MANAGEMENT OF ONLINE USES
It is trite to say that the Internet has radically changed the way in which copyright material
is created, distributed and licensed. The technology itself allows individual right holders
to connect with and license individual users1 reducing, but certainly not eliminating,
M4754-Math Eqn.indd

transaction costs.19 If copyright is to serve its purpose online, it will not be, at least not
primarily, by taking down content. Content must be available; it must be found, accessed
(by download, streaming, and so on), and the author or other right holder might then try
to get paid, whether per work or as part of a subscription. It is the era of what Professor
Ginsburg has referred to as ‘permitted but paid’ (Ginsburg, 2014). In that world, as I have
discussed elsewhere (Gervais, 2004), the scarcity is not of physical carriers of copyrighted
material, but of connections between works and users who will value them most. There is
another scarcity, that of creators who have the time to create professionally, and for this
they require an income stream.20
New intermediaries whose business is predicated on keeping users online to view a
maximum number of ads are essential players in creating connections and value. However,
unlike traditional publishers or producers who acted as copyright holders, they are not
licensed users of the ‘content,’ often claiming a safe harbor as mere connectors. Excluding
access to content that a right holder considers infringing is not important to them as it
might have been to a traditional copyright intermediary. It may be counterproductive.
Whether the content is authorized or not is irrelevant, provided the user accesses it with
paid ads. Naturally, the notice-and-take-down system and other similar ones negotiated
in exchange for safe harbors means that they do take down content. The point is that this
is not something they would do if they did not have to.
If healthy and sustainable financial flows are to be created, content must be available
and accessible. Taking it down is not the priority of authors, either. Therefore, monetiza-
tion is. Here, copyright can find a new way of achieving its purpose. When it was used
to exclude (e.g., pirated physical copies), excludability was essential to maintain physical
scarcity but the real end was to have a successful marketplace. The online marketplace
responds differently. If it is correct to picture it as the coexistence of a large zone of
free content (blogs; professional content viewed in exchange for accepting ads) and paid
content, then many services may compete to provide access to that content. They will need
to be licensed, and that points to repertoire licensing. If CMOs play their cards well, their
relevance could thus increase.
For CMOs, this is nothing new. Whenever a right is managed collectively, excludability
is often illusory (Gervais, 2012). Essentially, users pay to use works in the collective’s rep-
ertoire, and the CMO often has no legal right to stop a user willing to pay the applicable
fee (Maurushat and Gervais, 2003). Indeed, why would a CMO say no to a user willing
to pay the applicable fee?
What does this mean for CMOs and online copyright? Copyright should not be seen
as a tool to stop (exclude) content or end-users. Copyright remains an exclusion tool for
dealings between (competing) professional entities or against major pirates. In the case
of pirates, however, the reach of copyright in the case of non-physical, Internet-based
distribution is restricted by the technology itself. Copyright works best as an exclusion
tool when rules are internalized by stakeholders (professional publishers, producers or
19
There was indeed much hype around Digital Rights Management as a substitute for col-
lective management 10–15 years ago, but it seems to have at least partly dissipated (Korman and
Koenigsberg, 1986).
20
Professor Mackaay (1990) has argued that there is a related scarcity in the resources neces-
sary to gather and generate certain types of information.

broadcasters presumably want to be seen as obeying the rules of the road—they are also
easy to sue) or when physical objects are involved (the typical example would be pirated
CDs or DVDs). Pirates are excluded from this possibility, which explains the ‘whack-a-
mole’ issue with efforts to curtail online piracy. New legal services (e.g., e-books or music
streaming) need to be licensed and can be shut down if they do not operate legally. But,
they may reappear. Here, one could argue that the most efficient outcome will be reached
if anyone is free to compete easily in the space, that is, if licenses are available for anyone
willing to pay applicable fees, as opposed to having a few major right holders pick winners
and losers by using their exclusive right.
Now that futile judicial attempts to prohibit end-users from using the Internet’s power
have been mostly abandoned, copyright can and perhaps should best be seen as a basis
for an entitlement to remuneration when use reaches the level of interference with ‘normal
commercial exploitation,’ the main limit set by the three-step test (Geiger et al., 2015).
The test, which sets limits in many international intellectual property instruments for
exceptions and limitations to copyright and other rights, is useful in this context because
it rests on two keys notions: a dynamic notion of normalcy (of commercial exploitation)
and a notion of commercial harm. Their combination allows authors and other right
holders to claim payment for massive Internet uses not covered by an exception such as
fair use or fair dealing.
Collective licensing is not antithetical with adequate exceptions and limitations.
Indeed, it may be easier to safeguard space for those exceptions and limitations by set-
ting uniform tariffs than in adhesion-type end-user licenses that often restrict available
exceptions. Whether or not such terms are enforceable, they can have a chilling effect
on users. For instance, if a tariff was sought for reprographic use and Internet access
in schools and universities, the tariff-setting authority could exclude from the license
(or at least in calculating the payment) any use covered by research or educational
exceptions. This is precisely what happened in Canada in 2015 (Copyright Board of
Canada, 2015).
Exclusion was, and is, a means to an end. Part of that end, with respect to the online
exploitation of copyrighted content, is to ensure proper financial flows to professional
authors, those whose livelihood depends on monetizing access to their creative output,
when their works are successful in the marketplace (generating surplus). Copyright can
best serve this goal by allowing online uses but generating financial flows to creators.
This, in turn, depends on the availability of efficient intermediaries that can license the
necessary rights.
REFERENCES
Besen, Stanley, Sheila Kirby, and Steven Salop. 1992. ‘An Economic Analysis of Copyright Collectives,’ 78
Bouchard, Mario. 2010. ‘An Essay on Monetizing Copyright Over the Internet’, 11:6 Internet and E-Commerce
Law in Canada 45–52.
Cohen Jehoram, Herman. 2001. ‘The Future of Copyright Collecting Societies,’ 23 European Intellectual
Property Review 134–9.
Geiger, Christophe, Daniel Gervais, and Martin Senftleben. 2015. ‘The Three-Step Test,’ in Daniel Gervais,
ed., International Intellectual Property: A Handbook of Contemporary Research. Cheltenham, UK and

Gervais, Daniel. 2004. ‘The Price of Social Norms: Towards a Liability Regime for File-Sharing,’ 12 Intellectual
Property Journal 39–74.
Gervais, Daniel. 2011. ‘The Landscape of Collective Management,’ 34 Columbia Journal of Law and the Arts
591–618.
Gervais, Daniel. 2012. ‘Individual and Collective Management of Rights Online,’ in Johan Axhamn, ed.,
Copyright in a Borderless Online Environment. Stockholm: Norstedts Juridik.
Gervais, Daniel. 2013. ‘The Internet Taxi: Collective Management of Copyright and the Making Available
Right, after the Pentalogy,’ in Michael Geist, ed., The Copyright Pentalogy: How the Supreme Court of Canada
Shook the Foundations of Canadian Copyright Law. Ottawa: University of Ottawa Press.
Gervais, Daniel. 2015. ‘Authors, Online,’ 38 Columbia Journal of Law and the Arts 385–96.
Gervais, Daniel. 2018. ‘The Cultural Role(s) of Collective Management Organizations,’ 6 European Intellectual
Property Review 349–56.
Ginsburg, Jane C. 2014. ‘Fair Use for Free, or Permitted-but-Paid?,’ 29 Berkeley Technology Law Journal
1383–446.
Handke, Christian, and Ruth Towse. 2007. ‘Economics of Copyright Collecting Societies,’ 38 International
Review of Intellectual Property & Competition Law 937–57.
Hollander, Abraham. 1984. ‘Market Structure and Performance in Intellectual Property,’ 2 International Journal
of Industrial Organization 199–216.
Korman, Bernard, and Fred I. Koenigsberg. 1986. ‘Performing Rights in Music and Performing Rights
Societies,’ 33 Journal of the Copyright Society of the U.S.A. 332–67.
Kretschmer, Martin. 2002. ‘The Failure of Property Rules in Collective Administration: Rethinking Copyright
Societies as Regulatory Instruments,’ 24 European Intellectual Property Review 126–37.
Landes, William M., and Richard A. Posner. 2003. The Economic Structure of Intellectual Property. Cambridge,
MA: Harvard University Press.
Liebowitz, Stan. 2005. ‘MP3s and Copyright Collectives: A Cure Worse than the Disease?,’ in Lisa N. Takeyama,
Wendy J. Gordon, and Ruth Towse, eds., Developments in the Economics of Copyright: Research and Analysis.
Lunney, Glynn S. 2001. ‘The Death of Copyright: Digital Technology, Private Copying, and the Digital
Millennium Copyright Act,’ 87 Virginia Law Review 813–920.
Mackaay, Ejan. 1990. ‘Economic Incentives in Markets for Information and Innovation,’ 13 Harvard Journal
of Law and Public Policy 867–909.
Marshall, Rick. 2013. ‘The Quest for “Parity”: An Examination of the Internet Radio Fairness Act,’ 60 Journal
of the Copyright Society of the U.S.A. 445–74.
Massarsky, Barry M. 2013. ‘The Operating Dynamics behind ASCAP, BMI and SESAC, the U.S. Performing
Rights Societies,’ accessed March 27, 2019 at http://old.cni.org/docs/ima.ip-workshop/Massarsky.html.
Maurushat, Alana, and Daniel Gervais. 2003. ‘Fragmented Copyright, Fragmented Management: Proposals to
Defrag Copyright Management,’ 2 Canadian Journal of Law & Technology 15–34.
Melichar, Ferdinand. 1991. ‘Deductions Made by Collecting Societies for Social and Cultural Purposes in the
Light of International Copyright Law,’ 22 International Review of Industrial Property and Copyright Law
47–60.
Parisi, Francesco, and Ben Depoorter. 2003. ‘The Market for Intellectual Property: The Case of Complementary
Oligopoly,’ in Wendy J. Gordon and Richard Watt, eds., The Economics of Copyright: Developments in
Research and Analysis. Cheltenham, UK and Northampton, MA, USA: Edward Elgar Publishing.
Patry, William. 1995. ‘COPYRIGHT: American Geophysical Union v. Texaco, Inc.: Copyright and Corporate
Photocopying,’ 61 Brooklyn Law Review 429–51.
Richardson, Megan, and Sam Ricketson. 2017. ‘The Cultural Role of Copyright Collectives,’ in M. Richardson,
ed., Research Handbook on Intellectual Property in Media and Entertainment. Cheltenham, UK and
Rochelandet, Fabrice. 2003. ‘Are Copyright Collecting Societies Efficient Organizations? An Evaluation
of Collective Administration of Copyright in Europe,’ in Wendy J. Gordon and Richard Watt, eds., The
Economics of Copyright: Developments in Research and Analysis. Cheltenham, UK and Northampton, MA,
USA: Edward Elgar Publishing.
Smith, Douglas A. 1986. ‘Collective Administration of Copyright: An Economic Analysis,’ in John Palmer, ed.,
Research in Law and Economics. Greenwich, CN: JAI Press.
Snow, Arthur, and Richard Watt. 2005. ‘Risk Sharing and the Distribution of Copyright Collective Income,’
in Lisa N. Takeyama, Wendy J. Gordon, and Ruth Towse, eds., Developments in the Economics of Copyright:
Research and Analysis. Cheltenham, UK and Northampton, MA, USA: Edward Elgar Publishing.
Thorpe, Jeremy. 1998. ‘Regulating the Collective Exploitation of Copyright,’ 16 Prometheus: Critical Studies
in Innovation 317–29.
Tuma, Pavel. 2006. ‘Pitfalls and Challenges of the EC Directive on the Collective Management of Copyright
and Related Rights,’ 28 European Intellectual Property Review 220–29.

U.S. Copyright Office. 2015. Copyright and the Music Marketplace: A Report of the Register of Copyrights,
accessed March 27, 2019 at https://www.copyright.gov/policy/musiclicensingstudy/copyright-and-the-music-
marketplace.pdf.
Yu, Peter K. 2003. ‘The Harmonization Game: What Basketball Can Teach about Intellectual Property and
International Trade,’ 26 Fordham International Law Journal 218.
Case List
Am. Geophysical Union v. Texaco Inc., 60 F.3d 913 (2d Cir. 1994).
Phonographic Performances (NZ) Ltd v. Radioworks Limited, NZCOP 1 (2010).
17 U.S.C. § 101, §112(e)(5), § 115(b).

37 C.F.R. § 260 (2003).
Copyright Act (Act No. 63/1968) (Austl.).
Copyright Act (Act No. 7/2000) (St. Lucia).
Copyright Act, R.S.C., c. C-42, s. 2. (1985) (Can.).
Copyright Board of Canada. 2009. ‘Collective Administration of Performing Rights and of Communication
Rights,’ Statement of Royalties to Be Collected by SOCAN, NRCC and CSI in Respect of Multi-Channel
Subscription Satellite Radio Services, accessed March 27, 2019 at http://cb-cda.gc.ca/decisions/2009/20090408-
m-b.pdf (Can.).
Rights,’ Statements of Royalties to Be Collected by SOCAN and CMRRA/SODRAC INC. for the
Communication to the Public by Telecommunication or the Reproduction, in Canada, of Musical Works,
accessed March 27, 2019 at http://cb-cda.gc.ca/decisions/2012/socan-csi-reasons.pdf (Can.).
Rights,’ Statement of Royalties to Be Collected by Access Copyright for the Reprographic Reproduction, In
Canada, of Works in its Repertoire: Decision of the Board, accessed March 27, 2019 at http://cb-cda.gc.ca/
decisions/2015/DEC-2015-03-22.pdf (Can.).
Copyright Regulations (Act No. 9/2004) (Kenya).
Copyright Regulations (Act No. 239/2016) (Malta).
Directive 2014/26/EU of the European Parliament and of the Council of 26 February 2014 on collective
management of copyright and related rights and multi-territorial licensing of rights in musical works for
online use in the internal market, 84/72, 2014 O.J. (Eur.).
WIPO. 2003. Guide to the Copyright and Related Rights Treaties Administered by WIPO and Glossary of
Copyright and Related Rights Terms, accessed March 27, 2019 at http://www.wipo.int/edocs/pubdocs/en/
copyright/891/wipo_pub_891.pdf.

19. ‘The common law’ in the law and economics of
Shyamkrishna Balganesh * 21
Contents
I. Introduction
II. Efforts to Understand the Common Law as a Unified Whole
III. The Common Law as a Method of Lawmaking
IV. The Common Law as Judge-made Law
V. The Common Law as Substantive Law
A. Freestanding Common Law Intellectual Property
B. Interstitial Common Law Intellectual Property
VI. The Common Law as State Law
VII. Conclusion
References
I. INTRODUCTION
As applied to the context of intellectual property thinking today, ‘the common law’ refers
to a variety of different analytical ideas, each of which contributes to a particular norma-
tive vision for the institution. These myriad understandings of the term in turn draw from
different structural, institutional, and jurisprudential ideas that the common law has come
to be associated with over the last several centuries. Accordingly, this chapter attempts to
unbundle these different usages of the term in order to show what they each bring to the
discussion of intellectual property.
As a general matter, discussions of the common law within intellectual property can
be classified into four analytically distinct strands. This distinction certainly does not
suggest that these strands do not overlap or indeed that in practice the use of the term
does not implicate more than one. Instead, each strand brings a different dimension of
economic thought and analysis to the discussion, which in turn generates its own set of
normative implications when applied. In the first strand, the common law is seen as a
particular method of lawmaking, associated primarily with the ideals of pragmatism and
incrementalism, when employed as a method of reasoning. In the second, the common
law is treated as representing a particular locus of institutional authority for lawmaking:
courts. Here, it is used as a synonym for ‘judge-made law.’ While the first and second
strands ordinarily go together, it is important to recognize that they are nevertheless
* Professor of Law, University of Pennsylvania Law School. Many thanks to Lydia Franzek,
University of Pennsylvania Law School Class of 2020 for research assistance.
508

‘The common law’ in the law and economics of IP 509
distinct and that instances can and do arise when the overlap ceases. This divergence is
especially true in areas where an agency such as the United States Patent and Trademark
Office (USPTO) or the Copyright Office starts assuming a lawmaking role as well. The
third strand treats the common law of intellectual property as referring to bodies/areas of
substantive law that are usually produced through the mechanisms reflected in the first two
strands. In other words, this strand is associated with the product/output of the processes
and ideals reflected in the prior two. As discussed below, it in turn has two variants:
freestanding common and interstitial common law. The fourth strand is best described as
jurisdictional and associates the common law in intellectual property with the ability and
power of states to create intellectual property rules, either with or without the processes
and ideals reflected in the other strands.
Each of these strands—legal reasoning, institutional authority, substantive, and
jurisdictional—plays an important role in framing and directing discussions of the
common law in intellectual property and accordingly relates to different facets of the
economic analysis of law and lawmaking as applied to the common law. While each of
these legal strands finds a correlate in the law and economics literature, it is important
to note (as described more fully below) that the latter approaches the economic analysis
of the common law in overtly normative—rather than purely positive—terms. Economic
analyses of the common law thus begin with an effort to test a hypothesis about the con-
nection between the common law and efficiency/utility-maximization and to that end go
on to adopt varying conceptions of the common law and institutional dynamics therein.
This mode of analysis is worth bearing in mind in the correlations made below, since the
correspondence between the positive analysis seen in the general legal literature and the
normative analysis seen in the law and economics literature is attenuated as a result.
Yet, rarely ever do references to ‘the common law’ in intellectual property elucidate
on the precise variant being employed, in the process giving usages of the term a degree
of analytical imprecision that generates varying levels of suspicion among courts and
scholars. The discussion below is divided into six sections. Section II introduces some of
the seminal literature on ‘the common law,’ both in the traditional legal literature and in
the economic analysis of law. Sections III, IV, V, and VI collectively unbundle the term as
applied in the intellectual property setting, with each elaborating on a particular strand
of usage reflected in the idea, drawing on examples from within intellectual property and
relating them to established theoretical insights from law and economics as pertains to
judge-made law. Section VII then ends by discussing some fruitful directions that future
research in the area might take.
II. EFFORTS TO UNDERSTAND THE COMMON LAW AS A

UNIFIED WHOLE
Efforts to understand the common law as a unitary whole, that is, as ‘the common law,’
can be traced back to the 17th and 18th centuries, during which time Coke’s Institutes
and Blackstone’s Commentaries were produced. Due to the nature of lawmaking that
existed in England at the time, both works contain fairly little discussion of the compara-
tive costs and benefits of the common law and focus less on the common law from a
phenomenological perspective, instead treating the common law as but a stand-in for ‘law’

more generally. It therefore is not until the late 19th century, and owing to the work of
American legal scholars who were engaged in the task of critically evaluating the recep-
tion of English law in the United States, that we begin to see attempts to understand and
evaluate the common law as a distinct body of law and mechanism of lawmaking.
The Common Law by Oliver Wendell Holmes, Jr. (1881) is traditionally taken to be the
starting point of the modern American engagement with the common law. Yet, Holmes is
largely content with an analytical and historical exegesis of different substantive common
law rules, in some ways falling back on the models put forth by Coke and Blackstone.
Holmes also restricts himself to the main areas of the substantive common law and in
the process finds little time for any discussion of intellectual property. The first truly
analytical engagement with the common law, one that transcends its role as a repository of
substantive rules and looks to it as a method of lawmaking, is to be seen a few years later
in the work of Benjamin Cardozo. In The Nature of the Judicial Process and The Growth
of the Law, Cardozo (1921, 1924) develops an account of common law reasoning and the
role of the judge therein best described as constructive and pragmatic. In this account,
we also see the use of ideas from sociology, philosophy, and economics in the study of the
common law for the first time.
Cardozo’s work marked the onset of Legal Realism, an approach to legal thinking
that questioned the autonomy of legal discourse. Following Cardozo, several prominent
Legal Realists began engaging the common law process in an effort to elucidate their
core jurisprudential theses. Prominent among these efforts were Roscoe Pound’s (1906)
The Spirit of the Common Law, which was part historical and part normative, and Karl
Llewellyn’s (1960) The Common Law Tradition: Deciding Appeals. Llewellyn’s account is
particularly noteworthy in that it attempts to discern a rational coherence in the working
of the common law, using what he famously describes as the idea of ‘situation sense’ to
determine the applicability of prior precedent to any given dispute.
When the law and economics movement began to gain momentum in the 1970s, the
common law became an obvious area for it to test its core hypothesis. Richard Posner’s
(1973) Economic Analysis of Law famously argues that ‘[t]he common law method is to
allocate responsibilities between people engaged in interacting activities in such a way as
to maximize the joint value, or, what amounts to the same thing, minimize the joint cost
of the activities.’ This approach would become the core of the economic analysis of the
common law, with scholars within the tradition then attempting to discern why, how, and
where common law rules work to produce efficient outcomes.
A more recent effort to offer a unified account of the common law is seen in Mel
Eisenberg’s (1988) The Nature of the Common Law. While steeped in the Realist tradi-
tion, Eisenberg consciously attempts to transcend the economic analysis by offering an
account of ‘the institutional principles that govern the way in which the common law is
established.’ The work focuses on the interaction between ‘doctrinal’ and ‘social’ proposi-
tions, that in his account figure in all forms of common law adjudication.
Among the more recent non-economic efforts to understand the common law as a
whole, Alan Brudner’s (1995) classic, The Unity of the Common Law, deserves special
mention. In it, Brudner develops a theory of the common law that combines its formal
and instrumental aspects, as well as its roots in private law and public law. Relying on
Hegel’s concept of the geist, Brudner seeks to explain and justify the common law as a
coherent set of rules and practices that deserves independent legal scrutiny.

The law and economics literature on the common law has, for the most part, revolved
around trying to understand the systemic efficiencies of judge-made law as an enterprise
of law production. Well before the advent of law and economics, Hayek hypothesized that
the common law was more efficient than statutory law, and, in the years since, scholars
have attempted to either validate or refute this hypothesis. Prominent in this genre are
seminal works by Richard Posner (1992), Paul Rubin (1977), George Priest (1977),
Bob Cooter (1979), Lewis Kornhauser (1989), and Paul Mahoney (2001). Most of this
literature examines the efficiency-related hypothesis through individual legal doctrines,
undertakes a demand- or supply-side analysis of the process through which the common
law generates rules, looks at efficiencies relating to certain artifacts and processes within
the common law (e.g., stare decisis or multi-panel judges), or relates the idea of efficiency
to the common law’s ability to change and adapt to new circumstances in a timely manner.
Sadly enough, though, none of these unified studies of the common law, whether in the
traditional legal literature or in law and economics, make much of an effort to study intel-
lectual property through the common law, or indeed to examine the connection between
the two. While Posner’s (1973) work does contain a discussion of patent and copyright,
its primary goal is to study these subjects through the tools of economic analysis rather
than through the common law as such. Yet, this critique is hardly to suggest that efforts
to study aspects of intellectual property through the common law are altogether absent;
they just remain few and far between.
The Supreme Court’s 1918 decision in International News Service v. Associated Press,
248 U.S. 215 (1918) provided some momentum for the idea that the common law might
have some lessons of importance for intellectual property. A good amount of scholarship
attempting to make sense of the Court’s decision and its theory of ‘misappropriation’
therefore chose to engage with the interface between the common law and intellectual
property. Unfortunately, though, much of this scholarship was restricted to the narrow
domain of the misappropriation doctrine, and never quite examined what the common
law as a freestanding whole might bring to bear on intellectual property discussions.
Most recently, some scholars have begun examining the general connection between the
common law and intellectual property more directly, in the process engaging with some
of the themes on which the next few sections elaborate.
III. THE COMMON LAW AS A METHOD OF LAWMAKING
The first sense in which the phrase ‘common law’ is routinely used is largely analytical.
It refers to a particular method of contextualized lawmaking that English common law
courts were historically seen as adopting. This method was in turn characterized by being
both pragmatic and incremental. Common law courts came to generate legal rules and
principles from within the context of individual cases, using the particular dispute at
hand before them as the basis on which to either apply a rule from a prior case directly or
to adapt such a prior rule to fit the particulars of the dispute involved. The process thus
placed great emphasis on the use of precedent and stare decisis. Deductive and inductive
methods of reasoning took second place to analogical reasoning, and with it the processes
of abstraction and extension. All the same, the need to decide the individual dispute at
hand had an important disciplining effect on courts in individual cases. It ensured that in

the development of their rules/principles, courts were acutely aware of (i) how any par-
ticular rule/principle would play out in practice and (ii) the effects that it would generate
for the parties involved and other similarly situated parties. Indeed, it is for this reason
that Holmes famously observed that ‘[i]t is the merit of the common law that it decides
the case first and determines the principle afterwards’ (Holmes, 1870, p. 1).
The common law method of lawmaking has often been characterized as ‘pragmatic,’
a term that is used to capture three essential attributes of the process (Balganesh, 2010).
The first is its overt instrumentalism, or its emphasis on the consequences—immediate
and long-term—of the decision at hand. The second is its anti-foundationalism, or its
unwillingness to emphasize any single foundational value, such as autonomy or efficiency,
during the rule-development and adaptation processes. Instead, the process recognizes the
existence of multiple normative goals being served by the law and chooses to contextually
balance them out. The third attribute is its contextualization, wherein the rule/law is
developed by reference to its social situatedness rather than in the abstract. This attribute
relates back to the disciplining effect of case-based lawmaking referred to earlier. In
addition to being pragmatic, the common law method of reasoning and lawmaking is
also overtly incrementalist. Incrementalism refers to the gradual or accretive nature in
which the law (or an institution) develops over time. The process involves testing a rule in
individual cases over time and adapting it situationally with minimal modification, so as
to balance consistency and predictability with the need for flexibility. These four attributes
together result in the common law method of legal development being best characterized
as pragmatically incrementalist.
Within the law and economics literature, this strand of common law thinking was
initially seen in the work of Richard Posner (1973), who argued that the common law was
efficient through an examination of individual legal doctrines. These initial conclusions
produced a good amount of discomfort among economists, who were more predisposed
towards examining the efficiency or otherwise of a process, rather than the outcome of
that process. In this more process-oriented economic examination, the focus thus shifted
to two fruitful sources of analysis: the extent to which judges were utility-maximizers and
the extent to which the process of litigation (which generated the rules of the common
laws) was itself efficient. It is the latter of these two that found a close analog to non-
economic claims about the common law process (Rubin, 1999).
Paul Rubin (1980), who pioneered this form of economic analysis of the common law,
characterizes it as an ‘evolutionary’ approach. The core claim of these studies is that the
common law veers towards the selection of an efficient rule (over inefficient ones) because
of the adversarial process and the incentives of the litigating parties. Assuming the parties’
relative stakes in a dispute are symmetrical, a case gets settled when the expected value
to the plaintiff in the case is less than the expected cost to the defendant. When laws are
inefficient, the stakes become asymmetrical and generate a loss to one side that is greater
than the gain to the other, owing to its effect on future cases. This in turn makes litigation
more likely when rules are inefficient, making them likely to be litigated and modified/
overturned through the judicial process. Variants of this argument developed in the
literature in due course.
Evolutionary theorizing in the law and economics of the common law may not be the
dominant strand within that arena today, yet its principal contribution lay in emphasizing
‘the litigation decision as driving legal evolution’ (Rubin, 1999). As Rubin puts it, these

theories ‘were the first to focus on the motives of litigants, realizing that judges can decide
only those cases that come before them.’ Relating these theories back to the non-economic
literature, evolutionary theorizing thus claimed that the pragmatism and incrementalism
of the common law method of theorizing and rule development were as much an artifact
of the litigation process and the parties’ incentives therein as they were the product of
some innate or unspecified wisdom underlying the common law. In this important respect,
the non-economic and economic versions of this form of common law thus complement
each other fairly well.
Returning to intellectual property, one argument that deploys the common law here
is that the institution (i.e., various forms of intellectual property) would stand to benefit
from this pragmatic incremental method of rule development. The pragmatic incremen-
talism inherent in the process is seen as virtuous for the field, in so far as it is directed at
maximizing efficiency through minimizing inefficient rules (Balganesh, 2010).
To begin, the technologically dependent nature of intellectual property makes it ideally
suited to incremental rule development. Patent, copyright, trade secrets, and, to a lesser
extent, other forms of intellectual property attempt to regulate the production, con-
sumption, and dissemination of new technology in society. All the same, predicting the
precise direction and speed of such technological change is notoriously difficult given its
intrinsically stochastic nature. A one-size-fits-all approach that attempts to make radical
changes by predicting the future trajectory of technological development runs the risk of
proving to be either under- or over-determinative. Instead, an incremental approach that
takes ‘one case at a time’ and moves gradually based on empirical evidence drawn from
the circumstances of an individual case introduces an element of regulatory caution and
modesty in the face of pervasive technological uncertainty.
In addition, the normative value, pluralism, associated with intellectual property is best
served through the common law’s commitment to practical reasoning. While intellectual
property is today defended in principally utilitarian terms, the fact remains that several
other non-instrumental goals are just as important to its functioning, be they autonomy,
fairness, or distributive justice. The fundamental anti-foundationalism of the common
law method allows these plural normative goals to be balanced situationally through
the generation of a reflective equilibrium deeply embedded within the common law’s
commitment to practical reasoning. Indeed, many have characterized the common law
as epitomizing the ideal of practical reasoning, and intellectual property’s normative
uncertainty would stand to benefit from it.
Turning now to the law and economics literature on this conception of the common law,
to the extent that the evolutionary models’ assumptions about asymmetrical incentives
carry over to rules that are seen as systemically undesirable owing to their long-term effects
that can be understood in both economic and non-economic terms (not just owing to their
inefficiencies), we begin to see the resonance behind applying the common law ‘method’
to intellectual property rule development. We therefore begin to see a good degree of cor-
respondence between evolutionary theorization in the law and economics literature and
claims about the superiority of the common law method within traditional legal analyses.
We see this strand of common law thought at work in intellectual property most
prominently in copyright’s fair use doctrine, 17 U.S.C. § 107. As a creation of the federal
courts in the 19th century, Congress came to codify the doctrine in 1976. Yet, in so doing,
Congress continued to insist that its codification was hardly comprehensive, noting that:

The bill endorses the purpose and general scope of the judicial doctrine of fair use, but there is no
disposition to freeze the doctrine in the statute, especially during a period of rapid technological
change. Beyond a very broad statutory explanation of what fair use is and some of the criteria
applicable to it, the courts must be free to adapt the doctrine to particular situations on a case-
by-case basis. Section 107 is intended to restate the present judicial doctrine of fair use, not to
change, narrow, or enlarge it in any way. (H.R. 94-1476, 94th Cong. (2d Sess. 1976))
In essence, Congress was endorsing the common law method of lawmaking through
which the fair use doctrine had evolved and was insisting that the same method continue
to be applied in adapting the doctrine to new and evolving situations. As is apparent from
the record, ‘technological change’ was an obvious reason for this endorsement, given the
common law’s cautionary and gradual approach. The emphasis on ‘case-by-case’ further
highlights Congress’s insistence on the situated and contextual nature through which the
law develops. Additionally, as has been revealed over time, one of the added virtues of
utilizing the common law method for fair use proved to be the reality that fair use itself
serves a variety of important purposes relating to free speech, downstream economic
incentives, distributive justice, and the like. The common law’s case-by-case approach has
allowed these myriad goals to be balanced and simultaneously affirmed.
IV. THE COMMON LAW AS JUDGE-MADE LAW
Within intellectual property and beyond, ‘the common law’ is also routinely used to signify
an area (or body) of law or legal rules that is principally or exclusively judge-made. In
this sense, it refers primarily to rules developed by courts—either federal or statutory—
through the process of accretive rule development. The process is not just an expansive
interpretation of statutory directives, but instead involves the lawmaking process itself
coming to be delegated to courts. Of course, in an overwhelming majority of cases, the
common law method previously described accompanies the process of judicial lawmak-
ing. As an analytical matter though, the two need not always go hand in hand.
In developing the law on their own, or with limited legislative guidance, courts are
routinely driven by the normative ideals commonly believed to be associated with any
given area at hand. Learned Hand’s famous ‘Hand formula’ for negligence liability is a
good example as it developed almost entirely in the belief that negligence law was directed
at efficiently deterring dangerous activities (United States v. Carroll Towing Co., 159 F.2d
169 (2d Cir. 1947)). These ideals are in turn instantiated in the application and develop-
ment of a rule that is applied to the facts of the case before the court. Consequently,
judge-made law invariably develops incrementally, even though it need not adhere to
the ideals of pragmatism. Many bodies of American law remain principally (or entirely)
judge-made to this day: contract, tort, and property law in the state context and antitrust
and admiralty law in the federal.
Ordinarily, judge-made law in any given area is inherently susceptible to legislative
override. In other words, if the legislative branch is unhappy with any particular rule, or
indeed body of rules developed by a court, it can choose to override it by simply enacting
legislation to that effect. In this regard, judge-made law can be seen as operating with
the often tacit acquiescence of the legislature. In some areas, of course, the legislature
expressly delegates its lawmaking function to courts within a statute, thereby making an

overt commitment not to override courts, barring some major exceptions. These statutory
delegations are sometimes referred to as ‘common law statutes,’ and scholars continue to
debate the precise interpretive rules that ought to govern them.
Judicial lawmaking of this kind has certain inherent advantages. To the extent that it is
incremental in growth, it allows the law to move in new directions based on social changes
as refracted through the lens of the dispute before a court. In addition, it produces an
extended conversation over a period of time during the development of any rule, thereby
allowing the reform process itself to have multiple avenues for its ideas to be tested.
Perhaps most importantly, though, judicial lawmaking allows for a legal system to fill
certain important gaps in the system of formal rules, thereby ensuring that potentially
harmful (or inefficient) defaults that could result from the absence of legislative interven-
tion are minimized.
In addition to exhibiting obvious process-related benefits, judge-made law is also often-
times defended in terms of institutional competence and expertise. To the extent that the
judicial office is taken to embody certain virtues that inure to the benefit of the lawmaking
process in any given area, judge-made law in that area is seen as superior to legislative (or
indeed executive) rulemaking. A defense of the common law along these lines has become
less popular over the years, especially given concerns with anti-majoritarianism and the
undemocratic nature of judging. All the same, we see traces of it in early accounts of
the common law, where the system was defended as much for the virtues of its primary
protagonists (i.e., judges) as for its unique techniques and methods. As a situated member
of society, taken to embody a certain amount of common ‘wisdom’ in turn reflected
in a commitment to practical reasoning, judges, in this account, were seen as uniquely
positioned to develop the law in domains that tended to elide broad generalization. More
recently, this argument has renewed in a call for ‘virtue jurisprudence,’ or an Aristotelian
conception of law, which places just as much focus on the virtues of the lawmaker as it
does on the content and normative commitments of the law itself (Solum, 2004).
Judicial lawmaking, however, is hardly without its own flaws and critics. While some
of this criticism is institutional/political, emanating from a deep suspicion of unelected
judges and the usurpation of power by the judiciary, the more analytical side of the
critique originates in the idea that the common law’s focus on the individual dispute at
hand distorts the construction and framing of the law. Early versions of this argument are
seen in the writings of Jerome Frank (1930) and Karl Llewellyn (2011), who both claimed
that the particularized nature of the controversy before the court provided judges with the
real reasons for their decisions and rules. While Llewellyn and Frank saw this as largely
unproblematic, more recently some scholars have argued that this focus on the individual
characteristics of a case produces a series of cognitive biases among judges, which diverts
their attention away from understanding the true medium- and long-term consequences
of any legal rule that they are involved in framing. Fred Schauer (1989) thus argues that
this form of lawmaking invariably relies on the availability heuristic, an anchoring bias,
and issue framing, both of which result in the process being less than fully rational in
outcome.
To the extent that this account of the common law relies on the attributes of judges for
its validation, its closest analog in the law and economics literature was work that treated
judges as rational utility-maximizers and, in that vein, associated the common law with
the overall goal of utility-maximization, seen through the agency of judges. Richard

Posner’s (1973) early work epitomized this thinking, wherein he argued that judges were
insulated from personal factors and interest group politics, enabling them to focus on
efficiency maximization. Others have since followed suit and modified the initial model
by introducing elements of practical relevance to the role of judging, such as the judge’s
incentives relating to promotions and their employment conditions. Despite having
some traction within the world of law and economics, however, this strand of common
law thinking has been largely relegated to the margins, mostly because of its unverified
assumptions about judicial incentives, preferences, and behavior, all of which are hard to
theorize about and generalize from in the aggregate.
We nevertheless see important aspects of these debates at play in the context of judicial
lawmaking within intellectual property. On the one hand, the judicial crafting of a legal
rule from the context of an individual dispute may indeed be informed by a keener sense
of the direction in which future change is likely to take (or the sheer uncertainty of that
process instead). Indeed, it may well result in judges, who are presented with intellectual
property disputes with a degree of regularity, coming to develop an amount of specialized
expertise in an area, which renders them more capable of crafting an ideal rule, unlike
a legislative body that is by definition more generalized in its agenda. Some argue that
encouraging judicial expertise was precisely the thinking behind the creation of the Federal
Circuit within the patent law arena, though the evidence is obviously mixed on whether
patent law (and the legal system more generally) benefited from this specialization.
Similarly, copyright’s fair use doctrine was (and remains) principally the creation of
courts. Developed by Justice Story in Folsom v. Marsh, 9 F. Cas. 342 (C.C.D. Mass. 1841)
from within the context of a particular dispute, the boundaries of the doctrine continue
to be policed by courts and adapted to newer contexts. It is in this vein that the Supreme
Court came to adopt the ‘transformative use’ variant of the Court in Campbell v. Acuff-
Rose Music, Inc., 510 U.S. 569 (1994), a move that some argue breathed new life into the
fair use doctrine and allowed for its expansion to the digital context where the reuse and
modification of existing works occurs on a rather regular basis.
As an illustration of the converse phenomenon, even within the working of fair use,
consider more recent efforts to apply the doctrine to appropriation art, that is, art that is
created by modifying protected works in new forms—sometimes insubstantially. In Cariou
v. Prince, 714 F.3d 694 (2d Cir. 2013), the Second Circuit concluded that the defendant’s
artwork, which entailed taking the plaintiff’s ethnographic photographs and airbrushing
a few elements onto the images there, qualified as fair use because the defendant had
added new meaning to the work by converting the photograph into a work of art, even if
only with minimal modification. The court was unquestionably driven by the prominence
of the defendant’s art—which had received a good amount of notoriety in the press and
elsewhere—and, in so doing, paid surprisingly little attention to how/when the fair use
doctrine should be applied to other instances of appropriation art. As a consequence, the
extent to which fair use applies to appropriation art continues to remain unclear. None
of this discussion is to suggest that the costs of particularity in the judge-made approach
to fair use outweigh the benefits of nimbleness and potential judicial expertise. It merely
points to the advantages and disadvantages of the common law—understood as judicial
lawmaking—manifesting themselves in the intellectual property setting.

V. THE COMMON LAW AS SUBSTANTIVE LAW
A third and frequently used sense in which the term is employed identifies it with various
bodies of substantive legal rules and principles that are largely (even if not exclusively)
developed incrementally, and usually by courts. Whereas the prior two conceptions
focused on the process—reasoning and institutional basis—for the development of these
rules, this conception focuses on the product of that method and institutional authority.
Separating the two, that is, the process and the product, is nonetheless analytically critical
since the substantive law comes to assume an independent life of its own, once brought
into existence. In other words, once created, courts and scholars rarely ever look back into
the method through which the common law developed or the institutional basis behind
its creation.
Even when understood as substantive law, the common law comes in two different
forms (Balganesh, 2013). The first is freestanding substantive common law, which
refers to bodies of law that emerge spontaneously, and usually through courts’ reliance
on custom and norms within a particular domain of activity, to craft new rules when
presented with an individual dispute. In this form, the common law takes no guidance
(or delegation of authority) from the legislature and cares very little about potentially
conflicting with statutes or other forms of government regulation. As should be obvious,
most of the foundational areas of the law, such as property law, contract law, and tort law,
developed in this fashion, even though there exist numerous statutory adaptations and
incorporations of these original common law rules today (e.g., the Uniform Commercial
Code (U.C.C.)).
What is important to appreciate with freestanding common law is that the freestanding
nature is more than just an issue of temporal priority. In other words, it is not just that
the law came to be developed by courts before the legislature stepped in. Rather, the fact
that it was developed by courts, who assumed principal responsibility for the develop-
ment, adaptation, and maintenance of the law on its own, resulted in it earning a degree
of deference from the other branches of government such that they remain unwilling to
expressly override the common law unless it produces consequences or outcomes that
are seen as palpably egregious or inefficient. Even when the law in one of these domains
comes to be codified, the common law origins of the rule are seen as a license to courts
to continue to develop and adapt the law as circumstances demand.
In contrast to freestanding substantive common law is interstitial common law.
Interstitial common law refers to substantive law—also judge-made and usually incremen-
tally developed—produced within a domain that is primarily legislative in orientation and
focus. Statutes in a variety of different substantive areas consciously refrain from provid-
ing actors with elaborate guidance in the nature of bright line rules. Instead, they are
framed principally in terms of legal standards, which are then treated as active delegations
of law- and policy-making to courts to be done on an ex post basis. Oftentimes, a review
of the legislative history behind the enactment of a particular statute reveals Congress’s
express desire to delegate such rulemaking power to courts, either for institutional reasons
(of the kind discussed previously) or because courts had previously assumed primary
responsibility for the rules in that area and were seen as doing a good job in developing
them.
Interstitial common law, however, is oftentimes hard to identify as such. Since it exists

within a predominantly statutory domain, it is common practice for courts to couch

their lawmaking and rule development in the language of interpretation and application
of existing law, rather than in terms of making new law. The politics of this analytical
move and the problems it produces for the legal system have been the subject of extensive
debate by jurists for decades now. The Legal Realists famously claimed that this approach,
that is, masking rulemaking in the language of rule interpretation and application, was
deceptive and served to undermine the legitimacy of legal rules. In addition, they took it
to suggest that legal rules had no constraining (or cabining) effect on judicial reasoning,
which in turn fed into the basic legal indeterminacy thesis with which the Realists were
associated. Much of this extreme version of the argument has since been rejected. Legal
philosophers critical of positivism as a jurisprudential theory came to argue that the is/
ought divide in the law was artificial at best and that this did not imply that law and rules
were necessarily indeterminate or normatively vacuous in their content. While the precise
contours of these debates need not detain us here, they nonetheless highlight the reality
that interstitial common law is much more than just an outlier in the legal system and has
served as the site for several well-known disagreements about the very role of law, legal
reasoning, and rules.
Discussions of substantive common law within the law and economics literature
emerged during the first generation of theorizing about the common law therein. It
arose principally in the claims of those like Posner (1973), who argues that specific
doctrines of the common law—the product of the common law process—are efficiency-
maximizing and thus directed at promoting utilitarian and wealth-maximizing ideals.
In the years since, others have followed suit. As Rubin (1999) points out, this method
of legal analysis, while standard today in the law and economics literature, remains a
major departure from economic theorizing more generally, which focuses on a process
(e.g., competition or regulation) to measure its efficiency rather than a product. Yet, this
approach to theorizing in the law and economics literature remains fairly common and
has been extended to discussions about intellectual property as well. It is worth pointing
out that the law and economics literature does not, as best as one can tell, deal with the
distinction between free standing and interstitial common law, which one sees in the
traditional legal literature.
Both forms of substantive common law remain well-known and indeed, rather influen-
tial, in the world of intellectual property.
A. Freestanding Common Law Intellectual Property
Within intellectual property, freestanding common law is perhaps the easiest to identify.
Several well-known substantive areas of intellectual property are entirely the product
of judicial lawmaking, developed incrementally from within the context of individual
disputes. The law of trade secrets, publicity rights, common law copyright, hot news
misappropriation, and idea protection remain the foremost examples of such law. While
each of these areas today is conceived of as complex and specialized, in their origins
they came to be developed through courts’ reliance on certain basic rules and ideas from
other more foundational domains of the common law, such as tort and property law.
Consequently, their development showed strong signs of a phenomenon that currently
receives fairly little scholarly attention: inter-doctrinalism. Inter-doctrinalism refers

to the use of concepts, devices, and mechanisms in one area of law to (i) understand
disputes in another area, and in the process (ii) develop new rules for the latter area
through the incorporation of these devices therein. Inter-doctrinalism remained an
important hallmark of the early common law, where today’s rigid divisions between
different subject areas did not exist and courts felt at liberty to seek guidance from allied
or adjacent legal fields of inquiry. A good example is the rule of ‘foreseeability,’ which
contract law came to incorporate as a limiting principle for liquidated damages based on
its utility and role within tort law.
Trade secret law is in many ways the most well-developed and robust area of
freestanding common law intellectual property today. Situated at the interface of
contract law, property law, and tort law, its rules embody elements of all three areas,
which contemporary scholars consider a demerit for the most part. Indeed, some
even suggest that the inter-doctrinal nature of the area, and its obvious overlap with
other more foundational ones, should militate in favor of eliminating trade secret law
as an independent area altogether. This approach misunderstands the basic role of
inter-doctrinalism in the common law and sets a rather ambitious disciplinary bound-
ary for individual intellectual property areas—namely, that they each need to have
independently identifiable normative goals that cannot be served by cognate fields.
Trade secret law also exhibits a feature previously mentioned: the continuing role of
courts in the further development of the law, despite codification by statute. Today,
most states around the country have adopted the Uniform Trade Secrets Act (UTSA),
which seeks to harmonize and codify trade secret doctrine in an effort to stabilize it
and ensure national uniformity. Despite the reality of codification, courts continue to
play a primary role in the further development of the law. State legislatures rarely ever
have problems with this approach and are instead seen as acquiescing in the principally
court-driven approach to the area. Despite codification then, trade secret law remains
a body of common law.
Freestanding common law intellectual property also differs from its interstitial vari-
ant in one other important respect—it is susceptible to complete elimination by courts
over time. Much like legislation, where one institutional authority is responsible for the
creation and management of the law, to the extent that courts assume primary control
over lawmaking in an area, they remain perfectly at liberty to declare (usually over time)
that the area is no longer a viable cause of action or repository of rules that should be
followed. This of course assumes that there has been no legislative intervention in the
area, which would make such a move problematic. But, in the absence of legislation, such
an evisceration poses few problems.
We see such elimination and constriction in the areas of ‘hot news misappropriation’
and common law copyright, both areas of freestanding common law intellectual property.
In being entirely a creation of courts (initially federal, and then state), the true scope and
applicability of the misappropriation doctrine have remained something of a mystery.
Most recently, however, the Second Circuit in effect rendered the misappropriation
doctrine altogether inert by interpreting its domain as being excessively narrow, thereby
confining the scope to what are effectively exceptionally rare instances (Barclays Capital,
Inc. v. Theflyonthewall.com, Inc., 650 F.3d 876 (2d Cir. 2011)). It appears to be but a
matter of time before courts put the doctrine to rest altogether. A similar story may
perhaps be told with regard to common law copyright as well.

The law and economics literature on trade secret law represents a good example of
economic theorizing around freestanding common law. In their well-known work, David
Friedman, William Landes, and Richard Posner (1991) argue, through an examination
of individual trade secret doctrines, that the body of law as a whole represents good
economic sense. They further speculate that it might have ‘surprising efficiency p
roperties’
when viewed as a whole. Of note in their analysis is their observation that their approach
connects trade secret law to ‘broader issues in the positive economic theory of the
common law.’
B. Interstitial Common Law Intellectual Property
Interstitial common law is far more common within intellectual property and therefore
harder to pinpoint and identify. It is easiest to note in areas where a statute in question
is either altogether silent on a question or seen as actively delegating further lawmaking
to courts. Both within intellectual property and outside, interstitial common law is often
confused with judicial gap filling, and it is worth noting the difference between the two.
Whereas gap filling operates as essentially a second-best option to fill an oversight or a
void left by an incomplete statute, interstitial common law is both more enduring and
almost never seen as a lesser (or temporary) alternative to legislative action in the area. In
this sense, gap filling is an interpretive exercise in the expansive sense of the term, whereas
interstitial common law ordinarily does not hide behind the interpretation/lawmaking
divide and embraces its function within the latter category in that divide. There are of
course instances when gap filling morphs into active lawmaking, and there are myriad
statutory techniques and methods that courts use to this end. Yet, this practice is different
from interstitial substantive common law.
In intellectual property, the fair use doctrine is perhaps the best known example
of such interstitial common law. Within the otherwise comprehensive morass of the
Copyright Act, Congress consciously chose to delegate the crafting of limitations and
exceptions to copyright’s exclusive rights—under the rubric of ‘fair use’—to courts.
Further, Congress has done very little in the three decades since the enactment of the
statute to interfere with the courts’ approach in this area. Patent law’s doctrine of non-
obviousness is another good example of this phenomenon (35 U.S.C. § 103). While the
doctrine/requirement finds mention in the Patent Act, both courts and Congress agree
that it should be left to courts to build it through a ‘case-by-case development.’ Despite
strongly held views on the doctrine and its applicability to particular contexts, Congress
has continued to stay out of the process and acquiesced in the courts’ management of
the doctrine.
While both fair use and non-obviousness represent instances when a statute covers
an issue and actively delegates lawmaking in the sub-area to courts, there are other
areas where the statute in question is altogether silent. In these instances, Congress’s
acquiescence is inferred from the fact that, despite its repeated updating of the statute,
the statute continues to make no mention whatsoever of the topic, despite its centrality
to the area. Consider copyright law’s test for infringement. The copyright statute has
historically never once made any mention of how courts (and juries) are to approach the
analysis and comparison of works in order to determine infringement. And, it is hardly
the case that the question is insignificant, since the infringement analysis comes up in an

verwhelming majority of copyright infringement lawsuits. The infringement analysis

o
is, however, entirely a creation of courts. Since the Supreme Court has never weighed
in on the subject, we today have multiple common ‘laws’ on the issue, each limited to
a particular circuit. Owing to the importance of the issue, it can be fairly assumed that
Congress chose to have the statute stay out of the question and acquiesced in the courts’
continuing development of the standards for it.
A similar story may be told about patent law’s ‘doctrine of equivalents,’ patent law’s
analog to copyright’s substantial similarity analysis for non-literal infringement. Again,
the doctrine finds no mention in the patent statute, and yet it has been around for many
decades now, such that Congress’s failure to make mention of it in the statute can be
reasonably taken to represent its tacit approval of the judicial role in creating and main-
taining the doctrine.
Interestingly, scholars of law and economics have begun applying economic analysis to
several interstitial common law intellectual property doctrines in order to examine their
efficiency. The doctrine of equivalents, for instance, has been subjected to significant
scrutiny under the rubric of economic efficiency over the years, as has copyright’s fair
use doctrine.
VI. THE COMMON LAW AS STATE LAW
The last, and perhaps least frequently employed conception of the common law, is simply
a synonym for state law in any given substantive area. The reason for this nominalism
originates in US constitutional rules relating to federalism, specifically, the Erie doctrine
that originated in the Supreme Court’s case of Erie R.R. Co. v. Tompkins, 304 U.S. 64
(1938), where the Court famously declared that ‘there is no federal general common law.’
In so doing, the Court effectively concluded that federal courts could no longer apply rules
that they, as federal courts, spontaneously developed through the common law method,
but instead had to apply common law rules developed by state courts located in the state
jurisdictionally related to the dispute.
The obvious result of Erie was that the term ‘common law’ came to be associated
in certain contexts—especially those related to federalism—with state law rules in the
relevant substantive area. In due course, however, numerous substantive areas of state
law came to be codified by state legislatures. Nonetheless, the Erie doctrine, as a rule
of horizontal federalism, continues to apply in these settings, too. The doctrine cared
little about the distinction between judge-made state law and legislative state law, being
principally a rule that limited federal court (and Congressional) jurisdiction. In some
states, like California, previously common law areas have since come to be codified into
statutes, and courts developing the law there rely on a mix of common law precedent and
statutory text to craft new rules for the disputes that arise before them. In such instances,
it is equally common to simply refer to the state origins of the law as ‘the common law’
despite the codification. In this understanding then, the common law operates as a proxy
for a fundamental precept of federalism.
This conception of the common law is seen most frequently in areas that are predomi-
nantly federal in origin. Intellectual property law is therefore an obvious member of that
set. While patent law, copyright law, and trademark law are primarily federal in origin,

there nonetheless exist numerous other intellectual property regimes that are entirely
state law in origin. Examples include areas discussed previously such as trade secret law,
common law copyright, and misappropriation. It is therefore somewhat common for the
term ‘common law’ in the intellectual property context to refer to this set of areas, whose
most obvious commonality remains their state law origin.
Perhaps the most important legal principle that continues to determine the scope and
extent of state intellectual property regimes is that of federal preemption, or the idea that
federal law in any given area of overlap trumps and displaces state law. The principle
originates in a provision of the US Constitution, which categorically declares that US
law, that is, federal law, is to be the ‘supreme’ law of the land. The law of preemption is a
complex area of federal law, an area, like so many other constitutional questions, fraught
with a lack of clear guidance. Nonetheless, federal preemption is generally understood
to be of three different kinds, each of which has direct applicability in the intellectual
property context. The first is known as express preemption and comes into play whenever
a federal statute expressly preempts and displaces state law on an area of its coverage. The
federal copyright statute contains such a provision in 17 U.S.C. § 301, which provides in
relevant part that:
On and after January 1, 1978, all legal or equitable rights that are equivalent to any of the
exclusive rights within the general scope of copyright . . . and come within the subject matter of
copyright . . . whether created before or after that date and whether published or unpublished,
are governed exclusively by this title. Thereafter, no person is entitled to any such right or
equivalent right in any such work under the common law or statutes of any State.
Again, the precise contours of this provision have been the subject of various court
decisions, especially as they relate to common law copyright and state misappropriation
doctrine.
Quite independent of express preemption, however, is a principle known as implied
preemption, which, as the name suggests, covers situations where the statute is silent on
the question of displacement, but the broader context/intent behind the federal statute
strongly suggests such displacement. Implied preemption is in turn of two kinds: conflict
preemption—situations where a state law is in direct conflict/contravention of a federal
law such that it is impossible to comply with both or where a state law creates an obstacle
to the working of federal law—and field preemption—where the enactment of the federal
statute suggests that Congress chose to occupy the entire field such that its decision not
to engage certain areas or questions is deemed to be conscious. Implied preemption
remains an issue in the intellectual property setting as well. In one well-known case, the
Supreme Court invalidated a Florida state law that granted protection for unpatented
boat hull designs using implied preemption (Bonito Boats v. Thunder Craft Boats, 489
U.S. 141 (1989)). The Court’s opinion recognized that in the intellectual property setting,
Congress’s decision to protect certain areas/subject matter evinced a conscious decision
that areas/subject matter not protected were meant to remain in the public domain, that
is, continue unprotected.
Preemption within intellectual property is therefore a tricky question in so far as every
omission from the coverage of the federal patent, copyright, and trademark laws can be
understood as reflective of a Congressional decision to avoid protection for the areas
omitted. Determining when and how such state laws survive—especially in the absence of

an express preemption provision—remains a complex endeavor which has had the discern-
ible effect of impeding state legislatures’ and courts’ willingness to recognize new rules,
rights, and doctrines within state intellectual property. Consequently, federal preemption
law operates as more than just a jurisdictional principle in intellectual property. Instead, it
works to shape the substantive content of state intellectual property regimes by requiring
courts (and legislatures) to find creative ways of differentiating state rules from the federal
landscape and justifying such rules on grounds that remain compatible with the overall
balance struck by Congress in its enactments. A prime example of such maneuvering
is seen in the Second Circuit’s revival of the hot news misappropriation doctrine as a
part of New York common law in the case of NBA v. Motorola, 105 F.3d 841 (2d Cir.
1997). By situating the doctrine within the domain of unfair competition, even though it
indirectly related to an informational asset, the court successfully managed to carve out
some space for the doctrine to work and in the process effectively resurrected it in true
common law style. Later courts, however, have been less willing to exhibit such creativity
in their opinions.
The law and economics literature on the common law does not directly address or adopt
this conception of the common law in its analysis, despite intermittently and implicitly
examining different state law subjects for their efficiency (e.g., trade secret law, misap-
propriation). Beyond this substantive examination, however, there remains an important
and underappreciated respect in which the state law variant of the common law has
correspondence in the law and economics literature relating to the idea of jurisdictional
competition.
Deriving from the work of the economist Charles Tiebout (1956) on jurisdictional
competition, scholars of law and economics have developed theories of ‘legal federalism’
to show how variations in state laws promote robust competition among states, enabling
citizens and businesses to vote with their feet and produce an efficient market equilibrium
for the public goods at issue (Bratton and McCahery, 1997). Tiebout’s original model
begins with the idea that citizen mobility is driven by rational preference revelations such
that decisions about relocation are incentivized and motivated by preferences for public
goods, which encompass local laws and regulations. This in turn produces local (i.e.,
state) governments to compete with each other over their public goods offerings to attract
citizens making their locational decisions. In the legal literature, this economic idea has
been applied over the years to different subjects such as corporate law, banking, and tort
law, among others.
While it has only ever infrequently been extended to discussions about state intellectual
property regimes (i.e., common law intellectual property), the Tieboutian analysis has
obvious application within this domain. As states vary their laws relating to trade secrets,
publicity rights, misappropriation, and other areas of intellectual property protection,
they clearly do so with the intention of both protecting existing interests within their
jurisdiction and attracting out-of-state businesses and investments. While a few scholars
have begun to explore this line of analysis within intellectual property, it represents an
obvious one where insights from public goods economics can inform and motivate the
analysis fruitfully.

VII. CONCLUSION
As described in this chapter, the term ‘common law’ has multiple connotations within the
law and economics of intellectual property. Each of these connotations brings something
unique to the analysis, both descriptively and normatively, a reality that is worth bearing in
mind in discussions about the subject. Somewhat disappointingly, the subject has thus far
received little systematic analysis and scrutiny from scholars of intellectual property, despite
its embodying a variety of theoretical, normative, institutional, and practical lessons.
REFERENCES
Balganesh, Shyamkrishna. 2010. ‘The Pragmatic Incrementalism of Common Law Intellectual Property,’ 63
Balganesh, Shyamkrishna. 2013. ‘Stewarding the Common Law of Copyright,’ 60 Journal of the Copyright
Society of U.S.A. 103–26.
Blackstone, William. 1765. Commentaries on the Laws of England. Oxford: The Clarendon Press.
Bratton, William, and Joseph McCahery. 1997. ‘The New Economics of Jurisdictional Competition:
Devolutionary Federalism in a Second-Best World,’ 68 Georgetown Law Journal 201–78.
Brudner, Alan. 1995. The Unity of the Common Law. Berkeley, CA: University of California Press.
Cardozo, Benjamin. 1921. The Nature of the Judicial Process. New Haven, CT: Yale University Press.
Cardozo, Benjamin. 1924. The Growth of the Law. New Haven, CT: Yale University Press.
Coke, Edward. 1794. Coke’s Institutes of the Laws of England. London: E. and R. Brooke.
Cooter, Robert. 1979. ‘Liability Rules, Limited Information, and the Role of Precedent,’ 10 Bell Journal of
Economics 366–73.
Eisenberg, Melvin. 1988. The Nature of the Common Law. Cambridge, MA: Harvard University Press.
Frank, Jerome. 1930. Law and the Modern Mind. New York, NY: Brentano’s.
Friedman, David, William Landes, and Richard Posner. 1991. ‘Some Economics of Trade Secret Law,’ 5 Journal
of Economic Perspectives 61–72.
Holmes, Oliver Wendell. 1870. ‘Codes, and the Arrangement of the Law,’ 5 American Law Review 1–13.
Holmes, Jr, Oliver Wendell. 1881. The Common Law. Boston, MA: Little, Brown and Company.
Kornhauser, Lewis. 1989. ‘An Economic Perspective on Stare Decisis,’ 65 Chicago-Kent Law Review 63–92.
Llewellyn, Karl. 1960. The Common Law Tradition: Deciding Appeals. Boston, MA: Little, Brown and Company.
Llewellyn, Karl. 2011. The Theory of Rules. Chicago, IL: University of Chicago Press.
Mahoney, Paul. 2001. ‘The Common Law and Economic Growth: Hayek Might Be Right,’ 30 Journal of Legal
Studies 503–25.
Posner, Richard. 1973. Economic Analysis of Law. Boston, MA: Little, Brown and Company.
Posner, Richard. 1992. Economic Analysis of Law. Boston, MA: Little, Brown and Company. 4th ed.
Pound, Roscoe. 1906. The Spirit of the Common Law. Boston, MA: The Boston Book Company.
Priest, George. 1977. ‘The Common Law Process and the Selection of Efficient Rules,’ 6 Journal of Legal Studies
65–82.
Rubin, Paul. 1977. ‘Why Is the Common Law Efficient?,’ 6 Journal of Legal Studies 51–63.
Rubin, Paul. 1980. ‘Decision Making and the Efficiency of Law: A Comment on Rizzo,’ 9 Journal of Legal
Studies 219–334.
Rubin, Paul. 1999. ‘Judge-Made Law,’ in Boudewijn Bouckaert and Gerrit de Geest, eds., Encyclopedia of Law
and Economics, Volume V. The Economics of Crime and Litigation. Cheltenham, UK and Northampton, MA,
Schauer, Frederick. 1989. ‘Is the Common Law Law?,’ 77 California Law Review 455–72.
Solum, Lawrence. 2004. ‘Procedural Justice,’ 8 Southern California Law Review 181–231.
Tiebout, Charles. 1956. ‘A Pure Theory of Local Expenditures,’ 65 The Journal of Political Economy 416–24.
Case List
Barclays Capital, Inc. v. Theflyonthewall.com, Inc., 650 F.3d 876 (2d Cir. 2011).
Bonito Boats v. Thunder Craft Boats, 489 U.S. 141 (1989).
Campbell v. Acuff-Rose Music, Inc., 510 U.S. 569 (1994).

Cariou v. Prince, 714 F.3d 694 (2d Cir. 2013).

Erie R.R. Co. v. Tompkins, 304 U.S. 64 (1938).
Folsom v. Marsh, 9 F. Cas. 342 (C.C.D. Mass. 1841).
International News Service v. Associated Press, 248 U.S. 215 (1918).
NBA v. Motorola, 105 F.3d 841 (2d Cir. 1997).
United States v. Carroll Towing Co., 159 F.2d 169 (2d Cir. 1947).
17 U.S.C. § 107.
17 U.S.C. § 301.
35 U.S.C. § 103.
H.R. 94-1476, 94th Cong. (2d Sess. 1976).

20. In the shadow of the law: the role of custom in
Jennifer E. Rothman* 2
Contents
I. Introduction
II. The Role of Custom Outside IP
III. Industry Practices and Social Norms in the IP Space
A. Informal Industry Practices and Social Norms
1. Clearance culture
2. IP-adjacent norms
B. Formalized Practices and Guidelines
1. Classroom Guidelines
2. In-house guidelines
3. Best practices statements
4. Creative Commons
IV. The Role of Custom in Formal Law
A. Custom as Evidence of Market Effects, Commerciality, and Damages
B. Custom as a Proxy for What Should Be Done
C. Custom as a Proxy for What Is Reasonable
D. Custom as Evidence of What Is Usually Done
E. Custom as Evidence of What Parties Intended
V. Questioning Reliance on Custom
A. The Questionable Optimality of Industry-Driven IP Practices
B. Expectations Should Not Determine IP Rights
C. Autonomy Interests Do Not Justify Reliance on Custom
VI. Valuing Custom
A. Certainty of the Custom
B. Motivation for Custom
C. Representativeness
D. Implications
VII. Lessons for IP Policy
References
* Professor of Law and Joseph Scott Fellow, Loyola Law School, Loyola Marymount
University, Los Angeles. I consider many of the issues raised in this chapter in greater detail in ‘The
Questionable Use of Custom in Intellectual Property,’ 93 Virginia Law Review 1899 (2007).
526

In the shadow of the law: the role of custom in IP 527
I. INTRODUCTION
Custom encompasses many different things, from regularly occurring industry practices,
to social norms, to ongoing practices that have existed from time ‘immemorial.’ Custom
in all these senses has had a tremendous influence on intellectual property (‘IP’) law, from
affecting what happens outside of the courts in the trenches of the creative, and technol-
ogy- and science-based industries, to influencing how courts analyse infringement and
defenses in IP cases. For decades, many scholars overlooked or dismissed the impact of
custom on IP law. This happened in part because of a belief that the dominant statutory
frameworks that govern IP left little room for custom, and also because of a view that
‘intellectual goods’ lacked the ‘longstanding’ customs and norms that exist in the context
of real property (Carter, 1992, p. 131; Long, 2004, p. 484). In the last ten to fifteen years,
however, the landscape has shifted and more attention has been given to considering
how custom affects IP entitlements both outside and inside the courtroom. Scholars like
myself have brought attention to the profound impact custom has in IP. My work has
particularly focused on the theoretical frames that inform the incorporation of custom
into the law, as well as on documenting some of the practices and norms of various com-
munities that use IP (Rothman, 2007).
In this chapter, I provide an overview of the role of custom in IP, and the scholarship
in the field. I first situate the discussion in the broader context of the treatment of custom
outside of IP, and then consider some of the dominant practices and norms within IP,
focusing particularly on those involved in copyright law. I discuss the incorporation of
custom by the courts and criticize the often unreflected reliance on custom. After provid-
ing this background, I question the relevance of most customs to set legal standards in
IP disputes, and suggest limits on custom’s role in IP cases. Finally, I suggest implications
that flow from this analysis, as well as recommend future areas of research for scholars.
II. THE ROLE OF CUSTOM OUTSIDE IP
To understand the role of custom in IP, one must first contextualize the treatment of
custom in the law more broadly. The importance of custom in determining governing
law has a long tradition in Anglo-American law. One of the foundational features of
English common law was its use of custom to set legal rules. Prior to the institution of
an organized legal system, practices and norms regulated local behavior and facilitated
the resolution of disputes. As more formal legal systems developed in England, custom
shaped and sometimes defined the law. The incorporation of custom by courts served an
important role in getting communities to support the authority of the growing judiciary.
Under common law dating back to at least the late 1400s in England, ‘general customs’
formed the basis of the law. William Blackstone (1765), one of the foremost commentators
on the early common law, defined the common law in his influential eighteenth-century
Commentaries on the Laws of England as ‘[t]hat ancient collection of unwritten maxims
and customs [that] had subsisted immemorially.’ The two main advantages of using long-
standing community customs (either local or kingdom-wide) were that they were thought
to be ‘universally known’ and were viewed as originating with the communities and people
rather than being imposed by the king. Communities therefore were more willing to defer

to custom-based legal rules that largely reflected their prior understanding of appropriate
conduct (pp. 17, 45, 63–4, 67–8, 76–8; Baker, 2002, pp. 1–10; Postema, 1986, pp. 3–4).
Much of the Blackstonian discussion of custom focused on its role in defining the
scope of public use and access rights to private land. In contrast to property doctrines
like prescription, custom permitted access and use not by a particular person but by the
public at large. In a number of instances, the public obtained access and use rights to
private property on the basis of prior customary uses of that land. English courts held
that the public could hold annual dances, conduct horse races, play cricket, fish, gather
wood, and graze animals on private lands because they had customarily used the land for
those purposes (Blackstone, 1765, 76–8; Rose, 1986, pp. 739–41, 758–9). Carol Rose has
described some of these customary uses as ‘recreational’ in nature and preferred because
they supported social engagement and connections within a community (Rose, 1986,
pp. 723, 767–70, 779–81).
Many of the customary uses were also related to providing basic subsistence needs.
During the enclosure movement in England beginning in the sixteenth and seventeenth
centuries, landowners increasingly excluded citizens from land that they had previously
relied on for food and fuel. The English historian E.P. Thompson (1991) describes
custom during this period as a response to this enclosure of the land. The customary
use arguments challenged efforts by property owners to move property in the direction
of a virtually absolute right of the landowners with no permissible public use or access
(Thompson, 1991, pp. 106–84).
Rather than being the preferred starting point for legal rules, today the status of custom
is contested and debated. Different areas of law (and different inquiries within those
areas) treat custom differently. In tort law, for example, there are longstanding debates
about whether the development of customary safety precautions by a particular industry
should be an absolute defense to tort liability, no defense at all, or simply some evidence of
negligence or lack thereof. The dominant contemporary principle is that custom should
be some evidence of reasonable care, but not its measure (The T.J. Hooper, 60 F.2d 737
(2d Cir. 1932); Landes and Posner, 1987, pp. 132–3; Epstein, 1992a; Morris, 1942). In
contract law, there is a developed literature analysing whether industry practices should be
read into contracts as implied terms and also, less controversially, whether such practices
should inform the interpretation of existing contract terms (Bernstein, 1999; Epstein,
1999). In property law, custom primarily arises as a basis to assert public access to land
that has long been used despite competing private property claims, often in the context
of beaches (State ex. rel. Thornton v. Hay, 462 P.2d 671, 676–8 (Or. 1969); Rose, 1986,
pp. 713–14).
Scholars who have considered custom and the law largely have focused on how
custom can govern various communities without regard to formal laws or adjudica-
tory mechanisms. Robert Ellickson’s (1991) influential book Order without Law: How
Neighbors Settle Disputes considered the ranching practices of cattle ranchers in Northern
California, and determined that social norms and longstanding practices trumped more
formal legal rules and discouraged resort to the legal system. The political scientist Elinor
Ostrom (1990), in her book Governing the Commons: The Evolution of Institutions for
Collective Action, similarly analysed the way various communities develop systems of
self-government and self-organization to manage and control the use of common-pool
resources. Lisa Bernstein (1992; 2001), in a series of articles, documented a variety of

industry practices that govern relationships in different commercial fields, including the
diamond and cotton industries.
Over the last decade, scholars have begun to recognize that IP is not an exception to
these narratives about custom, but instead yet another example of the influence of custom
on both de facto and de jure rights. Custom has a powerful impact on what is happening
in the trenches of creative and other IP-dominated industries and also influences the
governing legal regimes. Just as the enclosure movement in England sparked arguments
in favor of granting customary use rights to the public, concerns over the increased
propertization of intangible works has generated efforts to articulate similar justifications
for public use of works that can form crucial pieces of our identities and culture.
III. INDUSTRY PRACTICES AND SOCIAL NORMS IN THE IP

SPACE
Numerous industry practices and norms govern how IP and quasi-IP rights function as
a de facto matter. In copyright law, custom has affected determinations of authorship,
ownership, copyrightability (such as whether something is original), and whether a use
is infringing—especially whether something is an idea or expression, or composed of
scènes à faire (unprotectable stock or commonplace elements). Uncertainties in IP law
incentivize the creation of custom. Some of this uncertainty is generated by the significant
impact that changing technology has had on the production and distribution of IP. Other
uncertainties in IP law are generated by flexible, and sometimes unpredictable, legal
standards. The best example of this is copyright’s fair use defense. A four-factor analysis
is used to evaluate whether a use is fair and therefore not infringing. The multi-factor
analysis has often been criticized as ‘muddled,’ ‘troublesome,’ and ‘ad hoc’ (Weinreb, 1990,
pp. 1138–140; Dellar v. Samuel Goldwyn, Inc., 104 F.2d 661, 662 (2d Cir. 1939); Leval,
1990, p. 1105). Because of the unpredictability and expense of litigating fair use defenses,
many players in the IP industries prefer to agree among themselves on some boundaries
of fair use or play it safe by conforming to industry practices, such as licensing, rather
than risk adverse court decisions if they guess wrong about a potential fair use (Rothman,
2014, pp. 1602–5). I consider a variety of practices that have developed in the IP space,
particularly those that have arisen in the context of copyright law, and the navigation of
its fair use defense.
A. Informal Industry Practices and Social Norms
1. Clearance culture

One of the most influential set of practices is the common licensing of works, marks,
inventions, and identities. Producers, publishers, and distributors often require creators
and inventors to license or ‘clear’ all potentially protected IP, even when there are strong
defenses for the use of works, trademarks, or inventions, or when the protectability of
the work, mark, or invention at issue is questionable. Instead of challenging the validity
of the copyright, trademark, or patent, or relying on fair use, First Amendment, or other
defenses, IP users often seek clearance. Patricia Aufderheide and Peter Jaszi (2004) have
dubbed this preference for licensing the ‘clearance culture’ (p. 22; Heins and Beckles, 2005,

pp. 5–6). The clearance culture is primarily motivated by efforts to avoid litigation and
operates without regard to what IP law requires or what, as a normative matter, should be
protected by IP rights. As the Sixth Circuit has observed, it is ‘cheaper to license than to
litigate’ (Bridgeport Music v. Dimension Films, 410 F.3d 792, 802 (6th Cir. 2005)). When
works, marks, or people’s identities cannot be licensed, gatekeepers often demand their
removal.
These clearance practices are firmly entrenched in all media, including music, fine
arts, and publishing. Clearance culture can be seen, for example, in limits on the content
of biographies. Even though courts have traditionally given great latitude to authors to
refer to individuals, trademarks, and copyrighted works without permission in histori-
cal, nonfiction works, publishers routinely demand clearance of a subject’s copyrights,
trademarks, and publicity rights (Max, 2006, pp. 34, 37–8). Many of the potential IP
claims in such circumstances are facially meritless, but risk-averse publishers and authors
nevertheless abandon projects or follow the restrictions set forth by alleged property
holders (Max, 2006, pp. 34, 37–8; Max, 2007, pp. 54, 66). The film and television industry
similarly clears potentially copyrighted or trademarked works or marks, as well as images
and references to individuals, especially well-known public figures, even when the uses
would likely be determined fair if litigated (Rothman, 2007, pp. 1912–15). Inventors and
developers also often license patents when the validity of a patent is questionable or the
infringing status of an inventor’s product is uncertain. Companies license to avoid costly
patent litigation and hold-up problems with a product that has already been developed or
marketed (Farrell and Merges, 2004, pp. 955–60; Lemley and Shapiro, 2007, pp. 1992–3).
Clearance culture practices have a profound influence on what gets made and the con-
tent of such works. When licensing is not an option, either because it is cost-prohibitive
or an IP owner does not like the way its IP will be used, creators and inventors often alter
their works or forgo some projects altogether. Clearance culture practices are enforced
extrajudicially by fear of litigation costs, in-house policies mandating clearance, concerns
over forfeiting large investment or start-up costs, and by limits on funding, insurance,
and distribution.
2. IP-adjacent norms

Several recent scholarly works have analysed communities in which traditional IP law
does not function well or participants choose not to pursue legal remedies, even when
they are available. Much of this literature focuses on critiquing the incentive rationale for
protecting IP by demonstrating that creative works and inventions are produced in the
absence of IP protection (Raustiala and Sprigman, 2006). But this scholarship also reveals
the myriad ways in which communities can erect customary protections for creative and
inventive works outside the judicial system and enforce them using community norms.
Economists Emmanuelle Fauchart and Eric von Hippel (2008) have documented norms
to protect food recipes in the absence of effective formal IP law. They found that a variety
of customs govern French chefs, including norms that the chefs should not copy or share
recipes without permission. The chefs also encourage and seek attribution for their recipes
and innovations. Shaming and ostracizing serve to enforce these norms. Dotan Oliar and
Christopher Sprigman (2008) have identified similar practices in the world of stand-up
comedy. They found norms that discourage joke-stealing and encourage originality. These
norms are enforced by obstructing employment, shaming, and the occasional use of

v iolence. David Fagundes (2012) has studied norms surrounding the adoption and polic-
ing of roller derby names using a master roster kept by community members and enforced
by positive reinforcement and shaming. Drawing on the work of Elinor Ostrom in the
world of common-pool resources, Michael Madison, Brett Frischmann, and Katherine
Strandburg (2010) have surveyed a variety of IP-based communities, such as those that
use open-source software, or that contribute to Wikipedia and the Associated Press. They
identify practices that help manage these ‘cultural commons.’
Norms governing the use of IP have also developed in a variety of subcultures that rely
on uses of others’ IP. In the world of online fan fiction, for example, norms require that
attribution be given when material is borrowed from another fan’s website. Copyright laws
have little sway in fan-fiction communities. Instead, fan-fiction authors conform to their
own social norms, such as that the sites be nonprofit, add creative material to the original
material, and provide appropriate credit. Deviation from the fan-fiction norms may lead
to shaming within that subculture, which is usually enough of a deterrent to keep the
norms intact (Tushnet, 2007).
B. Formalized Practices and Guidelines
More formal customary practices have also been used in an effort to avoid litigation.
Many of these guidelines and agreements have developed in the context of copyright’s
fair use doctrine. Some industries and user groups have sought to insulate themselves
from liability for copyright infringement by agreeing in a more formal manner to a set
of standard copying practices. I highlight some of the most influential of these policies,
guidelines, and ‘best practices’ statements.
1. Classroom Guidelines

The most influential of the extralegal copyright guidelines is the ‘Agreement on Guidelines
for Classroom Copying in Not-For-Profit Educational Institutions,’ commonly referred
to as the ‘Classroom Guidelines.’ While drafting the fair use section of the 1976 Copyright
Act, Congress recruited industry representatives, in particular publishers, to develop their
own guidelines for what constituted fair use of writings and music in educational settings.
The chairman and other members of the subcommittee working on the copyright revision
‘urged the parties to meet together independently in an effort to achieve a meeting of the
minds as to permissible educational uses of copyrighted material’ (H.R. Rep. No. 94-1476,
p. 67). Congress contended that ‘workable voluntary arrangements’ were the preferable
solution to questions regarding the scope of fair use, at least in the context of educational
uses (H.R. Rep. No. 90-83, pp. 33, 36; H.R. Rep. No. 94-1476, pp. 67–8).
The Classroom Guidelines take a narrow view of what sort of uses of copyrighted
works are permissible in the educational context. The Guidelines provide that single
copies may be made for or by teachers for use in teaching or research. These copies are
limited to those of ‘a chapter from a book, an article from a periodical or newspaper, a
short story, short essay or short poem [and a] chart, graph, diagram, drawing, cartoon or
picture from a book, periodical or newspaper.’ Multiple copies, not exceeding one copy
per enrolled student, are permitted under limited circumstances in which such uses meet
tests for brevity, spontaneity, and cumulative effect. Brevity is defined as less than 250
words of a poem and 500–2500 words of a prose work. The copies must also include a

notice of copyright (H.R. Rep. No. 94-1476, pp. 68–9). Although the Guidelines purport
to set forth the minimum allowable uses, many universities, other educational institu-
tions, and libraries have followed them as if they represent the maximum allowable uses
(Rothman, 2007, pp. 1920–21). Many universities have handed out the Guidelines to their
professors and mandated conformity with them. In 2006, William W. Fisher and William
McGeveran (2006) estimated that 80 percent of American universities comply with the
Guidelines (p. 57). Recently, a few universities have moved away from this conservative
approach. The University of Minnesota, for example, recently agreed to defend profes-
sors if they reasonably believe that their use of a copyrighted work is fair, even if the use
exceeds the Classroom Guidelines (University of Minnesota, 2017). At the beginning
of 2014, New York University also withdrew its requirement that faculty comply with
the Classroom Guidelines and now allows its faculty to conduct an independent fair use
analysis (compare New York University (2014) with New York University (1983)).
2. In-house guidelines

Many companies and organizations have developed internal guidelines that govern the
treatment of IP within their own institutions. In both the public and private sectors, guide-
lines have been developed to control internal copying and the use of others’ inventions,
works, marks, and identities. The clearance practices that occur as an informal practice are
often specifically mandated by in-house guidelines. In the film and television industry, for
example, networks, studios, and production companies develop ‘Standards and Practices’
which control content, including the use of copyrightable works, trademarks, names, and
images. Most film studios mandate the clearance of all copyrighted works regardless
of the manner in which they appear, the elimination of any references to trademarks in
dialogue, the removal of or blurring of trademarks that appear on screen, and the clear-
ance or removal of proper names. Many libraries also have developed in-house guidelines
to regulate photocopying, interlibrary loans, and journal purchases. The primary purpose
of these guidelines is to reduce the likelihood of a lawsuit, and, if sued, to reduce the
likelihood of findings of bad faith or ‘willfulness’ on the basis of compliance with such
internal guidelines. In the context of the entertainment industry, some of these practices
also facilitate product placement and advertising deals (Rothman, 2007, p. 1922).
3. Best practices statements

Scholars and various use communities have recently sought out custom as a way to define
and establish fair use. This interest in custom is not only driven by efforts to persuade
courts to accept defenses in individual cases, but also by efforts to encourage individuals
and organizations to assert fair use rather than to conform with the dominant, risk-averse
clearance culture. Most notably, the best practices statements developed by Peter Jaszi,
Patricia Aufderheide, and others at American University and its Center for Social Media
seek to establish boundaries of fair use in the context of communities that frequently use
copyrighted works owned by others.
The Documentary Filmmakers’ Statement of Best Practices (Center for Social Media,
American University, 2005) is the most well known of these statements. This statement
was the first one released by the Center and sets forth categories of uses of others’
copyrighted works that its drafters consider fair in the context of documentary films.
The privileged categories are critique or commentary, illustrative quoting, incidental

uses (those captured during the filming process), and uses in historical sequences. Each
of these categories contains a number of ‘limitations.’ In the context of the illustrative
category―in which uses are generally considered fair if the use ‘illustrate[s] an argument
or point’―such preferred uses are limited to instances in which documentarians:
assure that: the material is properly attributed . . . [; that] quotations are drawn from a range
of different sources[; that] each quotation . . . is no longer than is necessary to achieve the
intended effect; [and that] the quoted material is not employed merely in order to avoid the cost
or inconvenience of shooting equivalent footage.
These and the other limitations dramatically contract the scope of permissible uses under
these statements.
The Filmmakers’ Statement appears to have encouraged more reliance on fair use. After
its adoption, various film industry gatekeepers, such as errors and omissions (‘E & O’)
insurers and production companies, reconsidered their policies and have become more
willing to insure and distribute documentary films that have not licensed all copyrighted
material used in the films (Aufderheide and Jaszi, 2007). It is difficult, however, to know
exactly how much influence the statements had, given that a number of other changes
in the IP landscape occurred during this same time period. For example, legal clinics at
major law schools and nonprofit organizations offered to provide free legal representation
to those asserting fair use, and advocacy groups lobbied E & O companies to change their
procedures (Rothman, 2010a).
4. Creative Commons

Some efforts to encourage the use of copyrighted works have focused on creators, rather
than users. A subsection of creators prefer a more permissive copyright regime than the
default—one that makes it easier for third parties to use and share works. One of the most
successful of these creator-focused efforts is Creative Commons, a nonprofit organization
formed in 2001 with the idea of layering an alternative, formalized licensing regime on top
of existing copyright law. Hundreds of millions of works have been licensed using Creative
Commons licenses. Major bands and recording artists, such as Nine Inch Nails and
David Byrne, have used these licenses, as have Al Jazeera, Google, the California Digital
Open Source Library, and the White House (Creative Commons Brief as Amici Curiae,
Jacobsen v. Katzer, 535 F.3d 1373 (Fed. Cir. 2008); Creative Commons, 2015; Rothman,
2014, p. 1625). The most common Creative Commons licenses require attribution, but
allow noncommercial derivative works or adaptations if the new work is distributed in a
share-alike manner—meaning under the same Creative Commons licensing regime under
which it is licensed (Creative Commons, 2015). Thus, these licenses alter the usual baseline
of copyright protection which does not require attribution for uses to be determined fair,
and does not allow uses of noncommercial works that do not otherwise meet the criteria
for finding that they are a fair use of the work.
IV. THE ROLE OF CUSTOM IN FORMAL LAW
These customary practices have not only affected how things operate on the ground, but
also the formal law, largely by influencing courts in IP cases. Courts sometimes point to

nonconformity with industry practices as a basis to find infringement or to reject defenses

to infringement. Only rarely have courts referred to conformity with industry practices
as a possible basis for a defense. When courts incorporate custom, either implicitly or
explicitly, they often use customary practices as proxies for other considerations, such as
what constitutes a ‘reasonable’ or ‘ethical’ use of another’s IP or what will be the market
impact of allowing such uses. I consider here the primary ways that courts use custom
when deciding IP issues, and raise some concerns with the current treatment of custom.
A. Custom as Evidence of Market Effects, Commerciality, and Damages
Courts often consider what is customarily done as evidence of whether there is a negative
market effect caused by the use of another’s IP. The most prominent example of this is
when courts look at ‘customary pricing’ in evaluating copyright’s fair use defense. Two of
the four statutory factors for determining fair use involve consideration of the market for
a copyrighted work. The first factor looks at the character of the use and in particular
whether the use is commercial. A nonprofit or noncommercial use weighs in favor of
a finding of fair use, while a commercial use weighs against such a finding. The fourth
enumerated factor also considers the market by asking courts to consider how the relevant
use will affect the market for the copyrighted work (17 U.S.C. § 107). Courts view both
existing and potential licensing markets as an indication of whether a use is for profit (or
‘commercial’) and also whether a given use is likely to harm the market for the work at
issue. Courts have sometimes used custom to determine the existence of such markets.
In Ringgold v. Black Entm’t Television, 126 F.3d 70 (2d Cir. 1997), for example, the
United States Court of Appeals for the Second Circuit rejected a fair use defense when a
television sitcom used the plaintiff’s artwork in the background of a set without permis-
sion. The court pointed to the custom in the TV and film industries of licensing copy-
righted works used as set-dressing. If not for the consideration of these industry clearance
practices, Black Entertainment Television (‘BET’) had a convincing fair use defense—the
poster containing the plaintiff’s artwork was visible for less than 30 seconds, was never the
focus of any shot, and was not referred to in the dialogue. The Ringgold court, however,
concluded that BET had failed to pay the ‘customary price’ for using Ringgold’s work by
not licensing her art and therefore could not avail itself of the fair use defense.
Courts have relied on similar industry licensing practices to evaluate the legitimacy
of photocopying for educational and research purposes. In Princeton Univ. Press v.
Mich. Document Servs., 99 F.3d 1381, 1391 (6th Cir. 1996), the Sixth Circuit Court of
Appeals rejected a fair use defense in the context of university course packets in large part
because the defendant did not follow the industry practice of licensing works used in such
packets. Similarly, in Am. Geophysical Union v. Texaco, Inc., 60 F.3d 913 (2d Cir. 1994),
the Second Circuit rejected Texaco’s fair use defense largely because of noncompliance
with industry custom. The court was persuaded that the copying of journal articles by
Texaco’s research scientists was unfair at least in part because many major corporations
got licenses for similar copying. Nonconformity with industry practices in both cases
convinced the courts that the uses were commercial and caused market harm.
These courts’ analyses stem in part from the Supreme Court’s decision in Harper &
Row, Publishers, Inc. v. Nation Enters., 471 U.S. 539 (1985). In Harper & Row, the Court
concluded that: ‘[t]he crux of the profit/nonprofit distinction is not whether the sole

motive of the use is monetary gain but whether the user stands to profit from exploita-
tion of the copyrighted material without paying the customary price’ (p. 562). When a
defendant is found to have not paid the ‘customary price,’ the defendant’s use is often
judged ‘unfair.’
There are numerous problems with relying on the customary pricing analysis. The fact
that licensing may be common should not be used to determine that a use is for profit.
Even ‘educational’ and religious uses can be viewed by courts as commercial under such a
‘customary price’ analysis (Soc’y of Holy Transfiguration Monastery v. Gregory, 689 F.3d
29, 61 (1st Cir. 2012); Cambridge Univ. Press v. Patton, 769 F.3d 1232, 1261–268 (11th
Cir. 2014)). Several scholars, including myself, and a number of jurists have criticized
this reliance on licensing evidence and have warned of the circularity dangers inherent in
considering licensing opportunities as a basis for market harm. If a use is fair, there will be
no licensing market, and if a use is not fair, a licensing market will develop. When courts
rely on such licensing markets, particularly in an era of clearance culture, they abdicate
their role as independent evaluators of what uses are fair (Rothman, 2007, pp. 1933–4;
Princeton Univ. Press v. Mich. Document Servs., 99 F.3d 1381, 1391 (6th Cir. 1996); id.
at 1397 (Merritt, J., dissenting); id. at 1400–404 (Ryan, J., dissenting) (Am. Geophysical
Union v. Texaco, Inc., 60 F.3d 913, 929, 931 (2d Cir. 1994); id. at 936–9 (Jacobs, J., dis-
senting); Gibson, 2007, pp. 895–8; Pallas Loren, 1997, pp. 38–41).
B. Custom as a Proxy for What Should Be Done
Courts often view failure to conform with industry practices as both unethical and unfair.
In Roy Exp. Co. Establishment of Vaduz v. CBS, 503 F. Supp. 1137 (S.D.N.Y. 1980) and
Roy Exp. Co. Establishment of Vaduz v. CBS, 672 F.2d 1095 (2d Cir. 1982), both a district
court and the Second Circuit Court of Appeals held that failing to license film clips when
it is industry custom to do so was unethical and a basis for rejecting both fair use and First
Amendment defenses to copyright infringement. In Roy Export, CBS aired a retrospec-
tive on the great film actor and director Charlie Chaplin soon after his death. CBS incor-
porated in its broadcast footage from Chaplin’s films. In upholding a $700,000 jury verdict
against CBS, the district court found persuasive the fact that ‘CBS’[s] conduct violated not
only its own guidelines but also industry standards of ethical behavior.’ The court pointed
to the industry’s licensing practices as evidence of harm to the potential market for the
plaintiffs’ copyrighted works and of ‘bad faith’ (Roy Exp. Co. Establishment of Vaduz
v. CBS, 503 F. Supp. 1137, 1146–7 (S.D.N.Y. 1980)). Although the issue of fair use was
not raised on appeal, the Second Circuit, in affirming the district court decision, pointed
to CBS’s violation of its in-house guidelines and industry licensing practices as evidence
of ‘commercial immorality’ and a basis for rejecting its First Amendment defense against
unfair competition and copyright infringement claims (Roy Exp. Co. Establishment of
Vaduz v. CBS, 672 F.2d 1095, 1100, 1105 (2d Cir. 1982) (emphasis added)). The Supreme
Court in Harper & Row cited Roy Export when it set forth its ‘customary price’ standard,
suggesting that the customary price analysis is more of a normative concept than an
economic one (Harper & Row, Publishers, Inc. v. Nation Enters., 471 U.S. 539, 562 (1985);
Roy Exp. Co. Establishment of Vaduz v. CBS, 503 F. Supp. at 1144).
Even though the Classroom Guidelines are not legally binding and were intended to
state a minimum floor of allowable educational uses of copyrighted works, courts have

often viewed copying exceeding the Guidelines as unfair and done in bad faith. In Basic
Books v. Kinko’s Graphics, 758 F. Supp. 1522 (S.D.N.Y. 1991), a federal district court
considered the copy shop Kinko’s infringement in bad faith partly because Kinko’s in-
house handbook conceded that its copying practices for course packets exceeded those
permissible under the Guidelines (pp. 1544–5).
Defendants are also sometimes found to have acted willfully or recklessly when they
fail to conform with customary practices. Such findings can generate higher statutory
damages, punitive damages, and possibly criminal liability. By contrast, conformity with
custom often provides a basis for a finding of good faith even if infringement is ultimately
found, thereby limiting damages and avoiding the danger of criminal penalties.
C. Custom as a Proxy for What Is Reasonable
Sometimes courts try to figure out whether a use of another’s IP is reasonable or appropri-
ate, particularly in the context of copyright’s fair use defense. Although this is not strictly
speaking part of the fair use analysis, it often cuts to the heart of the evaluation a court is
trying to make. It is not easy to define what constitutes a reasonable use of another’s IP.
A reasonable use is not the same as a just or moral use; instead, like the reasonable person
standard in tort law, such a consideration asks more generally what is appropriate in a given
circumstance. Because it is difficult to determine when a use of another’s IP is reasonable,
courts often use custom as a shortcut or proxy for such a determination. Nowhere is this
approach more evident than in copyright’s fair use doctrine. The traditional common law
fair use standard asked courts to evaluate whether a use was ‘reasonable and customary’
(Shapiro, Bernstein & Co. v. P.F. Collier & Son Co., 26 U.S.P.Q. 40, 42 (S.D.N.Y. 1934);
Harper & Row, Publishers, Inc. v. Nation Enters., 471 U.S. 539, 550 (1985); De Wolf, 1925,
p. 143). Although this traditional formulation of fair use asks courts to consider both
what is reasonable and what is customary, modern courts often conflate the two inquiries
so that what is customary becomes what is reasonable.
Courts have judged uses fair solely on the basis that such uses have customarily been
practiced. This approach turns on a court’s conclusion that a use that has long been
allowed is reasonable. One example of this analysis is the common acceptance of the use
of copyrighted works in biographical works because such uses are considered ‘custom-
ary’ (New Era Publ’ns Int’l, ApS v. Carol Publ’g Grp., 904 F.2d 152, 157 (2d Cir. 1990);
Rosemont Enters. v. Random House, Inc., 366 F.2d 303, 307 (2d Cir. 1966)).
Another area where courts have relied on custom as a proxy for what is reasonable is
in the context of copyright’s work-for-hire doctrine. Many universities expressly allow
faculty members to retain copyrights over their lectures, course materials, and scholarly
works. In the few instances in which such faculty ownership has been contested,
courts have often concluded that such ownership is an appropriate exception to the
usual work-for-hire rules given the longstanding nature and commonality of the norm
allowing faculty to retain such rights (Hays v. Sony Corp. of Am., 847 F.2d 412, 416–17
(7th Cir. 1988); Weinstein v. Univ. of Illinois, 811 F.2d 1091, 1094 (7th Cir. 1987); but
see Forasté v. Brown Univ., 290 F. Supp. 2d 234, 238–9 (D.R.I. 2003); Pittsburgh State
Univ. v. Kansas Bd. of Regents, 122 P.3d 336, 345–7 (Kan. 2005)). Courts have also
looked to customary practices as an indication of what uses reasonable authors would
permit of their works―a consideration that some courts treat as highly indicative of

whether a use is fair (Harper & Row, Publishers, Inc. v. Nation Enters., 471 U.S. 539,
550–51 (1985)).
D. Custom as Evidence of What Is Usually Done
Sometimes courts look at evidence of custom simply to determine what is usually done
in a particular industry without using custom for second-order evaluations of what is
reasonable, ethical, or optimal. A number of legal issues in IP law legitimately require
consideration of industry practices. For example, whether particular features of a work
are common stock or standard elements, not capable of infringing another’s copyright
under the scènes à faire doctrine, turns in part on customary practices. Whether one can
use copyright law to protect characters that are mutant superheroes with special powers
turns on the conventions of the superhero genre. Because it is customary for superheroes
to have such qualities, a copyright holder cannot prevent others from also creating
superheroes who are mutants with extraordinary powers (Twentieth Century Fox Film v.
Marvel Enters., 155 F. Supp. 2d 1, 37–8, 42–3 (S.D.N.Y. 2001)). Similarly, in the context of
computer software, the doctrine of externalities denies copyright protection to aspects of
software that are standard programming features. Custom determines what programming
practices are considered standard programming features (Dun & Bradstreet Software
Servs. v. Grace Consulting, 307 F.3d 197, 214–16 (3d Cir. 2002); Comput. Assocs. Int’l
v. Altai, Inc., 982 F.2d 693, 709–10 (2d Cir. 1992)). Custom also matters in trademark
infringement and false endorsement cases, which turn on consumer confusion. Consumer
perceptions about whether permission is usually required for a particular use of a trade-
mark or celebrity name will likely influence whether consumers will think the use was
sponsored or endorsed.
E. Custom as Evidence of What Parties Intended
Finally, courts consider custom when determining parties’ intentions in explicit and
implied contracts related to IP rights, in claim construction of patents, and when
interpreting statutory language (May v. Morganelli-Heumann & Assocs., 618 F.2d 1363,
1367–8 (9th Cir. 1980); Jim Henson Prods. v. John T. Brady & Assocs., 16 F. Supp. 2d
259, 262, 267–77, 282, 285–6, 288–90 (S.D.N.Y. 1997)). Such straightforward uses of
custom are usually appropriate and not controversial (Rothman, 2007, pp. 1945–6, 1974,
1978–9).
V. QUESTIONING RELIANCE ON CUSTOM
Given courts’ frequent consideration and incorporation of custom, it is important to

consider both whether and when it is appropriate to rely on industry practices and social
norms to determine the scope of IP rights. Three main justifications have been advanced
for incorporating custom into other areas of the law: First, that a given industry can best
determine its optimal governing rules; second, that incorporating custom fulfills parties’
expectations; and finally, that individuals or industries should establish their own rules
to further their autonomy-based interests. To the extent that these are legitimate bases to

incorporate custom elsewhere, they are less convincing in the context of IP. I will briefly
consider each of these justifications, and how well they fit with considerations in the IP
realm.
A. The Questionable Optimality of Industry-Driven IP Practices
If an industry or community is likely to establish optimal practices, or at least rules

preferable to those that would independently develop in the courts or through legislation,
it makes sense to defer to industry practices. In the IP context, however, practices are likely
to develop in suboptimal ways and ultimately be inferior to court or legislative resolutions.
Defining what is meant by optimal practices in the context of IP is challenging.
Nevertheless, any IP regime requires a balancing of IP owners’ interests and those of
users. The incentive rationale (one of the primary justifications for IP) has a built-in
argument for allowing access to and use of works. Under the incentive-rationale theory,
copyrightable works and patentable inventions and discoveries are protected in part to
encourage their production and distribution. This justification rests not on authorial or
inventor rights, but instead on the encouragement of production for the public’s benefit.
Under this rubric, IP ownership rights should not be absolute, because then the goal
of promoting broader progress for the public good would be thwarted. The incentive
rationale also has a built-in ceiling—there is no need to provide additional protection once
no further incentive to produce is created. Accordingly, fulfilling the goals of the incentive
rationale requires consideration of the interests of both IP creators and users.
Personality-based justifications for IP protection also require the allowance of some
use of others’ copyrighted works both for creative, artistic purposes, and because users
and subsequent creators have their own personality-based interests in using others’ IP to
adequately express themselves and to describe their own reality (Rothman, 2010b). When
the integrity of the underlying work is not damaged and the creator’s interests are satisfied
through attribution, these personality-based approaches encourage the use of others’ IP.
Perhaps a pure labor-reward rationale—based solely on compensating creators or
distributors—would exclude all uses done without permission if it were the only justifica-
tion for IP laws. But it is not a stand-alone theory. In addition, under the logic of the
labor-reward rubric, users who exercise their own labor when reworking others’ creations
and inventions should reap the rewards of their own labor. Even if a labor-reward analysis
adequately explained IP laws and was the only justification for such laws, IP rights must
still be limited by free speech, free expression, and liberty interests. Accordingly, some uses
of others’ IP would be allowed.
The question then of how best to allocate IP centers on whether a decentralized,
industry-governed IP system is likely to adequately incentivize the production and distri-
bution of inventions and works and protect and reward authors and inventors, while at
the same time guaranteeing adequate use and access to other’s IP by authors, inventors,
and the public. Trademark protection similarly must balance the protection of businesses’
goodwill and the prevention of consumer confusion, with the need for both consumers
and competitors to reference and use others’ trademarks. In the context of publicity
rights, the law must balance both the need for the public to comment on and refer to real
people with the rights of those individuals to control and profit from the use of their
identities. How exactly one would divide up these rights is a matter of much debate, but

most people would agree that an optimal allocation of IP rights requires consideration of
these sometimes competing interests.
Given this model of optimal IP rules, customary practices on their own are not likely
to best reflect when exclusive rights should yield to permit access and use. Many of the
prevalent customs have developed to avoid litigation or preserve relationships by avoid-
ing conflict. These practices (such as those of the clearance culture) are not developed
to be optimal governing rules, but instead simply to promote harmony outside of the
legal system. The danger of allowing such risk-averse customs to define the scope of IP
rights should be apparent. As discussed, incorporating such behavior can greatly expand
infringement findings. Industry practices can establish a restrictive IP regime—one in
which virtually nothing is free and no uses are fair. IP owners and users do not view most
licensing practices and copying guidelines as optimal, nor as an expression of their pre-
ferred allocations of rights. Instead, even those who routinely license want the latitude to
contest, and sometimes litigate, when a license is not granted or is prohibitively expensive.
Inequalities in IP markets and the underrepresentation of the public increase the
likelihood that suboptimal practices will develop. One of the main arguments against the
incorporation of custom into tort law is that the market cannot adequately protect
the interests of third parties or the public because they have no role in the production of
the practices. Even plaintiffs who are in a direct relationship with a potential defendant,
such as a consumer of a defective product or an employee, may still lack bargaining power
or sufficient market options to exert pressure on potential tortfeasors. When parties do
not have equal bargaining power and are not in reciprocal positions, suboptimal practices
and norms are likely to develop. As Eric Posner has observed in his critique of the reliance
on norms as a source for legal rules, ‘once one abandons the unrealistic assumption that
parties have symmetrical positions, traditional theories of the efficiency of norms lose
their power’ (Posner, 1996, p. 1709). He suggests that ‘highly unequal endowments of
group members may be evidence of inefficient norms. The more powerful members may
prefer and enforce norms that redistribute wealth to them, even when those norms are
inefficient’ (Posner, 1996, p. 1727).
Incorporation of custom also can prevent the continued evolution of custom by
producing a lock-in effect—the incorporation of custom further entrenches the same
suboptimal customs. James Gibson, among others, has identified the troubling ‘doctrinal
feedback’ and rights accretion that stems from the consideration of licensing practices
in the context of IP. When courts consider licensing evidence, parties are more likely to
license, which makes courts more likely to once again rely on licensing evidence. As more
and more companies and individuals follow the licensing and other litigation-avoidance
practices, these customs drive conformity rather than the evolution of optimal practices
(Gibson, 2007, pp. 884–5; Rothman, 2007, p. 1955).
Even if various practices are efficient between the parties when weighing litigation
versus licensing costs, they should not be extrapolated to define IP rights more generally
or even in future transactions between the same parties. This is similar to the situation that
arises when parties to a contract wish to be bound by gap-filling terms based on custom
for efficiency’s sake. These gap-filling terms will not bind nonparties or outsiders, and
even the parties themselves can opt out of the custom in future contracts.
Some of the scholarly support for preferring custom over congressional and court-
made law is driven by reasonable concerns over the influence of special interest groups

in the drafting and passage of legislation. Richard Epstein, for example, contends that
reliance on custom ‘provides an effective bulwark against [the] bias and corruption’ that
pervade the legislative system (Epstein, 1992b, p. 86). Unfortunately, the same powerful
parties who influence legislation often also control the creation and development of cus-
tomary practices. Worse yet, the development of industry practices lacks the established
procedure for encouraging open debate and public commentary that exists in the context
of pending legislation. As Lloyd Weinreb has observed, the result of reliance on custom
is that ‘the better financed private interest’ will prevail, rather than that a ‘careful, sys-
tematic’ rule will develop to ‘serve the community as a whole’ (Weinreb, 1992, pp. 146–7).
Many of the practices and norms that I have discussed demonstrate this skewed devel-
opment. The Classroom Guidelines, for example, were negotiated and drafted primarily
by publishers and therefore unsurprisingly forward the agenda of publishing companies
rather than those of scholars, educators, students, or research institutions. The Guidelines
were adopted over the opposition of major universities and scholarly organizations, such
as the American Association of Law Schools (H.R. Rep. No. 94-1476, p. 72; Basic Books
v. Kinko’s Graphics, 758 F. Supp. 1522, 1535 (S.D.N.Y. 1991)). Similarly, the clearance
culture is driven by large corporations.
To the extent optimal customs have been identified in other areas of law, they usually
have developed in the context of close-knit communities in which community members
have ongoing relationships and in which the same types of transactions are repeatedly
conducted. Richard Epstein (1992b) has advocated that ‘custom should be followed in
those cases in which there are repeat and reciprocal interactions between the same parties,
for then their incentives to reach the correct rule are exceedingly powerful’ (p. 126). Robert
Ellickson (1991) has similarly concluded that close-knit communities are most likely to
develop welfare-maximizing norms (pp. 167, 187, 228, 267, 283). Henry Smith (2009),
using his information-costs model of property law, has also contended that custom’s
value dissipates outside of the close-knit communities in which it develops because parties
are no longer familiar with the practices. In criticizing the enforcement of norms in the
context of the Internet, Mark Lemley (1998) has emphasized that ‘[i]t is no accident that
virtually all of the empirical work on norms has taken place in small, close-knit communi-
ties with little change in membership over time.’ As a community becomes larger and
more diverse, there is less likely to be a ‘commonality of interest’ and norms are both less
likely to develop and more likely to develop without uniform agreement (Lemley, 1998,
pp. 1267–9).
IP markets are exactly these sorts of disaggregated spaces. Many IP transactions do
not involve repeat players or individuals who have any relationship with one another. A
documentary filmmaker may have no relationship with the Elvis estate or a major movie
studio, and neither the studio nor the Elvis estate is as likely to subsequently want to
license or use any material created by the documentarian. Nor does a person sitting at
home remixing video or music files usually have a relationship with particular production
companies, bands, or record companies, other than as a generic consumer.
Those who argue that the private sector is superior at allocating rights because it is
free to ‘independently’ develop ideal rules also overlook the legal shadow in which IP
transactions take place. In the context of IP, the governing customs are often generated in
response to legal regimes rather than on a clean slate. Despite such concerns, some, such as
Richard Epstein (2008), have argued in response to my work that the reliance on custom

in copyright cases involving negotiated licenses and other clearance culture practices
sometimes does reflect optimal private ordering based on mutually agreed upon pricing.
Epstein is more optimistic than I am that the clearance culture and other private-ordering
mechanisms will result in optimal delineations of use rights and pricing in the context of
copyrighted works. Negotiating licensing agreements is challenging, especially for smaller
players or when a potential user has a limited amount of time to obtain permission for
the use. Content owners sometimes cannot be located or do not respond (or at least not
in a timely manner) to requests. These challenges lead to significant transaction costs that
warp the market for these licenses. Content owners sometimes refuse to license at any price
or charge a prohibitively high or simply unreasonable fee for use.
In addition, because copyright furthers the overall public interest in generating more
works and more knowledge, we cannot simply look at an individual transaction and
evaluate the optimality between the owner and user, as compared to litigation costs—we
must also consider the costs to society more generally. Optimality in the sense of maximiz-
ing wealth is not the only consideration at issue. We must also maximize creativity and
knowledge. In short, custom is unlikely to independently establish optimal allocations of
IP, or to provide the best balance between the exclusionary rights of owners and the use
rights of other creators and the public.
B. Expectations Should Not Determine IP Rights
A second justification for incorporating custom is that doing so satisfies parties’ expecta-
tions. In tort law there has been significant scholarly debate about this very issue with
regard to the standard for negligence and, in particular, whether the negligence standard
should be governed by parties’ expectations or by a more objective standard. Richard
Posner views tort law not as furthering broad public policy goals, but instead as a
mechanism for fulfilling parties’ expectations when no formal contract governs a transac-
tion. ‘[T]he principle function of tort law,’ he has written from the bench, ‘is to protect
customers’ reasonable expectations that the firms with which they deal are complying with
the standard of care customary in the industry’ (Rodi Yachts v. Nat’l Marine, 984 F.2d
880, 889 (7th Cir. 1993)). Epstein similarly contends that customary practices should be
the standard of negligence when parties have some relationship to one another, even when
the relationship is inequitable, such as between employer and employee and retailer and
consumer (Epstein, 1992a, pp. 4–5, n. 14).
In the context of tort law, conformity with an industry practice may best reflect parties’
expectations, but as a matter of public policy there is concern that safety precautions will
lag if an expectations- or custom-based standard is adopted. The fact that a plaintiff was
on notice of a danger should not in and of itself end the inquiry. Tort law, like IP law,
is not another form of contract law in which individual parties’ expectations drive the
law. Instead, both bodies of law are in service to a higher purpose. Tort liability is not
solely about the parties before the court but is also about making society safer, protecting
third parties, and deterring bad or dangerous behavior. Given the bargaining power and
knowledge base of potential tort victims, it makes sense to protect consumers from the
possible race to the bottom that may result from deference solely to industry standards.
As a result, in most instances tort law requires an objective, external evaluation of what
is a reasonable standard of care.

Much of the literature supporting the incorporation of custom into contract law reflects
the goal of furthering contracting parties’ intentions and expectations. It makes sense to
incorporate custom into contracts, if it reflects the contracting parties’ understanding.
As Lisa Bernstein (1999) has noted, however, the mere existence of practices does not
indicate that parties would expect or want them to govern in ‘end-game’ disputes when
both a contract and relationship are breached.
In the context of traditional property rights, scholars have similarly debated whether
expectations should drive property rights. Property rights often have been justified on the
basis of expectations of entitlement to particular property. Carol Rose (2000a; 2000b)
has highlighted, however, that even though a party may expect certain property rights,
those rights should yield when they are unjust or otherwise not deserving of enforcement.
Rose emphasizes that expectations must often be frustrated to manage or protect scarce
resources or to promote social justice, tasks that often require limits on property rights
(2000a, pp. 485–6; 2000b, pp. 19, 22).
An analysis of IP law supports the views held by scholars critical of relying on
expectations-based models for tort, contract, and property law. Neither the expectations
of IP owners nor risk-averse IP users should govern the scope of IP rights. Patent and
copyright protection are provided by constitutional grant and explicitly require considera-
tion of the public interest separate from the property rights of IP owners. Copyright and
patent laws do not have as their primary purpose the promotion of authors’ rights, but
instead the promotion of the public interest more broadly. The United States Constitution
expressly states that the ‘exclusive Right to . . . Writings and Discoveries’ is granted for
the purpose of ‘promot[ing] the Progress of Science and useful Arts’ (U.S. Const. art. I, §
8, cl. 8). Accordingly, courts have often noted that they must ‘subordinate the copyright
[or patent] holder’s interest in maximum financial return to the greater public interest in
the development of art, science and industry’ (Rosemont Enters. v. Random House, Inc.,
366 F.2d 303, 307 (2d Cir. 1966)).
IP law therefore requires consideration of negative externalities worked on third par-
ties. An IP holder might expect, especially given the clearance culture, that no unlicensed
uses would be made of her work, yet public policy demands the use of some material for
commentary, scholarship, and for new works. If the public or IP owners have a narrow
view of the scope of IP rights, these expectations should not alter the congressional or
constitutional judgment about how best to balance IP holders’ rights with the public’s
right to use and access IP. For example, the New York Times licenses its news stories for
use in television shows and movies (Manly, 2006). The Times’ expectation that it can
extract compensation for news, however, should not alter copyright law’s guarantee that
facts remain in the public domain.
The more attenuated nature of the IP markets also suggests that there are fewer shared
expectations between the likely parties in IP disputes. Most IP cases involve parties who
have no direct relationship with one another and often no shared set of expectations.
One only needs to listen to the United States Patent and Trademark Office’s roundtables
on the legitimacy of mash-ups to see the lack of common expectations. Large content
providers, such as Disney, think virtually no uses are fair, while some user-focused groups,
such as the Organization for Transformative Works, think virtually all uses are fair (Dep’t
of Commerce Internet Policy Task Force, 2014, pp. 64–142).
Expectations also are driven by customs and therefore lock in existing property regimes,

even when they are unjust and even if the parties would prefer alternative rules. In sum,
furthering parties’ expectations does not justify the wholesale incorporation of custom
into IP law.
C. Autonomy Interests Do Not Justify Reliance on Custom
The final common justification for incorporating custom is the furtherance of autonomy
interests. Early justifications for the common law expressed a preference for communities
being governed by their own customary laws that had evolved over a period of time. These
laws not only furthered parties’ expectations, but also promoted self-governance and
autonomy in a world that otherwise was dominated by rules instituted by the monarchy
without any community input. Today, the democratic process allows communities to
contribute in a more orderly fashion to the creation of governing laws, so there is less of
a need to rely on private ordering to protect citizens in the law-making process.
Most importantly, autonomy interests do not support reliance on custom because there
are conflicting autonomy interests at stake. It is not just initial creators and inventors who
have autonomy interests. Users, and subsequent creators and inventors, also have liberty
and autonomy-based interests. In the context of copyright, for example, copyrighted
works become a part of the personal and cultural world of others. The autonomy interests
of an author must therefore sometimes yield to the competing autonomy and liberty-
based interests of her audience (Rothman, 2010b).
VI. VALUING CUSTOM
The foregoing analysis suggests that custom should rarely be dispositive of questions
involving the scope of IP rights. Nevertheless, custom may provide some guidance about
what is reasonable or appropriate in a particular context. In the past, I have developed six
vectors that courts and scholars should consider when evaluating practices or norms in
the context of IP (Rothman, 2007, pp. 1967–80). Here, I will focus only on four primary
areas of evaluation: (1) the certainty of the custom; (2) the motivation for the custom; (3)
the representativeness of the custom; and (4) the implications of adopting the custom.
A. Certainty of the Custom
To have any value, a custom must be identifiable, in terms of what constitutes the practice
itself, and the practice must be widely accepted and followed. This analysis tracks the
traditional common law requirement that practices be certain before meriting judicial
consideration (Blackstone, 1765, p. 78; Browne, 1875, pp. 21–4; Lawson, 1881, pp. 32–6;
Rothman, 2013). Several considerations help to evaluate how certain a particular custom
is. First, if there is unanimity as to the contours of the custom among diverse parties it is
more likely to exist and to have definable boundaries. Such agreement about the practice
also likely indicates the consent of the community. Second, customs that are longstanding
are more stable and have weathered the test of time.
Because the best practices statements, such as the Filmmakers’ Statement, are more
wishful than descriptive and have fuzzy boundaries, they are not particularly certain.

Although the best practices statements purport to set forth the practices of the relevant
communities, they instead set forth what the drafters think the community should be
doing. In the context of the Filmmakers’ Statement, the report leading up to the state-
ment and the statement itself both reveal that the dominant practice was to license or
cut out copyrighted materials from documentaries. Additionally, many of the inquiries
in the Filmmakers’ Statement and other best practices statements do not provide certain
guidelines. Instead, they leave the same ambiguities of the existing fair use system, but
add an additional layer of complexity to the already uncertain fair use analysis. For
example, the OpenCourseWare Code requires evaluations of whether the ‘extent of the
use [is] appropriate,’ and requires attribution when ‘reasonably possible’ (Center for Social
Media, American University, 2009, pp. 13–14). The Filmmakers’ Statement requires
that quotations be ‘no longer than is necessary’ (Center for Social Media, American
University, 2005, pp. 4–5).
In other instances, conflicting customs are at work. For example, in the Roy Export
case—in which courts rejected fair use and First Amendment defenses for the use of clips
of Charlie Chaplin films in a TV obituary of Chaplin—the district court rejected a fair use
claim in part because it decided that the defendant did not conform to customary licens-
ing practices (Roy Exp. Co. Establishment of Vaduz v. CBS, 503 F. Supp. 1137 (S.D.N.Y.
1980); Roy Exp. Co. Establishment of Vaduz v. CBS, 672 F.2d 1095 (2d Cir. 1982)). The
court failed to consider, however, that there was more than one custom at work. Clips
were not usually licensed for obituaries even though they were often licensed in other
contexts for news projects with more lead time or in scripted series. Such conflicting
customs suggest either that the court needed to more carefully scrutinize which custom
was applicable or that there was no single, dominant, and widely accepted custom worthy
of consideration.
B. Motivation for Custom
Although motivation was not a common law limit on custom; custom was long thought
to reflect the preferences of a particular community. In other words, if the community
had been asked to sit around and agree to what the rule should be, this is likely the rule
they would have come up with—or at least if such a rule had been suggested to them
they would have agreed to it. In the context of copyright, the most valuable practices
and norms reflect preferred allocations between copyright holders and users, rather than
litigation-avoidance or relationship-preservation strategies. Reactive and risk-averse
customs, like those of the clearance culture, are not the sort of aspirational, independently
developed customs that we should adopt.
When customs develop with aspirational motives they are better indications of what is
appropriate. In the context of fair use, practices and norms should primarily be relevant
only to the extent that they are indicative of what is actually deemed ‘fair’ by the relevant
community, rather than what that community thinks is safe, or simply ‘cheaper’ than
risking possible litigation.
When considering the motivation for a particular custom, courts should particularly
identify whether the custom was intended to provide a reasonable balance between
competing interests. As a check on this analysis, courts should independently evaluate
whether reasonable people would agree to such rules if they knew neither whether they

would be powerful or minor players in the market, nor whether they would own or instead
want to use the relevant content.
C. Representativeness
Customs that represent only one party’s or one group’s interests are suspect. By contrast,
when a custom develops with input and participation by both owners and users, and large
and small players, it is more valuable. Practices and norms should also only apply to those
within the community that developed them. The Classroom Guidelines and best practices
statements were both developed in an unrepresentative way. They are therefore less
valuable guidelines and should not be wielded against those who did not participate (or
were not represented) in their development. As discussed, the Classroom Guidelines were
developed without the participation of educators, universities, and students. Similarly,
the best practices statements were developed without representation of the content
providers whose work is most likely to be used in documentaries. The fact that some of
the participating users were also authors does not remedy this one-sidedness.
Even though the Department of Commerce has stated its support of such use
guidelines in theory, it has appropriately noted that such guidelines need to be developed
with the participation of a variety of parties on different sides of the issues before such
guidelines deserve serious consideration (Dep’t of Commerce Internet Policy Task Force,
2013, p. 23). The proponents of the best practices statements are likely correct that if they
had invited major content owners to participate, very little would have been agreed to. The
very likelihood that the parties could not have agreed on common principles, however,
raises serious flags about using the best practices statements to affect entitlements outside
of the community that developed them. Even within various use communities, there have
been reasonable objections to applying the standards against a party who was not directly
involved in their development (Rothman, 2014, pp. 1620–22).
D. Implications
Courts must also independently scrutinize the implications of adopting any customary
practice or norm as a legal rule. When evaluating the worth of a particular custom, a court
must consider what the end result of incorporating that custom would be. If followed
to its logical conclusion, will the custom result in a slippery slope, such that no uses will
be allowed, or, alternatively, that too many uses will be allowed? Consider, for example,
two extremes. If it is customary to license everything, then no fair uses remain. On the
flipside, consider the peak of peer-to-peer file-sharing in which the custom was not to pay
for music downloaded from the Internet. Using such a custom to set the standard of fair
use could have destroyed the market for online music, and perhaps have obstructed the
development of today’s successful streaming services.
In the best practices statement related to user-generated content (‘UGC’) (in the
context of online video) virtually any use is deemed fair because the commentary and
critique category is read very broadly. In the report supporting the Online Video Code,
the drafters suggest that a mash-up titled Clint Eastwood’s ‘The Office’—which mixed
together clips from the television series The Office with the movie Evan Almighty—falls
within the favored category of negative or critical commentary. This category and

its exemplars suggest that all mash-ups are fair (Center for Social Media, American
University, 2008, pp. 7–9). This means that there can be no market for licensing such
mash-ups; a conclusion that limits new media markets and makes copyright law largely
irrelevant in the context of UGC. Many mash-ups may well be fair ones, but one cannot
simply wish them all to be so.
In sum, if custom is certain, representative, motivated by aspirational purposes, and
would result in a reasonable allocation of use and ownership rights, then that custom
provides meaningful guidance. Otherwise, such practices and norms should be met with
skepticism and little deference.
VII. LESSONS FOR IP POLICY
Despite the many reasons discussed to be cautious about jumping on the custom band-
wagon, custom provides a number of lessons for IP policy. First, massive disobedience
of IP laws can signal market failure or overreaching by IP owners. The IP system needs
public buy-in to work. Public support requires people to think that on some level the law
is fair. When laws are wildly out of sync with community practices, there sometimes will
be value in interpreting the law to conform to those understandings or amending the laws
to better reflect some of those norms.
Second, customary uses may demonstrate a consensus about preferred rights that may
not currently be recognized under the law. Such locations of commonality are promis-
ing areas for legislation. For example, many norms in the copyright world favor giving
attribution to authors when their work is used, but the law does not generally recognize
such a right. Similarly, it may be worth addressing the potential conflict between the work-
for-hire provision and the widely accepted norms of faculty ownership of scholarship and
teaching materials.
Third, custom may demonstrate areas of need by users and creators that should be
accommodated either through a reasonable market mechanism or through fair use.
Finally, because the value of custom is based on its reflection of a commonly agreed upon
norm, it is important to dissent from dominant and restrictive practices in IP markets.
Although IP laws should continue to provide room for private ordering, these private
efforts should not alter IP’s boundaries. The clearance culture in the publishing and film
worlds should not influence courts’ independent analyses of whether particular uses are
fair. Nor should a small cross-section of documentary filmmakers decide when fair use
applies in that context. Creative Commons licenses can encourage the use of copyrighted
works in ways that creators support, but the fact that a use breaches such a license should
not weigh against a finding of fair use.
Recent scholarship has turned a prying eye on the worlds of fashion, chefs, comedians,
roller derby, fan fiction, and many other IP-based or IP-adjacent communities in which
creativity sometimes flourishes without reliance on the governing formal IP structures
(Raustiala and Sprigman, 2006; Fauchart and von Hippel, 2008; Fagundes, 2012;
Tushnet, 2007; Madison et al., 2010). Sometimes these industries are categorized as
producing ‘creativity without law’ and used to support arguments that creativity can take
place without needing IP rights and enforcement (Darling and Perzanowski, 2017). But
the law influences even these purportedly IP-free zones. The scope of IP laws determines

when and how alternative protection schemes develop. Aspects of recipes not protected
by copyright law will by necessity be protected by alternative mechanisms—such as
attribution enforced by shaming. Fashion designs that are not protected by trademark or
copyright law will be protected through other mechanisms or will be altered (for example,
through a cycle of rapid change, or in ways that allow for trademark or copyright protec-
tion). Those within communities that use alternative regimes to protect their creations
do not always shun more formal IP protection. Roller derby players, for example, use a
community-based registry of derby names, but try to register their names with the Patent
and Trademark Office when they become better known, and such registrations have also
become more common as roller derby has become more popular. Chefs and comedians
sometimes rely on copyright law when they write cookbooks and produce recordings of
their routines (Fauchart and von Hippel, 2008; Oliar and Sprigman, 2008). The norms
of these various communities develop with an awareness of IP law, operate in its shadow,
and often are efforts to work around its contours.
As we look back on more than a decade of robust attention to custom, norms, and
practices in the context of intellectual goods and IP laws, and celebrate many of them,
it is important to contextualize them in the broader framework of the interplay of law
with these extralegal activities. Those who favor expansive use rights cannot simply
point to all the practices that they like without acknowledging the practices that they
do not like—those that promote owners’ interests—as well as the way courts have long
considered custom. Courts are just as likely—perhaps more likely—to incorporate
restrictive practices (such as clearance culture norms and restrictive use guidelines, like
the Classroom Guidelines) rather than more permissive ones. To the extent that we want
to distinguish between customs, we cannot do so based on a gut instinct about which
practices are preferable. Instead, a detailed framework like the one set forth here must
be used to evaluate the worth of each custom to determine if it merits incorporation
into the law.
As the project of documenting norms in various creator, inventor, and user communities
continues, it would be helpful to gather data about what knowledge or awareness (accurate
or not) community members have about the law and to what extent those understandings
influence their practices and norms. It would also be useful to collect information about
whether the characteristics that make custom more or less worthy of consideration are
present. For example, are the practices uniformly known, certain, and longstanding? Are
the practices motivated by efforts to set appropriate boundaries for use and ownership,
or by risk aversion, relationship preservation, or wishful thinking (by owners or users)?
Are the practices representative of a variety of parties, or only of one particular group?
Custom can provide valuable information, but its usefulness depends on independently
evaluating the worthiness of the custom and particularly scrutinizing its reasonableness.
The unarticulated incorporation of custom threatens to swallow up IP law and replace it
with industry-led IP regimes that give the public and other creators more limited rights to
access and use IP. If we take seriously the notion that IP is protected in the public interest,
then we cannot abdicate the boundaries of IP rights to delineation by privately developed
customary practices. This does not mean that we cannot appreciate community norms
that have developed in the IP space; but it does demonstrate that we cannot simply bask
in their glory without recognizing their place in a larger ecosystem of custom in the IP
world and beyond.

REFERENCES
Aufderheide, Patricia, and Peter Jaszi. 2004. ‘Untold Stories: Creative Consequences of the Rights Clearance
Culture for Documentary Filmmakers,’ Center for Social Media, American University, accessed at http://
archive.cmsimpact.org/sites/default/files/UNTOLDSTORIES_Report.pdf (last visited April 7, 2019).
Aufderheide, Pat, and Peter Jaszi. 2007. ‘Fair Use and Best Practices: Surprising Success,’ October Intellectual
Property Today a–b.
Baker, J.H. 2002. An Introduction to English Legal History, 4th ed. London: Butterworths LexisNexis.
Bernstein, Lisa. 1992. ‘Opting Out of the Legal System: Extralegal Contractual Relations in the Diamond
Industry,’ 21 The Journal of Legal Studies 115–57.
Bernstein, Lisa. 1999. ‘The Questionable Empirical Basis of Article 2’s Incorporation Strategy: A Preliminary
Study,’ 66 The University of Chicago Law Review 710–80.
Bernstein, Lisa. 2001. ‘Private Commercial Law in the Cotton Industry: Creating Cooperation through Rules,
Norms, and Institutions,’ 99 Michigan Law Review 1724–90.
Blackstone, William. 1765. Commentaries on the Laws of England in Four Books, vol. I. Oxford: Clarendon Press.
Browne, J.H. Balfour. 1875. The Law of Usages and Customs: A Practical Law Tract. London: Stevens & Haynes,
Law Publishers.
Carter, Stephen L. 1992. ‘Custom, Adjudication, and Petrushevsky’s Watch: Some Notes from the Intellectual
Property Front,’ 78 Virginia Law Review 129–40.
Center for Social Media, American University. 2005. ‘Documentary Filmmakers’ Statement of Best Practices
in Fair Use,’ accessed at http://archive.cmsimpact.org/sites/default/files/fair_use_final.pdf (last visited April
7, 2019).
Center for Social Media, American University. 2008. ‘Recut, Reframe, Recycle: Quoting Copyrighted Material in
User-Generated Video,’ accessed at http://cmsimpact.org/wp-content/uploads/2016/04/CSM_Recut_Reframe_
Recycle_report.pdf (last visited April 7, 2019).
Center for Social Media, American University. 2009. ‘Code of Best Practices in Fair Use for OpenCourseWare,’
accessed at http://cmsimpact.org/wp-content/uploads/2016/01/10-305-OCW-Oct29.pdf (last visited April 7,
2019).
Creative Commons. 2015. ‘State of the Commons,’ accessed at https://stateof.creativecommons.org/2015/ (last
visited April 7, 2019).
Creative Commons Brief as Amici Curiae Supporting Plaintiff-Appellant, Jacobsen v. Katzer, 535 F.3d 1373
(Fed. Cir. 2008) (No. 2008-1001) (filed Dec. 28, 2007).
Darling, Kate, and Aaron Perzanowski. 2017. Creativity Without Law: Challenging the Assumptions of
Intellectual Property. New York: New York University Press.
De Wolf, Richard C. 1925. An Outline of Copyright Law. Boston: John W. Luce & Company.
Dep’t of Commerce Internet Policy Task Force. 2013. ‘Copyright Policy, Creativity, and Innovation in the
Digital Economy,’ accessed at https://www.uspto.gov/sites/default/files/news/publications/copyrightgreenpa-
per.pdf (last visited April 7, 2019).
Dep’t of Commerce Internet Policy Task Force. 2014. ‘Reporter’s Transcript of Proceedings, Dep’t of Commerce
Internet Policy Task Force Green Paper Roundtable on Copyright Policy, Creativity, and Innovation in the
Digital Economy,’ accessed at https://www.uspto.gov/sites/default/files/ip/global/copyrights/la_transcript.pdf
(last visited April 7, 2019).
Ellickson, Robert C. 1991. Order Without Law: How Neighbors Settle Disputes. Cambridge, MA: Harvard
University Press.
Epstein, Richard A. 1992a. ‘The Path to The T. J. Hooper: The Theory and History of Custom in the Law of
Tort,’ 21 The Journal of Legal Studies 1–38.
Epstein, Richard A. 1992b. ‘International News Service v. Associated Press: Custom and Law as Sources of
Property Rights in News,’ 78 Virginia Law Review 85–128.
Epstein, Richard A. 1999. ‘Confusion about Custom: Disentangling Informal Customs from Standard
Contractual Provisions,’ 66 The University of Chicago Law Review 821–35.
Epstein, Richard A. 2008. ‘Some Reflections on Custom in the IP Universe,’ 93 Virginia Law Review In Brief
223–30.
Fagundes, David. 2012. ‘Talk Derby to Me: Intellectual Property Norms Governing Roller Derby Pseudonyms,’
Farrell, Joseph, and Robert P. Merges. 2004. ‘Incentives to Challenge and Defend Patents: Why Litigation Won’t
Reliably Fix Patent Office Errors and Why Administrative Patent Review Might Help,’ 19 Berkeley Technology
Fisher, William W., and William McGeveran. 2006. ‘The Digital Learning Challenge: Obstacles to Educational

Uses of Copyrighted Material in the Digital Age,’ The Berkman Center for Internet & Society at Harvard Law
School, https://papers.ssrn.com/sol3/papers.cfm?abstract_id=923465 (last visited April 8, 2019).
Gibson, James. 2007. ‘Risk Aversion and Rights Accretion in Intellectual Property Law,’ 116 The Yale Law
Journal 882–951.
Heins, Marjorie, and Tricia Beckles. 2005. ‘Will Fair Use Survive? Free Expression in the Age of Copyright
Control,’ Brennan Center for Justice, NYU School of Law, accessed at https://www.aaup.org/sites/default/
files/files/WillFairUseSurvive.pdf (last visited April 7, 2019).
Landes, William M., and Richard A. Posner. 1987. The Economic Structure of Tort Law. Cambridge, MA:
Lawson, John D. 1881. The Law of Usages and Customs, with Illustrative Cases. St. Louis: F.H. Thomas &
Company.
Lemley, Mark A. 1998. ‘The Law and Economics of Internet Norms,’ 73 Chicago-Kent Law Review 1257–94.
1991–2049.
Leval, Pierre N. 1990. ‘Toward a Fair Use Standard,’ 103 Harvard Law Review 1105–136.
Loren, Lydia Pallas. 1997. ‘Redefining the Market Failure Approach to Fair Use in an Era of Copyright
Permission Systems,’ 5 Journal of Intellectual Property Law 1–58.
Madison, Michael J., Brett M. Frischmann, and Katherine J. Strandburg. 2010. ‘Constructing Commons in the
Manly, Lorne. 2006. ‘Arts Briefly: Times Hires Talent Agency,’ The New York Times (Aug. 17, 2016), accessed
at https://www.nytimes.com/2006/08/17/arts/arts-briefly-times-hires-talent-agency.html (last visited April 7,
2019).
Max, D.T. 2006. ‘The Injustice Collector: Is James Joyce’s Grandson Suppressing Scholarship?,’ The New Yorker
(Jun. 19, 2006), accessed at https://www.newyorker.com/magazine/2006/06/19/the-injustice-collector (last
visited April 7, 2019).
Max, D.T. 2007. ‘Final Destination: Why Do the Archives of So Many Great Writers End Up in Texas?,’ The
New Yorker (Jun. 11, 2007), accessed at https://www.newyorker.com/magazine/2007/06/11/final-destination
(last visited April 7, 2019).
Morris, Clarence. 1942. ‘Custom and Negligence,’ 42 Columbia Law Review 1147–68.
New York University. 1983. ‘University Policy on Photocopying Copyrighted Materials, accessed at https://
www.nyu.edu/content/dam/nyu/provost/documents/Archived%20Policies/University%20Policy%20on%20Pho
tocopying%20Copyrighted%20Materials%20May%201983%20-%20ARCHIVED.pdf (last visited April 7,
2019).
New York University. 2014. ‘Statement of Policy and Guidelines on Educational and Research Uses of
Copyrighted Materials,’ accessed at http://www.nyu.edu/content/dam/nyu/compliance/documents/Copyrighted
Materials.1.6.14.pdf (last visited April 7, 2019).
Oliar, Dotan, and Christopher Sprigman. 2008. ‘There’s No Free Laugh (Anymore): The Emergence of
1787–867.
Ostrom, Elinor. 1990. Governing the Commons: The Evolution of Institutions for Collective Action. Cambridge:
Posner, Eric A. 1996. ‘Law, Economics, and Inefficient Norms,’ 144 University of Pennsylvania Law Review
1697–744.
Postema, Gerald J. 1986. Bentham and the Common Law Tradition, 2004, repr. Oxford: Clarendon Press.
Rose, Carol. 1986. ‘The Comedy of the Commons: Custom, Commerce, and Inherently Public Property,’ 53 The
Rose, Carol M. 2000a. ‘Left Brain, Right Brain and History in the New Law and Economics of Property,’ 79
Oregon Law Review 479–92.
Rose, Carol M. 2000b. ‘Property and Expropriation: Themes and Variations in American Law,’ 2000 Utah Law
Review 1–38.
Rothman, Jennifer E. 2007. ‘The Questionable Use of Custom in Intellectual Property,’ 93 Virginia Law Review
1899–982.
Rothman, Jennifer E. 2010a. ‘Best Intentions: Reconsidering Best Practices Statements in the Context of Fair
Use and Copyright Law,’ 57 Journal of the Copyright Society of the U.S.A. 371–87.
Rothman, Jennifer E. 2010b. ‘Liberating Copyright: Thinking Beyond Free Speech,’ 95 Cornell Law Review
463–534.
Rothman, Jennifer E. 2013. ‘Copyright, Custom, and Lessons from the Common Law,’ in Shyamkrishna
Balganesh, ed., Intellectual Property and the Common Law. New York: Cambridge University Press.

Rothman, Jennifer E. 2014. ‘Copyright’s Private Ordering and the “Next Great Copyright Act,”’ 29 Berkeley
Smith, Henry E. 2009. ‘Community and Custom in Property,’ 10 Theoretical Inquiries in Law 5–41.
Thompson, E.P. 1991. Customs in Common. London: Penguin Books.
Tushnet, Rebecca. 2007. ‘Payment in Credit: Copyright Law and Subcultural Creativity,’ Spring 70 Law and
University of Minnesota. 2017. ‘University of Minnesota Instructors Rights & Responsibilities,’ accessed at
https://www.lib.umn.edu/copyright/university-minnesota-instructors/ (last visited April 7, 2019).
Weinreb, Lloyd L. 1990. ‘Fair’s Fair: A Comment on the Fair Use Doctrine,’ 103 Harvard Law Review 1137–61.
Weinreb, Lloyd L. 1992. ‘Custom, Law, and Public Policy: The INS Case as an Example for Intellectual
Property,’ 78 Virginia Law Review 141–7.
17 U.S.C. § 107 (2015).

H.R. Rep. No. 90-83 (1967).
H.R. Rep. No. 94-1476 (1976).
Cases
Am. Geophysical Union v. Texaco, Inc., 60 F.3d 913 (2d Cir. 1994).
Basic Books v. Kinko’s Graphics, 758 F. Supp. 1522 (S.D.N.Y. 1991).
Bridgeport Music v. Dimension Films, 410 F.3d 792 (6th Cir. 2005).
Cambridge Univ. Press v. Patton, 769 F.3d 1232 (11th Cir. 2014).
Comput. Assocs. Int’l v. Altai, Inc., 982 F.2d 693 (2d Cir. 1992).
Dellar v. Samuel Goldwyn, Inc., 104 F.2d 661 (2d Cir. 1939).
Dun & Bradstreet Software Servs. v. Grace Consulting, 307 F.3d 197 (3d Cir. 2002).
Forasté v. Brown Univ., 290 F. Supp. 2d 234 (D.R.I. 2003).
Harper & Row, Publishers, Inc. v. Nation Enters., 471 U.S. 539 (1985).
Hays v. Sony Corp. of Am., 847 F.2d 412 (7th Cir. 1988).
Jim Henson Prods. v. John T. Brady & Assocs., 16 F. Supp. 2d 259 (S.D.N.Y. 1997).
May v. Morganelli-Heumann & Assocs., 618 F.2d 1363 (9th Cir. 1980).
New Era Publ’ns Int’l, ApS v. Carol Publ’g Grp., 904 F.2d 152 (2d Cir. 1990).
Pittsburgh State Univ. v. Kansas Bd. of Regents, 122 P.3d 336 (Kan. 2005).
Princeton Univ. Press v. Mich. Document Servs., 99 F.3d 1391 (6th Cir. 1996).
Ringgold v. Black Entm’t Television, 126 F.3d 70 (2d Cir. 1997).
Rodi Yachts v. Nat’l Marine, 984 F.2d 880 (7th Cir. 1993).
Rosemont Enters. v. Random House, Inc., 366 F.2d 303 (2d Cir. 1966).
Roy Exp. Co. Establishment of Vaduz v. CBS, 503 F. Supp. 1137 (S.D.N.Y. 1980).
Roy Exp. Co. Establishment of Vaduz v. CBS, 672 F.2d 1095 (2d Cir. 1982).
Shapiro, Bernstein & Co. v. P.F. Collier & Son Co., 26 U.S.P.Q. 40 (S.D.N.Y. 1934).
Soc’y of the Holy Transfiguration Monastery v. Gregory, 689 F.3d 29 (1st Cir. 2012).
State ex. rel. Thornton v. Hay, 462 P.2d 671 (Or. 1969).
The T.J. Hooper, 60 F.2d 737 (2d Cir. 1932).
Twentieth Century Fox Film v. Marvel Enters., 155 F. Supp. 2d 1 (S.D.N.Y. 2001).
Weinstein v. Univ. of Illinois, 811 F.2d 1091 (7th Cir. 1987).

21. Infrastructure theory and IP
Brett Frischmann*
Contents
I. Introduction
II. An Overview of Infrastructure Theory
A. A Shift in How We Think About Infrastructure
B. Characteristics of Infrastructural Resources
C. Demand-Side Market Failures that Infrastructural Resources Help
Resolve and Why They Arise
D. Why Commons Management of Infrastructure Resources can Help Avert
Market Failures
III. Applying Infrastructure Theory to Intellectual-Cultural Resources
A. The Cultural Environment as Infrastructure
B. Economic Characteristics of Intellectual Resources
1. Conventional supply-side problems
2. Intellectual resources and activities, products, and processes
C. Intellectual Property Laws as Semi-Commons Arrangements
IV. Conclusion
References
I. INTRODUCTION
This chapter explores how infrastructure theory applies to cultural-intellectual resources.

It begins with a summary of infrastructure theory and then discusses how the theory
applies to intellectual-cultural resources. Applying the theory reveals a series of demand-
side complications for conventional law and economic theories of intellectual property
and related governance institutions. These complications arise vividly in modern debates
about intellectual property rules that exclude certain subject matter or otherwise limit the
scope of intellectual property rights and sustain commons (public access).
The chapter is an adaptation of ‘Intellectual Infrastructure,’ chapter 12 of Frischmann
(2012). Accordingly, it summarizes the theory developed in Frischmann (2012) and does
not provide as comprehensive an intellectual history as other chapters or the original
source, which has an extensive bibliography covering various fields of economics, law, and
innovations studies.1 Benkler (2013) describes many of the important works in intellectual
* Charles Widger Endowed University Professor in Law, Business and Economics Charles
Widger School of Law, Villanova University.
1
I included many of the sources that bear on intellectual infrastructure and intellectual prop-
erty in the reference list at the end of this chapter.
551

property that strongly influenced infrastructure theory. Besides the intellectual property
laws themselves, this chapter does not examine various infrastructures that support the
production and distribution of intellectual resources.
II. AN OVERVIEW OF INFRASTRUCTURE THEORY
A. A Shift in How We Think About Infrastructure
‘Infrastructure’ generally conjures up the notion of a large-scale physical resource made

by humans for public consumption. Standard definitions of infrastructure refer to the
underlying framework or foundation of a system. Familiar examples of ‘traditional infra-
structure’ include (1) transportation systems, such as highway systems, railway systems,
airline systems, and ports; (2) communication systems, such as telephone networks and
postal services; (3) governance systems, such as court systems; and (4) basic public services
and facilities, such as schools, sewers, and water systems. The list could be expanded con-
siderably; daily, we rely on a wide range of traditional infrastructure resources. Following
the suggestion of the National Research Council (1987), Frischmann (2012) argues a
more capacious understanding of infrastructure is warranted for functional economic
analysis. In particular, recognizing that infrastructural resources are shared means to
many ends expands the definitional scope beyond traditional infrastructure categories.
A host of ‘nontraditional infrastructure’ resources exist and also serve as the underlying
frameworks or foundations of various economic and social systems. Examples include
environmental infrastructure, such as oceans and the atmosphere, as well as intellectual
infrastructure, such as basic research, ideas, and languages. Both traditional and nontra-
ditional infrastructures function as infrastructural capital (Button, 1996; Frischmann,
2005; Frischmann and Lemley, 2007; Frischmann, 2012; Benkler, 2013; on different forms
of capital, see Bourdieu, 1986).
In various fields of economics, there is widespread recognition that infrastructure is
special. Macroeconomics recognizes that infrastructure is important for economic devel-
opment and a key ingredient for economic growth (Aschauer, 2000; Ghosh and Meagher,
2004). Microeconomics recognizes that infrastructural resources often generate substantial
social gains and that markets for infrastructure often fail and call for government inter-
vention in one form or another (Branscomb and Keller, 1996; David and Foray, 1996;
Justman and Teubal, 1996; Kahn, 1988). Increasingly, there is recognition that market
failures arise on both the supply and demand sides and demand-side failures also create
obstacles for government provisioning (Frischmann, 2012; Frischmann and Hogendorn,
2015). Development economics, which applies both micro- and macroeconomics to the
process of development, recognizes similar ideas. Yet there are some gaping holes in our
understanding of how infrastructural resources generate substantial social gains and
contribute to development and economic growth. In particular, the importance of open
infrastructure and user-generated spillovers at the microeconomic level and their relation-
ships to technological innovation and growth at the macroeconomic level deserve attention
(Tassey, 1996; Helpman, 1998; see also Romer, 1986, 1996; Jones, 1998; Schmidt, 2003).
Legal scholar Carol Rose (1986) recognized the importance of these issues in her
seminal article, ‘The Comedy of the Commons: Custom, Commerce, and Inherently

Infrastructure theory and IP 553
Public Property.’ She explained that a ‘comedy of the commons’ arises where open access
to an infrastructure resource leads to scale returns—greater social value with greater use
of the resource. For example, managing road systems in an open, nondiscriminatory
manner is the key to sustaining and increasing participation in commerce, and commerce
is itself a productive activity that generates significant positive externalities. Commerce
generates private value that is easily observed and captured by participants in economic
transactions, as buyers and sellers exchange goods and services, but it also generates social
value that is not easily observed and captured by participants—value associated with
socialization, cultural exchange, trust, and other such processes.
Following Rose, intellectual property and communications scholars extended her
ideas to information. Yochai Benkler (2013) describes the intellectual history, tracing
the important contributions of Fisher (1988), Litman (1990), Samuelson (1990), Boyle
(1996), Cohen (1998), Benkler (1998a, 1998b, and 1999), Lessig (2001), and Lemley
(2005), among many others.2 These and other scholars recognized that intellectual and
cultural resources of various kinds were special in a manner similar to infrastructure
(without necessarily making the comparison explicit) because the resources are strictly
nonrivalrous in consumption and often inputs into further production of such resources
(Nelson, 1959; Arrow, 1962; Merges and Nelson, 1990; Scotchmer, 1991). Section II
explores this connection more fully. Benkler (2013) also situates this line of scholarship
in the broader and in some respects different literature on commons (Hess and Ostrom,
2003, 2007; Ostrom, 1990, 2005, 2007; Frischmann et al., 2014). This chapter does not
cover the commons literature.
B. Characteristics of Infrastructural Resources
Infrastructural resources are intermediate capital resources that often serve as critical
foundations for productive behavior within economic and social systems. Infrastructural
resources effectively structure in-system behavior at the micro level by providing and shap-
ing the available opportunities of many actors. In some cases, infrastructural resources
make possible what would otherwise be impossible, and in other cases, infrastructural
resources reduce the costs and/or increase the scope of participation for actions that are
otherwise possible. In addition, infrastructural resources structure in-system behavior in
a manner that leads to spillovers. That is, infrastructural resources facilitate productive
behaviors by users that affect third parties, including other users and even non-users of the
infrastructure. The third party effects often are accidental, incidental, and not especially
relevant to the infrastructure provider or user. To put it another way, the social returns on
infrastructure investment and use may exceed the private returns because society realizes
benefits above and beyond those realized by providers and users (Steinmueller, 1996;
Frischmann and Lemley, 2007).
According to Frischmann (2012),
Infrastructural resources satisfy the following criteria:
(1) The resource may be consumed nonrivalrously for some appreciable range of demand.
2
Benkler (2013) discusses other works as well. I have highlighted those that relate most
closely to the infrastructure theory.

(2) Social demand for the resource is driven primarily by downstream productive activities that
require the resource as an input.
(3) The resource may be used as an input into a wide range of goods and services, which may
include private goods, public goods, and social goods.
The first criterion captures the consumption attribute of public and impure public
goods (Samuelson, 1954; Musgrave and Musgrave, 1984; Cornes and Sandler, 1996).
In short, this characteristic describes the ‘sharable’ nature of infrastructural resources.
Infrastructural resources are sharable in the sense that the resources can be accessed and
used by multiple users at the same time. Infrastructural resources vary in their capacity
to accommodate multiple users, and this variance in capacity differentiates purely non-
rival (infinite capacity) resources from partially nonrival (finite but renewable capacity)
resources (Frischmann, 2012; Benkler, 2013 (‘nonscarce over substantial ranges of their
use’ instead of partially (non)rival)). Nonrivalry opens the door to widespread access and
productive use of the resource. For nonrival resources of infinite capacity, the marginal
cost of allowing an additional person to access the resource is zero (Cornes and Sandler,
1996). For partially nonrival resources of finite capacity, the cost-benefit analysis is
more complicated because of the possibility of congestion (Cornes and Sandler, 1996;
Frischmann, 2012; Benkler, 2013).
The second and third criteria focus on the manner in which infrastructural resources
create social value. The second criterion emphasizes that infrastructural resources are
capital goods that create social value when utilized productively downstream and that
such use is the primary source of social benefits. Thus, societal demand for infrastructure
is derived demand. The third criterion emphasizes both the variance of downstream
outputs (the genericness of the input) and the nature of those outputs, particularly public
goods and social goods.3
C. Demand-Side Market Failures that Infrastructural Resources Help Resolve and Why
They Arise
Infrastructural resources enable many (market and nonmarket) systems to function and
satisfy demand derived from many different types of users. Infrastructural resources are
not special-purpose resources, optimized for a particular user or use to satisfy the demand
derived from a particular downstream market. Instead, and importantly, they provide
basic, multipurpose functionality.
An electricity grid, for example, delivers power to the public, supporting an incred-
ibly wide range of uses, users, markets, and technologies. It is not specially designed
or optimized for any particular use, user, market, technology, or appliance; it provides
nondiscriminatory service for a toaster and a computer, for Staples and a pizzeria, and so
on. The same could be said of the national highway system, the Internet, contract law, or
3
‘Social goods generate value through their impact on social systems and interdependencies
. . ..Their value is inherently social because it depends on the existence and nature of interdepend-
ent social relations. . ..Social goods change environmental conditions and social interdependencies
in ways that affect social welfare’ Frischmann, 2012, ch. 3 (providing a detailed examination of
various social goods, including nonmarket goods, merit goods, social capital, and irreducibly social
goods).

Infrastructure
User Activities
within economic, cultural, political, and other systems
User Generated Outputs

Private Goods Public Goods Social Goods
Output Markets
Spillovers
Figure 21.1 Infrastructure
even a language. These resources provide basic functionalities and thereby support and
structure more complex systems of user activities, but the resources (or the owners and
managers of such resources) do not determine what users do. See Figure 21.1.
Users determine what to do with the capabilities that infrastructure provides.
Genericness implies a range of capabilities, options, opportunities, choices, freedoms.
Subject to standardized compatibility requirements, users decide what to plug in, run,
use, work with, play with. Users decide which roads to travel, where to go, what to do,
who to visit. Users choose their activities; they can choose to experiment, to innovate, to
roam freely. Users decide whether and what to build. Users decide how to use their time
and other complementary resources.
To understand societal demand and how value is created and realized, it is necessary
to pay closer attention to the nature of the user activities and the outputs users produce;
the value of an infrastructural resource is realized by producers and consumers of these
outputs. It is thus the demand for these outputs that determines demand for the infra-
structure. For this reason, the third criterion emphasizes the nature of the downstream
outputs and, more specifically, the potential production of public and social goods.
When infrastructure supports these productive activities, markets will not work well
in assessing demand and supplying infrastructural resources; a gap between private
demand and social demand arises because the social value created by allowing users to
access and use the resource to produce public and social goods is underrepresented in
the prices people are willing to pay for infrastructure (Musgrave and Musgrave, 1984).
The social value may be substantial but extremely difficult to measure. The reason for
the gap is relatively straightforward: Infrastructure users’ willingness to pay reflects
private demand—the value that they expect to realize—and does not take into account
value that others might realize as a result of their use. That is, it does not account for
external effects associated with the production of public and social goods. This means
that infrastructure users who produce public and social goods are not necessarily optimal
purchasers of access and use rights, because they do not themselves capture the full social
value of their use.

Special purpose input? Commercial infrastructure?
Infrastructure
Infra
User Activities
within economic, cultur
cultural, political, and other systems
User Gen
Generated Outputs
P
Private Goods Pu
Public Goods Social Goods
Output Markets
Spillovers
Figure 21.2 Infrastructure with optimization
Not only is there a gap between private and social demand, but to make matters even
more complex, the shape of the social demand curve can be quite different from that
of the private demand curve due to the variance in producers and outputs. This added
complexity gives rise to specific and substantial demand-side market failures that have
static and dynamic consequences; however, they are beyond the scope of this chapter
(Frischmann, 2012).
Difficulties in measuring and appropriating value generated in output markets
translates into a problem for infrastructure suppliers and, consequently, the public. This
‘demand-manifestation’ problem may affect infrastructure allocation, design, investment,
and management, as well as other supply-side decisions. At least in market contexts, infra-
structure suppliers base such decisions in large part on the prospect of foreseeable returns
in downstream markets. Demand signals manifested in those markets, aggregated and, in
a sense, communicated upstream by dynamic operation of the price mechanism, provide
critical raw information for making assessments about prospective returns. To society’s
detriment, demand-manifestation problems can lead to the undersupply of infrastructure
essential to various producers of public and social goods, and this undersupply can lead to
an optimization of infrastructure design or prioritization of access and use of the infra-
structure for a narrower range of uses than would be socially optimal. See Figure 21.2.
D. Why Commons Management of Infrastructure Resources can Help Avert Market

Failures
In this context, commons management is a mode of governance for infrastructure that

can be especially attractive from an economic and social perspective. The core principle
of commons management is sustainable sharing on nondiscriminatory terms within a

defined community (Frischmann et al., 2014). With respect to infrastructure,4 this means
nondiscrimination among users and uses, and the corresponding preservation of equality,
flexibility, and general purposiveness. From an economic perspective, it makes sense to
manage certain infrastructural resources in an openly accessible manner because doing
so permits a wide range of downstream producers of private, public, and social goods to
flourish. As Benkler (2001) noted, ‘[t]he high variability in value of using both transporta-
tion and communications facilities from person to person and time to time have made a
commons-based approach to providing the core facilities immensely valuable.’ The point
is not that all infrastructural resources (traditional or nontraditional) should be managed
in an openly accessible manner. Rather, for certain classes of resources, the economic
arguments for managing the resources in an openly accessible manner vary in strength
and substance (Frischmann, 2012; Benkler, 2013).
By precluding optimization and prioritization based on market demands alone,
commons management sustains the social option value of the infrastructure; it retains
‘flexibility’ (Benkler, 2013) or ‘breathing room’ (Cohen, 2012) in the face of uncertainty.5
But commons management is not only a buffer from market pressures. It also serves as
a buffer from government pressures to optimize or prioritize infrastructure (use), which
may originate from market actors lobbying for government interventions or from other
pressures to regulate. Finally, managing public infrastructure as commons also reduces
reliance on government to pick winners and direct subsidies to users who produce public
and social goods.6 In the end, as shown in Figure 21.3, the basic idea is that commons
management leverages nonrivalry and the sharable nature of the infrastructure (first
criterion) to sustain the general-purpose nature of the infrastructure and support the
wide range of user activities that generate private, public, and social goods (second and
third criteria).
III. APPLYING INFRASTRUCTURE THEORY TO

INTELLECTUAL-CULTURAL RESOURCES
Applying the infrastructure criteria to intellectual-cultural resources delineates a broad

set of resources that create benefits for society primarily through the facilitation of
downstream productive activities, many of which generate spillovers. The definition of
infrastructure can be reduced as follows to fit intellectual infrastructure: nonrival input
into a wide variety of outputs. This seems incredibly capacious. Like many other types of
4
The defined community may be the public, in which case commons management converges with
open access. The community might also be more limited in scope, in which case there may be members
and nonmembers (Ostrom, 1990, 2005; Hess and Ostrom, 2007; Frischmann, 2012; Frischmann et al.,
2014).
5
Where future sources of social value are uncertain, there are good reasons to sustain infra-
structure as commons. See Frischmann, 2012, ch. 5.
6
In other words, there is less reliance on government grants and subsidies. The government also
faces demand-manifestation problems because it does not know the shape of the demand curve
and may struggle to identify which public or social goods to subsidize and which producers to fund.
Government is often better off supporting open infrastructure as a means for supporting the capabili-
ties of productive users. See id.

Infrastructural Resources
1. The resource may be consumed nonrivalrously;
COMMONS MANAGEMENT
le sta
ve in
su
ra in
2. social demand for the resource is driven
gi g
ng
primarily by downstream productive activity that
requires the resource as an input; and
3. the resource is used as an input into a wide range

of goods and services, including private goods,
public goods and/or social goods.
Figure 21.3 Infrastructure with commons management interface
infrastructure, intellectual infrastructure can be identified and analysed at various levels

of abstraction, ranging from the meta-environment itself to a discrete general-purpose
input, such as a basic idea, to a specific expression that has broad communicative power
and social meaning. The resource set is much more highly and diversely populated
than traditional infrastructure. Each of the following categories, for example, contains
innumerable examples:
● Basic research
● Ideas
● General-purpose technologies
● Languages
These categorical examples of intellectual infrastructure create benefits for society pri-
marily by facilitating a wide range of downstream productive activities, including infor-
mation production, innovation, and the development of products and services, as well as
education, community building and interaction, democratic participation, socialization,
and many other socially valuable activities. The foundational role of these intellectual
infrastructures in cumulative, dynamic, and complex systems merits attention. Courts and
commentators frequently refer to intellectual infrastructure resources as ‘building blocks’
to capture their role as basic inputs. The ‘building blocks’ metaphor is evocative, but it
oversimplifies and fails to fully reflect the complex relationships among participants in
intellectual systems that derive value from intellectual infrastructures as producers, users,
consumers, or incidental beneficiaries.
Applying infrastructure concepts to cultural-intellectual resources is more difficult in
some respects than applying them to other resources. The difficulties stem in part from the
fluid, continuous, and dynamic nature of cultural-intellectual systems. Distorting reduc-
tionism seems inevitable when we attempt to delineate clear boundaries around discrete
cultural-intellectual resources or to separate resources (inputs, outputs, products, things)
from activities (processes, practices, uses) (Madison, 2005). Intellectual infrastructures,
such as basic research, often seem to be both resources and activities.

Difficulties also stem from the fact that infrastructure appear to exist on many different
scales within cultural-intellectual systems.7 The infrastructure concept has a fractal nature
when applied to intellectual resources because one can identify infrastructure at various
scales and observe repeating patterns, similar characteristics, and governance dilemmas.
The dynamic, cumulative, and interactive features of cultural-intellectual resources and
practices may mean that more cultural-intellectual resources potentially function as
infrastructure. Keep in mind that the infrastructure criteria describe functional economic
relationships and are only part of the broader resource management question. That
the functional economic relationships repeat at different scales means that choosing the
appropriate scale for evaluating resource management options and trade-offs is critical.
Different types of governance rules may be appropriate at different scales; for example, a
default open-access-style rule (ex ante, broadly applicable) may be appropriate for intel-
lectual infrastructure at high levels of abstraction, while a less demanding rule, such as
an essential-facilities-style rule (ex post, case-specific), may be appropriate for intellectual
infrastructure at lower levels of abstraction.
This Section is organized into three subsections.8 Subsection A begins with the idea
of the cultural environment as infrastructure. This discussion establishes a context for
examining intellectual-cultural resource systems and governance institutions. It also
explains some of the complex relationships between intellectual-cultural resources and
people—for example, the mutual dynamic shaping that takes place as society lives within,
interacts with, shapes, and is shaped by its interactions with the cultural environment.
Subsection B describes the economic characteristics of intellectual resources. The eco-
nomic analysis of intellectual resources is quite complex, and other chapters discuss most
of the general issues, particularly as they relate to the institutions of intellectual property.
Accordingly, this subsection only highlights a few specific concerns. First, it explains how
(non)excludability and nonrivalry give rise to distinct economic considerations concerning
systematic risk of undersupply of some intellectual resources. Next, it considers the added
layer of complexity associated with ‘information [being] both input and output of its own
production process. . ..known to economists as the “on the shoulders of giants” effect’
(Benkler, 2006, p. 37; Scotchmer, 1991). The added complexity centers on the dynamics of
intellectual processes and systems and how intellectual progress occurs. An appreciation
of the ‘on the shoulders of giants’ effect is critical to understanding how productive use
of intellectual infrastructure generates spillovers (Benkler, 2006; Scotchmer, 1991).
Subsection C considers intellectual property laws. It examines intellectual property
laws as a semi-commons regime (Smith, 2000; Heverly, 2003; Loren, 2007). It shows first
how intellectual property laws are targeted exceptions/interventions to the default com-
mons regime for the cultural environment. The laws enclose and regulate a select (albeit
very broad) set of intellectual resources to overcome supply-side market failures. Given
7
Of course, these difficulties arise for most theories.
8
Again, this chapter is an adaptation of chapter 12 of Frischmann (2012). In addition to
abbreviating much of the discussion, it leaves out the entire section that delineated intellectual infra-
structure, focused on ideas as an example, and examined how the First Amendment, copyright, and
patent laws govern ideas as commons. In addition to the need for brevity, the reason for excluding
this material is that it would have required an extensive discussion of recent developments in patent
law regarding patentable subject matter, and that would have been beyond the scope of this chapter.

tremendous difficulties in establishing and maintaining boundaries around this set and
the dynamic and complex nature of cultural-intellectual resource systems, the laws also
sustain semi-commons arrangements for enclosed resources. This leads to a second point.
Many intellectual infrastructure resources are excluded from the enclosed set—that is, the
resources remain in the public domain and are not patentable or copyrightable, but many
other intellectual infrastructures are patentable or copyrightable. For these resources,
intellectual property laws seem to recognize the social demand for commons management
and mediate access to intellectual infrastructure (Lee, 2008).
A. The Cultural Environment as Infrastructure
We can identify infrastructure at a very high level of abstraction and broad scale—the
cultural environment—and at progressively lower levels of abstraction, on smaller scales,
and within different systems or fields. In parallel with such efforts, we can study different
commons management institutions. Thus, we analyse both infrastructure and commons
as ‘nested phenomena’ operating at different levels that may interact with one another
(Madison et al., 2010, p. 674; Ostrom, 2005, pp. 58–62).
At a relatively abstract level, the basic similarities between the natural and cultural
environments concern the functional and relational meanings of the common term ‘envi-
ronment.’ An environment might be defined as a complex system of interconnected and/
or interdependent resources (or even resource systems) that comprise the ‘surroundings,’
‘setting,’ or ‘context’ that we inherit, live within, use, interact with, change, and pass on to
future generations. We inherit the natural physical environment; we live within, use, inter-
act with, and change it; and we pass it on to future generations. Similarly, we inherit, live
within, use, interact with, change, and pass on to future generations a cultural-intellectual
environment, comprised of many overlapping sub-environments, if one would like to
distinguish culture(s), science(s), and so on. The world we live in comprises multiple,
complex, overlapping, and interdependent resource systems with which we interact and
that constitute our environments—the natural environment is one type, and (socially)
constructed environments, such as the cultural environment, are another.
Thus, we can envision a cultural environment that consists of the various cultural,
intellectual, and social resource systems that we inherit, use, experience, interact with,
change, and pass on to future generations (Boyle, 2003, 1997, 1996; Opderbeck, 2009).
The cultural environment provides us with resources and capabilities to act, participate,
be productive, and ‘make and pursue life plans that can properly be called our own’
(Benkler, 2006, p. 146). It also shapes our very beliefs and preferences regarding our lives
(life plans) and relationships with each other and the world we share (DeMartino, 2000,
pp. 77–9; Krackhardt, 1998; North, 2005, pp. 46–7). Human beings are not born with fully
formed preferences, knowledge, and beliefs about the world they enter; rather, preferences,
knowledge, and beliefs are learned and experienced and thus contingent to a degree on the
cultural environment a person experiences (Cohen, 2001, 2007, 2012).
We have an incredibly complex and dynamic relationship with the cultural environment.
Science and culture, for example, are cumulative and immersive systems that develop with
society, while simultaneously developing society. Put another way, the cultural environ-
ment provides for, shapes, and reflects us, and at the same time, we provide, shape, and
reflect it. The cultural environment has a normative dimension that is sometimes lost

in the sense that it can be understood as a society’s answer to a series of ‘fundamental

questions’ about what it values.
‘Cultural’ captures the contextual, contingent, and social/relational aspects of the
resources that constitute the meta-environment; the resources are resources vis-à-vis their
meaning to and among people. As Benkler (2006, p. 282) suggests, ‘[Culture] is a frame
of meaning from within which we must inevitably function and speak to each other, and
whose terms, constraints, and affordances we always negotiate. There is no point outside
of culture from which to do otherwise.’
Viewed through the infrastructure lens, two points are clear: First, the cultural
environment constitutes mixed infrastructure. Human beings produce an incredible
diversity of private, public, and social goods simply by living in and interacting with the
cultural environment. It would be a mistake to assume that such productivity is inevi-
table or natural. In some respects, some may be; human beings experience their lives,
and ‘[e]xperience constitutes an important intellectual resource that simultaneously
relates human beings to their inherited and evolving environment(s) and constitutes
a resource that may shape the intellectual environment’ (Madison et al., 2010, p. 685).
But the fact that the cultural environment shapes both our capabilities to be culturally
or intellectually productive and our beliefs and preferences regarding the exercise of
such capabilities is a reason to pay close attention to the dynamic relationships between
users and the cultural environment (Sen, 1993, 2009; Frischmann, 2017). Whether
people are active participants in intellectual, cultural, and social resource production,
actively shaping the cultural environment, or passive consumers shaped by the cultural
environment (or by those who are shaping it) depends on the cultural environment and
how it is managed (Benkler, 2006). Different evolutionary paths are possible—for our
environment and for us:
[People] may become more aware, conscious of their (potential) roles as listeners, voters, and
speakers, but also as consumers and producers, as political, cultural, and social beings, as
members of communities. They may learn to be productive—or learn to want to be productive,
if such desire is not simply latent. This very awareness that one can play different roles and
that the environment is not fixed or fully determined by others is encouraging. It encourages
participation and the development of facilitative social practices, and perhaps over time, the
adoption of a participatory culture. . ..(Frischmann, 2007a, pp. 1123–8)
Second, the case for commons management is incredibly strong. Managing the cultural
environment as a public commons maintains flexibility and maximizes the social option
value of the infrastructure. At this scale and level of abstraction, commons management
aims to limit both government and market shaping of the environment and our lives,
plans, beliefs, and preferences. In other words, commons management is a strong default
position for the cultural environment because users—autonomous individuals as well as
social groups and communities—get to shape the environment and choose what to say
and do and how to plan their lives, experiences, and interactions with each other and the
environment. The cultural environment is spillover-rich because of the many different
user activities that produce, distribute, use, and reuse public and social goods.
Support for commons as a default governance regime at this macro level seems to be
reflected in both the First Amendment, which restricts the exercise of government power
to control the cultural environment, and the related conception of a robust public domain,

which limits private ownership, control, and exclusion over swaths of cultural, intellectual,
and social resources (Litman, 1990; Benkler, 1999; Lessig, 2001; Samuelson, 2006).
Markets and government have played and will continue to play incredibly important
roles in shaping the cultural environment. These social systems and institutions depend
on and are essential to the cultural environment. As Benkler (2006) aptly describes (and
critiques), the reality of our modern existence is that the industrial information economy
and mass media system have had a tremendous influence on both the cultural environ-
ment and the American people. Yet this does not undermine the case for commons
management; government and market interventions and institutional structures should
be understood as targeted exceptions.
The cultural environment as infrastructure has an intergenerational dimension. Each
generation is blessed beyond measure with the intellectual and cultural resources it
receives from past generations; each generation experiences and changes the cultural
environment and passes it on to future generations. That we ‘stand on the shoulders
of giants’ is often noted to emphasize the cumulative nature of cultural or scientific
progress (Scotchmer, 1991, p. 29). But ‘the expression also reflects an understanding
of intergenerational dependence: each generation is both dwarf and giant; the current
generation stands on the shoulders of the past and also serves as the shoulders for
the future’ (Frischmann and McKenna, 2015; Gordon, 1993). Essentially, ‘shoulders’
refers to the ‘fundamental blessings’ of resources preserved, created, and transmitted.
While each generation faces supply-side problems (discussed below), it does so within
the existing cultural environment, while also shaping the cultural environment for the
future.
B. Economic Characteristics of Intellectual Resources
The economic characteristics of intellectual resources are complex. To start, intellectual

resources are public goods, often a form of capital, and often the source of various types
of externalities. Intellectual resources often are infrastructural and face the demand-side
concerns common to infrastructure. An added layer of complexity is the fact that intel-
lectual resources are part of cultural, intellectual, and social progress, and thus a part
of our complex and evolving cultural environment. This subsection briefly explains the
supply-side problems that flow from the public goods nature of intellectual resources;9 it
highlights how the conventional supply-side problems are complicated by demand-side
concerns. It then discusses the added layer of complexity.
1. Conventional supply-side problems

Intellectual resources face well-known supply-side problems, common to public goods,
discussed in other chapters. First, the inability to (cheaply) exclude competitors and
nonpaying consumers (free riders) presents a risk to investors perceived ex ante (prior to
production of the good), and this risk may lead to undersupply. This problem is a function
of (non)excludability. Second, even if exclusion is feasible at low cost, nonrivalry suggests
that markets still undersupply various intellectual resources.
9
Other chapters discuss these issues, but the infrastructure lens reveals nuances worth noting.

a. (Non)excludability (Non)excludability is not a fixed or inherent characteristic of

a resource; the costs of exclusion vary considerably with technology and context. In the
absence of some institutional solution, there would be a significant underinvestment in
some types of intellectual resources because of the risk that competitors would appropri-
ate the value of the resources and undermine the ability of investors to recover their costs.
Whether private incentives are in fact inefficiently suppressed by this potential misap-
propriation risk depends substantially on the type of investment, the intellectual resource
in question, and the particular context. Many intellectual resources are not subject to
this particular supply-side concern; we generate the resources without being disabled
by concerns over misappropriation. In fact, many people make investments because the
expected private benefits exceed the fixed costs, regardless of whether or not others free
ride. Appropriating benefits through market exchange of the intellectual resource or some
derivative product may not be relevant to the investor. For example, we engage in many
intellectually productive activities because participation itself provides sufficient private
benefits. Participation can be fun, intellectually stimulating, educational, or service-
oriented, among other things. Participation may not be effortless or free; it may require
substantial investment. Regardless, the private value derived from participation may be
sufficient, and external benefits conferred to others that use or consume the output (i.e.,
the intellectual resource) may be irrelevant to incentives to invest. Even if those benefits
could be internalized, such internalization could potentially decrease incentives to invest
and prove quite costly (Pink, 2010; Amabile, 1996; Benkler, 2006). In many situations,
people create, invent, and innovate because the anticipated returns from their own use
of the results are sufficient to justify the investment. Von Hippel (2005) pioneered a rich
literature on user innovation that demonstrates quite clearly how many significant innova-
tions result from users seeking to solve their own particular problems, needs, or curiosities
without disabling concern over free riding. In some contexts, people produce intellectual
resources and welcome free riding by others. Sharing intellectual resources can be a viable
strategy for increasing returns generated through other means. Benkler (2006) describes
different examples, ranging from lawyers who write articles to attract clients to software
developers who share software and make money by providing services to users. Sharing
may help attract attention, build a reputation, or lead to reciprocal sharing.
Even where free riding is a concern and appropriating benefits through market
exchange of the intellectual resource or some derivative product is relevant to investment
decisions, self-help mechanisms, such as lead-time advantages and barriers to entry, may
provide sufficient protection against free riding by competitors to support the investment.
In some instances and in certain industries, self-help mechanisms are preferred for gaining
a competitive advantage. Surveys of R&D managers show that factors such as securing
lead-time advantages, increasing learning, developing complementary products, and
ensuring secrecy are more relevant to incentives to invest in R&D than any perceived abil-
ity to secure traditional intellectual property protection (Mansfield et al., 1981, Mansfield,
1986; Levin et al., 1987; Cohen et al., 2000; Barnett, 2004). The extent of the benefits that
inure from such self-help mechanisms vary by industry and depend on a multitude of
external factors, including technology-enhanced access and copying opportunities. Such
mechanisms are imperfect and do not suffice for many types of intellectual resources, but
when relevant, they can be quite important and should be evaluated in comparison with
each other and alternative institutions.

Intellectual property laws are a prominent but by no means exclusive means of address-
ing the supply-side problem where free riding is a concern and appropriating benefits
through market exchange of the intellectual resource or some derivative product is relevant
to investment decisions. Consider patent law. In the absence of patent law, there would
be a significant underinvestment in some types of inventions because of the risk that
competitors would appropriate the value of the inventions. Granting inventors patents
lessens the costs of exclusion, raises the costs of free riding, encourages licensing, and, as
a result, makes a greater portion of the surplus generated by the invention appropriable by
the inventor. The exclusivity provided by patent law does more than affect investment in
invention, however. Patents affect the supply-side functioning of markets for inventions as
well as markets for derivative products, including additional improvements, innovations,
and commercial end-products. The reward, prospect, and commercialization theories of
patent law take patent-enabled exclusivity as the relevant means for fixing a supply-side
problem—the undersupply of private investment in the production of patentable subject
matter or in the development and commercialization of patentable subject matter that
would occur in the absence of patent-enabled exclusivity (Landes and Posner, 2003; Kitch,
1977; Kieff, 2001; Ghosh, 2004, pp. 1353–7).The theories differ largely in terms of where
in the supply chain patent-enabled exclusivity is needed, and in terms of the degree of
control/exclusivity needed. In reality, these needs vary by industry and context, giving
each of the theories some support. As the next subsection explains, intellectual property
laws also set boundaries around intellectual resources in a manner that reduces transac-
tion costs and reduces information costs. This boundary-setting function creates legal
‘things’ that can be more easily subject to market exchange (Madison, 2005).
The supply-side benefits provided by patent law are not costless. The appropriation
of a greater portion of the surplus presumes an increase in price. Absent exclusivity,
competitive distribution and use of the invention would drive the price to marginal cost
(zero), in which case consumers would capture the full consumer surplus. When relevant,
patents may enable pricing above marginal cost and, as a result, introduce deadweight
losses. Keep in mind, however, that the magnitude of the deadweight losses depends on
both the strength of the legal rights conferred and the market conditions. Other chapters
explain this further.
b. Nonrivalry Addressing the excludability problem for intellectual resources through

intellectual property or other means does not eliminate the nonrival nature of the
resources or ensure efficient market provisioning (Lunney, 2008; Samuelson, 1954). Some
scholars have suggested that private property rights convert the public good into a private
good, but this is not correct (Samuelson, 1958, p. 335). (Non)excludability should not be
confused or conflated with nonrivalry. It may be the case that exclusion can prevent shar-
ing, but that in no way affects the capacity of the resource or the corresponding option to
share among many users. Consider three important implications.10
First, nonrivalry enables sharing and an extra degree of freedom in managing or
allocating the intellectual resource. For purely consumptive ideas, this prompts the classic
10
These implications and their relationships to the infrastructure theory are why the chapter
discusses (non)excludability and nonrivaly despite the coverage given by other chapters.

trade-off between static and dynamic efficiencies—for an existing idea, open sharing
generally maximizes social welfare because the marginal cost of sharing with someone
is zero, but such sharing may have consequences for dynamic efficiency if it lessens
investment incentives. Exclusion does not eliminate this trade-off; it simply provides the
entity with the capability to exclude with the opportunity to decide whether or not to do
so. For ideas, nonrivalry prompts a more complicated trade-off among static efficiency
and various types of dynamic efficiencies. In sum, the private and public opportunity to
leverage nonrivalry remains an important economic consideration, even when the costs
of exclusion are minimal.
Second, even when the costs of exclusion are minimal, demand-measurement problems
still lead to undersupply by markets. There are two notable demand-measurement
problems, one focused on ‘optimality conditions’ and difficulties in accurately measuring
consumer preferences, and one focused on externalities. With respect to the first demand-
measurement problem, Paul Samuelson (1954) noted that a second type of free riding
occurs when consumers strategically misrepresent their true preferences in the hope that
other consumers will bear a greater proportion of the costs. This strategic behavior prob-
lem, however, is independent of exclusion. Demand-measurement problems associated
with externalities do not depend upon strategic behavior or exclusion. The bottom line is
while exclusion facilitates market provisioning, markets still systematically undersupply
some public goods because market demand fails to accurately reflect social demand.11
Third, reducing exclusion costs fixes an important supply-side problem and brings the
supply-side analysis of market provisioning of intellectual resources in line with provi-
sioning of most other infrastructural goods with similar cost structures—high fixed costs
coupled with low marginal costs (Frischmann and Hogendorn, 2015). The cost structure
of supply still can impact incentives to invest and impose deadweight losses during fixed
cost recovery, and excludability does not eliminate these issues.
The economic baseline for evaluating the sufficiency of market incentives to invest
should be average cost recovery. (There has been a vigorous debate on this point: Lemley,
2005; Duffy, 2005; Frischmann, 2007b; Frischmann and Lemley, 2007; Demsetz, 2008;
Frischmann, 2009b; Barron, 2010; see also Demsetz, 1967.) Sufficient incentives to invest
depend on an expectation of recovering total costs, including a competitive return on
capital investment. The cost structure suggests that incentives to invest will be insufficient
and undersupply will result, unless pricing above marginal cost and (at least) approximat-
ing average cost is sustainable. Without exclusion enabled by intellectual property or other
means, it might be impossible for suppliers to recover their average costs because free-
riding competitors would drive prices to marginal cost. Exclusion can enable sustainable
11
Some argue in favor of exclusion coupled with perfect price discrimination. I do not address
this argument in this chapter, but I do elsewhere. Simply put, the argument only works in very
limited contexts with much of which is most important in reality assumed away for purposes of
making the argument (Lunney, 2008; Lipsey and Lancaster, 1956). One of the lessons from the
infrastructure theory is that the allocation of surplus between producers and consumers can have
important implications for efficiency, productivity, and social welfare and should not be casually
disregarded as a mere wealth transfer, pecuniary externality, or distributional issue (Frischmann,
2012). I am working with economists Alain Marciano and Giovanni Ramello on ‘Relevantly
Irrelevant Externalities,’ a paper that will elaborate on this point.

average cost pricing and competition, in a sense facilitating markets. Enabling average
cost pricing does not ensure actual cost recovery, however. There are a number of practi-
cal obstacles to effectively implementing average cost pricing (Frischmann, 2012, ch. 8).
Moreover, competition and innovation can jeopardize cost recovery, for example, when
a new entrant figures out a way to compete with lower fixed costs. If exclusion is limited
in scope to actual misappropriation (free riding on the fixed cost investment of the first
entrant), then whether the first entrant is capable of recovering its costs will depend on
the fixed cost investments that others must make to enter the market as well as lead-time
advantages and other possible barriers to entry.12 On the other hand, exclusion also can
enable monopoly pricing and eliminate competition. What exclusion enables depends on
the strength and scope of exclusion and the market context. The legal right to exclude
can be narrowly or broadly constructed along various dimensions. For example, it can be
limited to actual copying of an entire intellectual work, broadened to block copying of
parts of the work, broadened to block similar but not identical copying, or broadened
beyond instances of copying to block independent creation, among other things. It can
also vary in other dimensions such as duration, the strength of remedies, and so on.
At the extreme, government could grant monopoly franchises with legal entry barriers.
Exclusion can vary in strength, scope, and market impact.
A difficult supply-side question is whether patent or copyright law should do more
than address actual free-riding risks (Le, 2004; Lemley, 2005; Duffy, 2005; Frischmann
and Lemley, 2007; Liivak, 2010). If patent and copyright laws aim to facilitate competi-
tive markets for intellectual resources and derivative products, a narrow focus on such
risks would be appropriate. However, if the laws aim to induce investment in intellectual
resources above and beyond what competitive markets would provide, a broader focus on
conveying market power and the ability to appropriate supra-competitive returns might
be appropriate. If the latter objective is chosen, however, then one would have to both
justify the need for extra inducement—why put a thumb on the scale in favor of invest-
ments in intellectual resources rather than other types of investments? Is the increase in
deadweight losses worth the gain?—and explain from a comparative institutional stand-
point why intellectual property laws are the preferred institution for making this social
investment—why intellectual property rights rather than government subsidies, a prize
system, or other alternatives (Frischmann, 2000; Barnett, 2004; Fisher, 2004; Burstein,
2012; Hemel and Ouellette, 2013; Sarnoff, 2013; see also Komesar, 1994, 2001). This is an
area in need of sophisticated and sustained research (Frischmann and McKenna, 2015).
2. Intellectual resources and activities, products, and processes

The subsection focuses on the added complexity associated with ‘the other crucial
quirkiness . . . that information is both input and output of its own production process.’
12
People sometimes emphasize the magnitude of fixed costs. In many cases, this does not
really matter so long as a second comer would have to sink the same amount. High fixed costs
can be a decent barrier to entry, provided that misappropriation is precluded and average cost
pricing is feasible. What seems to matter in such circumstances is the rate of fixed-cost-reducing
innovation—whether a second comer can figure out a way to enter more cheaply. Of course, this is
true in all sorts of markets. Certainly in some cases, incredibly high fixed costs may exceed capital
constraints for any single firm, but that raises a different problem altogether.

As noted, this effect is interesting and complex because it reveals necessary depend-
ence among generations, but there is more to it than that. It implicates the cumulative,
dynamic, and evolutionary nature of progress in intellectual-cultural systems, or, more
broadly, in the cultural environment.
Benkler (2006) focuses on how the ‘on the shoulders of giants’ effect makes ‘property-
like exclusive rights less appealing’ because it increases the deadweight losses from pricing
above the marginal cost of zero by making productive use of the nonrival resources
more costly. He notes: ‘Today’s users of information are not only today’s readers and
consumers. They are also today’s producers and tomorrow’s innovators.’ Simply put, users
are both consumers and producers. The fact that many intellectual resources are a form
of nonrival capital that supports production of even more nonrival capital suggests the
possibility of increasing returns to investing in such resources and leveraging nonrivalry
(Romer, 1996; Ochoa, 1996; Schmidt, 2003).
To the extent that this effect is taken seriously in economic and legal scholarship, atten-
tion is devoted to the relationship between two stages, the first- and second-generation
producers, the pioneer and improver. For example, Scotchmer (1991, 1996) has focused
on this effect. She emphasizes the importance of licensing intellectual property between
first- and second-generation inventors, and adequately compensating and maintaining
investment incentives to both stages of inventorship, given that many products are the
result of numerous improvements on previous inventions (Scotchmer, 1991, 1996; Green
and Scotchmer, 1995; Lemley, 1997; Merges and Nelson, 1994, 1990).
Yet we often take the ‘on the shoulders of giants’ effect for granted. For example,
we often take for granted the intellectual or cultural environment within which and
on which we (and others) build; this may be due to a romantic notion of authorship,
an inflated sense of self, or any number of things (Litman, 1990; Boyle, 2010; Lessig,
2001; Vaidhyanathan, 2001). Similarly, we often take for granted the various intellectual
outputs that emerge from our experience and engagement with the cultural environment;
we only have so much time and attention. Regardless, we use, make, and reuse intel-
lectual resources continuously in our lives. This seemingly trivial observation has some
interesting implications.
First, we need intellectual inputs to be intellectually productive and to make intellectual
progress in our lives. Second, the intellectual resources to which we have access will shape
the intellectual outputs we are capable of producing as well as our beliefs and desires
about what to produce; in a sense, they shape who we become (our beliefs, knowledge,
preferences) as we engage with the environment. Third, each producer and producer’s
output is thus dependent or contingent on various infrastructural inputs. ‘In order to
write today’s academic or news article, I need access to yesterday’s articles and reports.
In order to write today’s novel, movie, or song, I need to use and rework existing cultural
forms, such as story lines and twists’ (Benkler, 2006, p. 37). As Julie Cohen suggests, we
are situated within the cultural environment, shaping it while being shaped by it (Cohen,
2005, p. 370; Cohen, 2012).
Continuous situated engagement implies a stream of input-output relationships (i.e.,
input → output/input → output/input . . .). In many contexts, it may not be worth the
effort to pay attention to the continuous streams of relationships. Surely, we do not
need to acknowledge and consider each incremental addition associated with sensory
experience or thinking. Instead, we may conflate many input-output relationships into a

process (activity or practice) and pay attention only to particular outputs that are worthy
of attention.
It is common to talk about intellectual resources as identifiable, discrete things with
known properties and boundaries. The very notion of a ‘resource’ or ‘public good’ implies
such features. But this is a significant oversimplification. Intellectual resources often
have a dual nature—creation, invention, and innovation may be resources and activities.
Consider basic research: Is it a thing—a result, an input, an output, both—or is it a
process or activity that one engages in? It is both, right? Maybe this seems like a semantic
point, but isolating one from the other (product from process) for purposes of the law,
economic analysis, or just discussion loses something quite valuable.
Can we ‘discretize’ cumulative intellectual processes of creation, invention, and innova-
tion in a manner that makes analytic sense? We try to do so regularly within copyright and
patent law, but are we truly granting patents and copyrights over discrete outputs—over
discrete ‘things’? When we are dealing with streams of input-output relationships that
may or may not culminate in a consumer good, it can be difficult to isolate the ‘thing’ we
might identify as the invention or work of authorship, much less the intellectual contribu-
tion made by the person claiming patent or copyright (Madison, 2005). We recognize and
enforce (artificial) boundaries for purposes of constructing property rights and facilitat-
ing exclusion, coordination, and market provisioning, but our focus on ‘things’ (inputs,
outputs, resources, goods, and so on) often obscures the continuity and complexity of
the system.
Even if we reduce the number of input-output relationships we are willing to entertain,
we must acknowledge that intellectual progress involves a stream of such relationships,
and this requires acknowledgment of a potential stream of spillovers, or what we might
refer to as cascading spillovers. This is the case even if we assume a simple string of single
public good input-output relationships, where a single public good is produced at each
stage. In reality, each stage of production may involve multiple input and outputs, each
of which can be used productively to produce different outputs and potentially support
different production paths by many people. Many intellectual and cultural activities yield
social goods as well, in which cases the diffusion of a different set of externalities cascade
as well.
The conventional model of intellectual production represents progress in a linear
fashion, for example, from basic research to applied research and finally to commercial
application; or, alternatively, from idea conception to invention to commercial develop-
ment; or something similar. Linear models are intuitive and qualitatively appealing
because many economic and social policy questions that follow seem to have straight-
forward answers. For example, the government should support basic research as a form
of public goods production; the basic research pool should supply inputs for applied
research; private firms should step in at some point and bring the benefits of research
results to the public through commercialization. Yet the linear model is not an actual
scientific model of innovation or intellectual progress. Rather, as Benoît Godin (2006)
explains, a host of different actors—scientists seeking funding, economists advising
government agencies—constructed the linear model of innovation to classify research
activities, establish a connection between basic and applied research and eventually
commercial activities, and advance political and other agendas. Godin explains that the
simple three-stage ‘basic research → applied research → development’ model became

standardized when official government statisticians appropriated the three-stage model

as a means for classification of research to aid in statistical categorization, measure-
ment, and quantitative analysis (Godin, 2006; OECD, 1962). Yet the linear model
has been roundly criticized and rejected (Kline and Rosenberg, 1986). As Nathan
Rosenberg claimed in 1994, ‘Everyone knows that the linear model of innovation is
dead’ (Rosenberg, 1994, p. 139). Yet as Godin shows, the linear model remains intact
in the discourse despite its many criticisms. He observes that alternative models have
struggled to replace the linear model because they pose more difficult measurement
issues, and ‘with their multiple feedback loops look more like modern artwork or “a
plate of spaghetti and meatballs” than a useful analytical framework’ (Godin, 2006,
p. 639).
Intellectual production processes, and intellectual progress more generally, are often
nonlinear, multidirectional, stochastic, full of feedback loops, and difficult to model
(Dreyfuss, 2010; Godin, 2006; Knudsen, 2003, pp. 13, 24). There are various nonlinear
innovation models that incorporate dynamic interactions between different types of
research and even nonresearch activities as well as the background cultural environment
within which such interactions take place. For example, the ‘Chain-Linked Model,’ devel-
oped by S.J. Kline, incorporates feedback loops between research and the ‘existing corpus
of knowledge’ and emphasizes the importance of various different activities, procedures,
and external influences that play a role in innovative progress; there are multiple paths,
feedback loops, and various actors (Kline and Rosenberg, 1986; Kline, 1991a, 1991b). As
Kline and Rosenberg point out, the linear model’s omission of feedback loops, learning
from ‘short-coming and failures,’ and other features renders it incapable of dealing with
radical and incremental innovation.
The dynamic nature of progress often leads to unexpected spillover effects. For exam-
ple, an idea developed in one sector may lead to beneficial progress in another unrelated
(or marginally related) sector.
The practice of science is becoming increasingly interdisciplinary, and scientific progress in one
discipline is often propelled by advances in other, often apparently unrelated, fields. For example,
who would have thought that nuclear physics research (the study of the inner workings and
properties of the atomic nucleus) and data gathering techniques developed for experiments on
elementary particles (quarks and such) would lead to a device that has advanced the boundaries
of biomedical research and health care? Yet both of these lines of inquiry led ultimately to
Magnetic Resonance Imaging (MRI), a tool now used in laboratories and hospitals around the
world both to conduct basic biological research and also to diagnose illness. Such cross-over
between fields is yet another example of the unexpected payoffs that can come from basic
research. (Staff of House Comm. On Science, 1998)
There are countless examples in science, technology, and innovation, but these phenom-
ena are equally relevant in cultural systems too (Boyle, 2010, pp. 122–5; McLeod and
DiCola, 2011; Schatz, 1981, 1977, p. 44; Altman, 1987; Doll and Faller, 1986; Miller,
2007).
The cultural environment and its constituent innovation, science, culture, knowledge,
and other systems are dynamic evolutionary systems. Since how and what direction
the systems, environment, and consequently society evolve are not predetermined or
inevitable, institutions and social policies matter considerably; the cultural environment
we construct and sustain reflects deep normative values.

The dynamic, nonlinear, and multidirectional nature of these processes/activities

involves considerable uncertainty. This can be daunting, and for many, it is something we
hope to control, diminish, or eliminate over time.
Infrastructure theory suggests a very different outlook. When coupled with the nonrival
nature of intellectual resources, it suggests considerable social opportunity. The nonrival
nature of the resources means that intellectual capital generated at different points in the
‘stream’ may flow along many paths, potentially being used simultaneously by different
people in different settings as an input into multiple intellectual-cultural processes. There
are a variety of obstacles to the free flow and use of intellectual capital (e.g., limited
absorptive capacity, education, or capabilities to productively use the resources) (Lobel,
2015; Bontis, 2005, pp. 134–5). In particular contexts, there may be good reasons to restrict
the free flow and use of intellectual capital (e.g., to prevent misappropriation and protect
supply-side incentives to invest). But in light of nonrivalry and the ‘on the shoulders of
giants’ effect, the social opportunity deserves recognition and further attention.
The complex, dynamic, nonlinear, and multidirectional nature of intellectual progress
and the prevalence and variety of external effects in the cultural environment suggest that
focusing on optimality conditions may be a red herring (or worse). We are inevitably in
what economists call a second-best world because of the incredible number of incomplete
and missing markets in the cultural environment (Lipsey and Lancaster, 1956). Rather
than focus on achieving optimal government or market selection of public good invest-
ments, society is likely much better off focusing on indirect interventions that (1) support
public capabilities to participate in intellectual-cultural activities and (2) aim to lower the
costs of public goods production for a wide range of public goods while (3) maintaining
flexibility in the opportunities available to potential participants. In other words, support-
ing open infrastructure would better leverage nonrivalry, facilitate progress along many
paths, and sustain a spillover-rich cultural environment in which and with which members
of society are capable of interacting productively. This shift in focus has important
implications for appreciating intellectual property laws as semi-commons arrangements.
Here I summarize what this would mean for intellectual property laws: First, the laws
should focus on misappropriation risks with an aim to facilitate average cost recovery and
competition, rather than market power or monopoly; to the extent that certain areas war-
rant subsidies, then targeted subsidies seem more direct and less distorting than adjusting
the legal system for all areas, and such subsidies can be directed at infrastructural invest-
ments in the targeted area—for example, basic research in biotech or even directly funding
shared clinical trial infrastructure. This first point suggests that exclusion is important
but should be limited in scope. Second, in addition to facilitating exclusion, intellectual
property systems should aim to reduce information and transaction costs because such
cost reductions would apply to a wide range of public goods investments. This can be
done in a manner similar to traditional property law systems, by providing recordation,
registration, dispute resolution, and so on. Also, given the multitude of intellectual
property owners, private ordering solutions to collective management problems should
be facilitated as well. Third, and related to the first point, it should be understood that
the division of surplus often has efficiency consequences (rather than mere distributional
or equity consequences) because consumers are often productive users, even if their
productive use does not immediately generate a marketable good. This has consequences
for a variety of economic issues in intellectual property law. Fundamentally, and in stark

contrast with the conventional economic perspective, intellectual property systems should
be understood as exceptional, targeted interventions that construct semi-commons and
sustain a spillover-rich cultural environment.
C. Intellectual Property Laws as Semi-Commons Arrangements
The intellectual property laws construct semi-commons arrangements, complex mixtures

of interdependent private rights and commons (Heverly, 2003; Madison et al., 2010;
Frischmann and Lemley, 2007; Frischmann, 2007b; Yu, 2005, pp. 6–8; Loren, 2007;
Vetter, 2007; Smith, 2000). Semi-commons exist at different scales.
At a macro level, the cultural environment constitutes mixed infrastructure that
should be managed as a commons. As discussed, commons management aims to limit
both government and market shaping of the cultural environment and our lives, plans,
beliefs, and preferences. Commons management is a strong default position for the
cultural environment because users—autonomous individuals as well as social groups and
communities—get to shape the environment and choose what to say and do and how to
plan their lives, experiences, and interactions with each other and the environment.
Yet, as in the context of the natural environment, a pure open-access or commons
regime can lead to tragedy, in this context associated with undersupply of certain types of
intellectual resources. Consequently, intellectual property systems enclose and regulate a
select (albeit very broad) set of intellectual resources. Thus, an identifiable semi-commons
emerges at the macro level, with commons being the default form of management and
intellectual property enclosure being exceptional, albeit of broad scope and significant
importance. The unenclosed and enclosed are highly interdependent; much of which
exists in either space/environment/category depends substantially on complex interac-
tions and various inputs/contributions from the other space/environment/category. Given
tremendous difficulties in establishing and maintaining boundaries, and the dynamic and
complex nature of cultural-intellectual resource systems, the intellectual property laws
also mediate the relationships between the enclosed and unenclosed.
With respect to ideas, for example, the First Amendment, copyright, and patent interact
with each other and the public domain.
The discussion of ideas demonstrates the semi-commons structure and associated
interdependence between private and public at the macro level; at the same time, it reveals
the semi-commons structure at the meso (or intermediate) level of the copyright and
patent systems. Though both legal systems construct private rights and enclose a set of
intellectual resources, neither constitutes pure private rights nor enclosure. Rather, both
copyright and patent laws themselves are semi-commons arrangements that mix both
private rights and commons. Both legal systems are designed to sustain incentives and
spillovers (Frischmann and Lemley, 2007; Frischmann, 2007b, p. 659).
Copyright law creates a semi-commons arrangement—a complex mix of private rights
and commons. The rights granted by copyright law—specifically, the 17 U.S.C. § 106
rights to reproduce, display, perform, distribute, and make derivative works—provide
incentives to create and disseminate works by facilitating transactions and lowering the
costs of excluding competitors from using the expression. The supply-side incentives
affected by copyright extend beyond the initial investment in creation to investments in
content development and dissemination. What must be encouraged is not only works’

creation but also their publication, dissemination, and productive use. Like traditional
property rights, copyright facilitates transactions over certain uses of creative expression,
and thereby enables rights holders to appropriate some of the surplus generated by their
investments in creation, development, and dissemination. In this fashion, the private
rights component of copyright law improves investment incentives through the opera-
tion of the market mechanism; in a sense, it uses the market to achieve a broader set of
economic and social ends.
The commons component of copyright law promotes spillovers; or, to put it another
way, the commons component of copyright law avoids market or government allocation
of resources for certain ranges of uses and for certain elements of a copyrighted work.
Through a variety of leaks and limitations on the private rights granted, copyright law
sustains common access to and use of resources needed to participate in a wide variety
of intellectually productive activities. Many of these activities generate socially valuable
spillovers: benefits realized by consumers, users, and third parties that are external to a
creator’s decision to produce the work and to any transactions involving the work. For
example, due to its limited duration, copyright has generated temporal externalities. A
work that enters the public domain is free for public use, and any value derived from this
use is external to both the creator’s decision to produce the work and any transactions
involving the work. Similarly, due to copyright’s limited scope, copyright generates
externalities that accrue to other creators, even competitors, as these entities can freely
use various unprotected elements of a work, such as an idea, theme, or functional feature.
Copyright’s limited scope may also generate externalities in complementary technology
markets: for example, companies can design and build products such as DVD players
and iPods that facilitate the enjoyment of copyrighted works. Finally, copyright produces
externalities when consumers productively use or reuse works. Creating and consuming
creative expression of different types develops human capital, educates, and socializes in
a manner that benefits not only creators and consumers but also nonparticipants.
Patent law, like copyright, is a semi-commons that promotes both ownership of rights
and spillovers, but the particular ways in which patent law and copyright law permit
‘leakage’ differ significantly. Patent law protections have a much shorter duration than
copyright, permitting inventions to enter the public domain more quickly. Patent law also
excludes some inventions from protection because requirements for obtaining protection
are stricter. Once inventors do obtain protection, however, the right they obtain is much
stronger and less leaky than that afforded by copyright law.
Patent law promotes spillovers in several ways. Patents generate externalities by facilitat-
ing learning and disclosure. Indeed, patent law, unlike copyright law, requires the patent
owner to teach the public how to make and use the invention, and this is often identified as
a central function of the patent system, though in practice it is considerably less important
than the system’s incentive effects. Patents lead to temporal externalities—spillovers
occur when the patent expires. Temporal spillovers are quite significant. For example, the
overwhelming majority of the social benefit associated with the telephone (and, for that
matter, the paper clip) occurred after the basic patents on those technologies expired.
These legal systems sustain commons by excluding resources and designating them
unprotectable, but also by sustaining public access to privately owned resources for certain
types of uses (Litman, 1990; Landes and Posner, 2003). In a sense, the legal systems also
construct semi-commons at the micro level of the protected expression or invention.

At this micro level, copyright law appears to be more sensitive to and accommodating
of social demand for commons management of infrastructural expression than patent
law is with respect to infrastructural invention. ‘Copyright encourages and sustains
participation in intellectually productive activities that both generate and use expressive
works to communicate, entertain, teach, and engage us in many different ways. Many of
these activities—e.g., education, community development, democratic discourse, political
participation—generate socially valuable . . . spillovers’ (Frischmann, 2007b, p. 672). For
example, fair use is a particularly important copyright law doctrine that aims to preserve
public capabilities to use copyright protected expression in various ways (Frischmann and
Lemley, 2007, pp. 286–90). As Lemley and I explain:
Many paradigmatic uses deemed fair involve use of a work to engage in activities that yield
diffuse, small-scale spillovers to a community. Using a work for educational purposes, for
example, not only benefits the users themselves, but also, in a small way, benefits others in the
users’ community with whom users have interdependent relations—reading and learning builds
socially valuable human capital. Critiquing a work similarly benefits not only the user but also, in
a small way, others in the users’ community—not only because those others may read the critique
itself, but also because engaging in critical commentary is a form of creative and cultural activity
that builds socially valuable human capital. We recognize that observing and measuring these
spillover benefits is probably an impossible task. That is our point, in fact. As a society, on the
whole, we recognize the value of active, wide-spread participation in these types of activities,
and we know that creative expression is essential to participation. Thus, we encourage common
access to and use of expression for these types of activities.
Other features of copyright law, such as constraints on the exclusive scope of rights (e.g.,
private display and performance is permissible) or judicial willingness to provide ‘thin-
ner’ protection for certain types of works, also provide breathing space. The levels-of-
abstraction framework described above permits judges to adjust the scope of copyright
based on the context and nature of the work, and judges routinely filter infrastructural
elements of a work in addition to ideas (e.g., stock literary elements) (Lee, 2008). Patent
law is not as sensitive to social demand for commons management of infrastructural
invention. Lee (2008) provides an excellent account of how trademark, copyright, and
patent law accommodate social demand for access to infrastructural works. The primary
commons components of patent law are its mechanisms for exclusion and conferral to
the public domain (e.g., disclosure, strict qualification criteria, and duration). There is
no fair use or functionally equivalent doctrine. Courts implicitly engage in levels-of-
abstraction analysis when construing claims and determining infringement, and that
provides an opportunity to adjust patent scope, but the analysis is not sensitive to social
demand or the infrastructural characteristics of the invention. Scholars have advanced
arguments for a patent fair use doctrine, a more robust experimental use defense, and
adjusting remedies when patents on infrastructural inventions are infringed (O’Rourke,
2000 (fair use); Strandburg, 2011 (same); Mueller, 2001, 9–10 (experimental use);
Caruso, 2003 (same); Lee, 2008 (exploring the concept of intellectual infrastructure, and
proposing that ‘courts should consider the infrastructural use of a patented invention
when determining infringement remedies and, in certain circumstances, allow such use
to continue by a downstream user contingent upon providing compensation to the
patentee’)).

IV. CONCLUSION
Infrastructure theory reflects a combination of ideas drawn from multiple disciplines

within economics and law. Though not fully explored in this short chapter, it also is
intimately connected to the commons field of study. This adaptation of a book chapter
necessarily provided a limited view. Readers are encouraged to draw on the extensive
reference list.
REFERENCES
There is an extensive literature that spans multiple disciplines. Most of this short chapter is derived directly
from Frischmann (2012), which provides a much more extensive bibliography. In the spirit of compiling a list
of sources for those interested in infrastructure theory, I have provided some additional sources not cited in
the text of the chapter.
Altman, Rick. 1987. The American Film Musical. Bloomington, IN: Indiana University Press.
Amabile, Teresa M. 1996. Creativity in Context. Routledge.
Arrow, Kenneth. 1962. ‘Economic Welfare and the Allocation of Resources for Inventions,’ in R.R. Nelson, ed.,
The Rate and Direction of Inventive Activity. Princeton, NJ: Princeton University Press.
Arrow, Kenneth J. 1970. ‘The Organization of Economic Activity: Issues Pertinent to the Choice of Market
versus Nonmarket Allocation,’ in Robert H. Haveman and Julius Margolis, eds., Public Expenditure and Policy
Analysis. Chicago, IL: Markham Publishing Co.
Aschauer, David Alan. 2000. ‘Public Capital and Economic Growth: Issues of Quantity, Finance, and
Efficiency,’ 48 Journal of Economic Development and Cultural Change 391–406.
Balkin, Jack M. 2012. ‘The First Amendment is an Information Policy,’ 41 Hofstra Law Review 1–30.
Barnett, Jonathan. 2004. ‘Private Protection of Patentable Goods,’ 25 Cardozo Law Review 1251–313.
Barron, Anne. 2010. ‘Copyright Infringement, “Free-Riding” and the Lifeworld,’ in Lionel Bently, Jennifer
Davis, and Jane Ginsburg, eds., Copyright and Piracy: An Interdisciplinary Critique. Cambridge: Cambridge
University Press.
Benkler, Yochai. 1998a. ‘Overcoming Agoraphobia: Building the Commons of the Digitally Networked
Environment,’ 11 Harvard Journal of Law and Technology 287–400.
Benkler, Yochai. 1998b. ‘The Commons as a Neglected Factor of Information Policy,’ Remarks at the
Telecommunications Policy Research Conference, Arlington VA (Sept. 1998), available at http://www.benkler.
org/commons.pdf (last visited Jan. 1, 2012).
Benkler, Yochai. 1999. ‘Free as the Air to Common Use: First Amendment Constraints on Enclosure of the
Public Domain,’ 74 New York University Law Review 354–446.
Benkler, Yochai. 2001. ‘Property, Commons, and the First Amendment: Towards a Core Common Infrastructure,’
White Paper for the First Amendment Program, Brennan Center for Justice at NYU Law School, available
at http://www.benkler.org/WhitePaper.pdf (last visited Jan. 1, 2012).
Benkler, Yochai. 2013. ‘Commons and Growth: The Essential Role of Open Commons in Market Economies,’
80 University of Chicago Law Review 1499–555.
Bontis, Nick. 2005. ‘National Intellectual Capital Index: The Benchmarking of Arab Countries,’ in Ahmed
Bounfour and Leif Edvinsson, eds., Intellectual Capital for Communities. Oxford, UK: Butterworth-
Heinemann.
Bourdieu, Pierre. 1986. ‘The Forms of Capital,’ in John G. Richardson, ed., Handbook of Theory and Research
for the Sociology of Education. Westport, CT: Greenwood Press.
Boyle, James. 1996. Shamans, Software, and Spleens: Law and the Construction of the Information Society.
Boyle, James. 1997. ‘A Politics of Intellectual Property: Environmentalism for the Net?,’ 47 Duke Law Journal
87–116.
Boyle, James. 2003. ‘The Second Enclosure Movement and the Construction of the Public Domain,’ 66 Law
Boyle, James. 2010. The Public Domain: Enclosing the Commons of the Mind. New Haven, CT: Yale University
Press.
Branscomb, Lewis M., and James H. Keller. 1996. ‘Introduction,’ in Lewis M. Branscomb and James H. Keller,

eds., Converging Infrastructures: Intelligent Transportation and the National Information Infrastructure.
Buchanan, James M., and William Stubblebine Craig. 1962. ‘Externality,’ 29 Economica 371–84.
Burk, Dan, and Mark A. Lemley. 2003. ‘Policy Levers in Patent Law,’ 89 Virginia Law Review 1575–696.
Burstein, Michael. 2012. ‘Exchanging Information without Intellectual Property,’ 91 Texas Law Review 227–82.
Button, Kenneth. 1996. ‘Ownership, Investment and Pricing of Transport and Communications Infrastructure,’
in David F. Batten and Charlie Karlsson, eds., Infrastructure and the Complexity of Economic Development.
Heidelberg: Springer.
Chiang, Tun-Jen. 2011. ‘The Levels of Abstraction Problem in Patent Law,’ 105 Northwestern University Law
Review 1097–152.
Christensen, Clayton M. 1997. The Innovator’s Dilemma. Boston, MA: Harvard Business School Press.
Coase, Ronald H. 1946. ‘The Marginal Cost Controversy,’ 13 Economica 169–82.
Coase, Ronald H. 1947. ‘The Marginal Cost Controversy: Some Further Comments,’ 14 Economica 150–53.
Coase, Ronald H. 1960. ‘The Problem of Social Cost,’ 3 Journal of Law and Economics 1–44.
Cohen, Julie E. 1998. ‘Lochner in Cyberspace: The New Economic Orthodoxy of “Rights Management,”’ 97
Michigan Law Review 462.
Cohen, Julie E. 2001. ‘Examined Lives: Informational Privacy and the Subject as Object,’ 52 Stanford Law
Review 1373–437.
Cohen, Julie E. 2005. ‘The Place of the User in Copyright Law,’ 74 Fordham Law Review 347–74.
Cohen, Julie E. 2007. ‘Creativity and Culture in Copyright Theory,’ 40 U.C. Davis Law Review 1151.
Cohen, Julie E. 2012. Configuring the Networked Self: Law, Code, and the Play of Everyday Practice. New Haven,
CT: Yale University Press.
Cohen, Wesley, Richard Nelson, and John Walsh. 2000. ‘Protecting Their Intellectual Assets: Appropriability
Conditions and Why U.S. Manufacturing Firms Patent (or Not),’ NBER Working Paper 7552, available at
http://www.nber.org/papers/w7552 (last visited Jan. 1, 2012).
Colander, David, ed. 1996. Beyond Microfoundations: Post Walrasian Economics. Cambridge: Cambridge
University Press.
Cornes, Richard, and Todd Sandler. 1996. The Theory of Externalities, Public Goods, and Club Goods (2nd edn).
David, Paul A., and Dominique Foray. 1996. ‘Information Distribution and the Growth of Economically
Valuable Knowledge: A Rationale for Technological Infrastructure Policies,’ in Morris Teubal, Dominque
Foray, Moshe Justman, and Ehud Zuscovitch, eds., Technological Infrastructure Policy: An International
Perspective. Dordrecht: Springer Science+Business Media.
DeMartino, George. 2000. Global Economy, Global Justice. New York, NY: Routledge.
Demsetz, Harold. 1967. ‘Toward a Theory of Property Rights,’ 57 The American Economic Review 347–59.
Demsetz, Harold. 1970. ‘The Private Production of Public Goods,’ 13 Journal of Law and Economics 293–306.
Demsetz, Harold. 2008. ‘Frischmann’s View of “Toward a Theory of Property Rights,”’ 4 Review of Law and
Economics 127–32.
Doll, Susan, and Greg Faller. 1986. ‘Blade Runner and Genre: Film Noir and Science Fiction,’ 14 Literature/
Film Quarterly 89–100.
Dreyfuss, Rochelle Cooper. 2010. ‘Does IP Need IP? Accommodating Intellectual Production Outside the
Duffy, John F. 2004. ‘The Marginal Cost Controversy in Intellectual Property,’ 71 University of Chicago Law
Review 37–56.
Duffy, John F. 2005. ‘Comment: Intellectual Property Isolationism and the Average Cost Thesis,’ 83 Texas Law
Review 1077.
Eisenberg, Rebecca S. 2001. ‘Bargaining Over the Transfer of Proprietary Research Tools: Is this Market Failing
or Emerging?,’ in Rochelle C. Dreyfuss, Diane L. Zimmerman, and Harry First, eds., Expanding the Boundaries
of Intellectual Property: Innovation Policy for the Knowledge Society. Oxford: Oxford University Press.
Fisher III, William W. 2004. Promises to Keep: Technology, Law, and the Future of Entertainment. Stanford, CA:
Stanford University Press.
Flores, Nicholas E. 2003. ‘Conceptual Framework for Nonmarket Valuation,’ in Patricia A. Champ, Kevin J.
Boyle, and Thomas C. Brown, eds., A Primer on Nonmarket Valuation. Dordrecht: Springer.
Freeman III, A. Myrick. 2003. ‘Economic Valuation: What and Why,’ in Patricia A. Champ, Kevin J. Boyle, and
Thomas C. Brown, eds., A Primer on Nonmarket Valuation. Dordrecht: Springer.
Frischmann, Brett M. 2000. ‘Innovation and Institutions: Rethinking the Economics of U.S. Science and
Technology Policy,’ 24 Vermont Law Review 347–416.
Frischmann, Brett M. 2005. ‘An Economic Theory of Infrastructure and Commons Management,’ 89

Frischmann, Brett M. 2007a. ‘Cultural Environmentalism and the Wealth of Networks,’ 74 University of Chicago
Frischmann, Brett M. 2007b. ‘Evaluating the Demsetzian Trend in Copyright Law,’ 3 Review of Law and
Economics 649–77.
Frischmann, Brett M. 2008. ‘Speech, Spillovers, and the First Amendment,’ 2008 University of Chicago Legal
Forum 301–33.
Frischmann, Brett M. 2009a. ‘The Pull of Patents,’ 77 Fordham Law Review 2143–67.
Frischmann, Brett M. 2009b. ‘Spillovers Theory and Its Conceptual Boundaries,’ 51 William and Mary Law
Review 801–24.
Frischmann, Brett M. 2012. Infrastructure: The Social Value of Shared Resources. Oxford: Oxford University Press.
Frischmann, Brett M. 2017. ‘Capabilities, Spillovers, and Intellectual Progress: Toward a Human Flourishing
Theory for Intellectual Property,’ 14 Review of Economic Research on Copyright Issues 1–38.
Frischmann, Brett M., and Christiaan Hogendorn. 2015. ‘Retrospectives: The Marginal Cost Controversy,’ 29
The Journal of Economic Perspectives 193–205.
Frischmann, Brett M., and Mark A. Lemley. 2007. ‘Spillovers,’ 100 Columbia Law Review 101–43.
Frischmann, Brett M., and Mark McKenna. 2015. ‘Comparative Analysis of (Innovation) Failures and
Institutions in Context,’ Working Paper.
Frischmann, Brett M., Michael Madison, and Katherine Strandburg, eds. 2014. Governing Knowledge Commons.
Ghosh, Arghya, and Kieron Meagher. 2004. ‘Political Economy of Infrastructure Investment: A Spatial
Approach,’ in North American Econometric Society Summer Meeting at Brown University, Providence, RI.
Ghosh, Shubha. 2004. ‘Patents and the Regulatory State: Rethinking the Patent Bargain Metaphor after Eldred,’
Ghosh, Shubha. 2008. ‘Decoding and Recoding Natural Monopoly, Regulation, and Intellectual Property,’ 2008
University of Illinois Law Review 1125–84.
Gifford, Daniel J., and Robert T. Kudrle. 2010. ‘The Law and Economics of Price Discrimination in Modern
Economies: Time for Reconciliation?,’ 43 University of California Davis Law Review 1235–93.
Godin, Benoît. 2006. ‘The Linear Model of Innovation: The Historical Construction of an Analytical
Framework,’ 31 Science, Technology, and Human Values 639–67.
Gordon, Wendy J. 1993. ‘A Property Right in Self-Expression: Equality and Individualism in the Natural Law
of Intellectual Property,’ 102 Yale Law Journal 1533–609.
Green, Jerry, and Suzanne Scotchmer. 1995. ‘On the Division of Profits between Sequential Innovators,’ 26
Hamilton, Walter H. 1930. ‘Affection with Public Interest,’ 39 Yale Law Journal 1089.
Helpman, Elhanan. 1998. ‘Introduction,’ in Ehanan Helpman, ed., General Purpose Technologies and Economic
Growth. Cambridge, MA: MIT Press.
Hemel, Daniel Jacob, and Lisa Larrimore Ouellette. 2013. ‘Beyond the Patents-Prizes Debate,’ 92 Texas Law
Review 303–82.
Hess, Charlotte, and Elinor Ostrom. 2003. ‘Ideas, Artifacts, and Facilities: Information as a Common-Pool
Resource,’ 66 Law and Contemporary Problems 111–45.
Hess, Charlotte, and Elinor Ostrom, eds. 2007. Understanding Knowledge as a Commons: From Theory to
Practice. Cambridge, MA: MIT Press.
Heverly, Robert. 2003. ‘The Information Semicommons,’ 18 Berkeley Technology Law Journal 1127–89.
Hotelling, Harold. 1938. ‘The General Welfare in Relation to Problems of Taxation and of Railway and Utility
Rates,’ 6 Econometrica 242–69.
Jones, Charles. 1998. Introduction to Economic Growth. New York, NY: W.W. Norton and Company.
Justman, Moshe, and Morris Teubal. 1996. ‘Technological Infrastructure Policy (TIP): Creating Capabilities
and Building Markets,’ in Morris Teubal, Dominique Foray, Moshe Justman, and Ehud Zuscovitch, eds.,
Technological Infrastructure Policy: An International Perspective. Dordrecht: Springer Science+Business
Media.
Kahn, Alfred E. 1988. The Economics of Regulation: Principles and Institutions. Cambridge, MA: MIT Press.
Review 697–754.
Kitch, Edmund W. 1977. ‘The Nature and Function of the Patent System,’ 20 Journal of Law and Economics
265–90.
Kline, Stephen J. 1991a. ‘Models of Innovation and Their Policy Consequences,’ in David Kingery, ed.,
Japanese/American Technological Innovation. Amsterdam: Elsevier Science Publishing Co., Inc.
Kline, Stephen J. 1991b. ‘Styles of Innovation and Their Cultural Basis (Part 1),’ 21 Chemtech 472–80.
Kline, Stephen J., and Nathan Rosenberg. 1986. ‘An Overview of Innovation,’ in Ralph Landau and Nathan
Rosenberg, eds., The Positive Sum Strategy. Washington, D.C.: The National Academies Press.

Knudsen, Christian. 2003. ‘The Essential Tension in the Social Sciences: Between the “Unification” and
“Fragmentation” Traps,’ in Hans Siggaard Jensen, Lykke Margot Ricard, and Morten Thanning Vendelø,
eds., The Evolution of Scientific Knowledge. Cheltenham, UK and Northampton, MA, USA: Edward Elgar
Publishing.
Komesar, Neil K. 1994. Imperfect Alternatives: Choosing Institutions in Law, Economics, and Public Policy.
Chicago, IL: University of Chicago Press.
Komesar, Neil K. 2001. Law’s Limits: The Rule of Law and the Supply and Demand of Rights. Cambridge:
Krackhardt, David. 1998. ‘Endogenous Preferences: A Structural Approach,’ in Jennifer Halpern and Robert
Stern, eds., Debating Rationality: Non-Rational Aspects of Organizational Decision Making. Cornell, NY:
Cornell University Press.
Landes, William M., and Richard Posner. 2003. The Economic Structure of Intellectual Property Law.
Landsburg, Steven E. 2008. Price Theory and Applications. Mason, OH: Thomson Southwestern.
Le, Net. 2004. ‘Sunk Costs, Free-Riding Justifications, and Compulsory Licensing of Interfaces,’ 1 Review of
Economic Research on Copyright Issues 29–53.
Lee, Peter. 2008. ‘The Evolution of Intellectual Infrastructure,’ 83 Washington Law Review 39–120.
Lee, Peter. 2009. ‘Interface: The Push and Pull of Patents,’ 77 Fordham Law Review 2225–35.
Lee, Robin S., and Tim Wu. 2009. ‘Subsidizing Creativity through Network Design: Zero Pricing and Net
Neutrality,’ 23 Journal of Economic Perspectives 61–76.
Lemley, Mark A. 1997. ‘The Economics of Improvement in Intellectual Property Law,’ 75 Texas Law Review 989.
Lemley, Mark A., and David McGowan. 1998. ‘Legal Implications of Network Economic Effects,’ 86 California
Lessig, Lawrence. 2001. The Future of Ideas: The Fate of the Commons in a Connected World. New York, NY:
Random House.
Lessig, Lawrence. 2004. Free Culture: How Big Media Uses Technology and the Law to Lock down Culture and
Control Creativity. New York, NY: Penguin Press.
Lessig, Lawrence. 2005. ‘Re-Marking the Progress in Frischmann,’ 89 Minnesota Law Review 1031–43.
Levin, Richard, Alvin Klevorick, Richard Nelson, and Sidney Winter. 1987. ‘Appropriating the Returns from
Industrial Research and Development,’ 3 Brookings Papers on Economic Activity 783–831.
Lichtman, Douglas. 2000. ‘Property Rights in Emerging Platform Technologies,’ 29 Journal of Legal Studies 615–48.
Liivak, Oskar. 2010. ‘Rethinking the Concept of Exclusion in Patent Law,’ 98 Georgetown Law Journal 1643–91.
Lipsey, R.G., and Kelvin Lancaster. 1956. ‘The General Theory of the Second Best,’ 24 Review of Economic
Studies 11–32.
Litman, Jessica. 1990. ‘The Public Domain,’ 39 Emory Law Journal 965–1023.
Liu, K.C. 2008. ‘Rationalising the Regime of Compulsory Patent Licensing by the Essential Facilities Doctrine,’
7 International Review of Intellectual Property and Competition Law 757–74.
Lobel, Orly. 2015. ‘The New Cognitive Property: Human Capital Law and the Reach of Intellectual Property,’
Loren, Lydia Pallas. 2007. ‘Building a Reliable Semicommons of Creative Works: Enforcement of Creative
Commons Licenses and Limited Abandonment of Copyright,’ 14 George Mason Law Review 271–328.
Lunney, Glynn S., Jr. 2008. ‘Copyright’s Price Discrimination Panacea,’ 21 Harvard Journal of Law and
Madison, Mike J. 2005. ‘Law as Design: Objects, Concepts, and Digital Things,’ 56 Case Western Reserve Law
Review 381–478.
Madison, Mike J., Brett M. Frischmann, and Katherine Strandburg. 2010. ‘Constructed Commons in the
Mandel, Greg. 2008. ‘When to Open Infrastructure Access,’ 35 Ecology Law Quarterly 205–14.
Mansfield, Edwin. 1988. ‘Intellectual Property Rights, Technological Change, and Economic Growth,’ in
Charles Walker and Mark Bloomfield, eds., Intellectual Property Rights and Capital Formation in the Next
Decade. Lanham, MD: University Press of America.
Mansfield, Edwin, Mark Schwartz, and Samuel Wagner. 1981. ‘Imitation Costs and Patents: An Empirical
Study,’ 91 Economic Journal 907–18.
Merges, Robert P. 2004. ‘A New Dynamism in the Public Domain,’ 71 University of Chicago Law Review
183–203.

Merges, Robert P., and Richard R. Nelson. 1994, ‘On Limiting or Encouraging Rivalry in Technical Progress:
The Effect of Patent Scope Decisions,’ 25 Journal of Economic Behavior and Organization 1–24.
Miller, Zane L. 2007. ‘Urban Life and Urban Landscape,’ in Erik Bond, ed., Reading London: Urban Speculation
and Imaginative Government in Eighteenth-Century Literature. Columbus, OH: Ohio State University Press.
Musgrave, Richard. 2008. ‘Merit Goods,’ in Steven N. Durlauf and Lawrence E. Blume, eds., The New Palgrave
Dictionary of Economics (2nd edn). New York, NY: Palgrave Macmillan.
Musgrave, Richard, and Peggy Musgrave. 1984. Public Finance in Theory and Practice (4th edn). New York,
NY: McGraw-Hill.
National Research Council. 1987. Infrastructure for the 21st Century: Framework for a Research Agenda.
Washington, D.C.: National Academy Press.
297–306.
Nelson, Richard R. 1982. ‘Government Stimulus of Technological Progress: Lessons from American History,’
in Richard Nelson, ed., Government and Technical Progress: A Cross-Industry Analysis. New York, NY:
Pergamon Press.
Nelson, Richard R. 1987a. Understanding Technical Change as an Evolutionary Process (F. De Cries Lectures in
Economics). Amsterdam: Elsevier Science Ltd.
Nelson, Richard R. 1987b. ‘Roles of Government in a Mixed Economy,’ 6 Journal of Policy Analysis and
Management 541–66.
North, Douglass C. 1990. Institutions, Institutional Change and Economic Performance. New York, NY:
North, Douglass C. 2005. Understanding the Process of Economic Change. Princeton, NJ: Princeton University
Press.
O’Rourke, Maureen A. 2000. ‘Toward a Doctrine of Fair Use in Patent Law,’ 100 Columbia Law Review
1177–2195.
Oakland, W.H. 1987. ‘Theory of Public Goods,’ in Alan J. Auerbach and Martin Feldstein, eds., Handbook of
Public Economics (vol. 2). Amsterdam: Elsevier.
Ochoa, Orlando A. 1996. Growth, Trade, and Endogenous Technology: A Study of OECD Manufacturing. New
York, NY: Palgrave Macmillan.
OECD. 1962. The Measurement of Scientific and Technical Activities: Proposed Standard Practice for Surveys of
Research and Development. Paris: OECD Publishing.
OECD. 2000. Knowledge Management in the Learning Society. Paris: OECD Publishing.
Olson, Mancur. 1965. The Logic of Collective Action: Public Goods and the Theory of Groups. Cambridge, MA:
Opderbeck, David. 2009. ‘Deconstructing Jefferson’s Candle: Towards a Critical Realist Approach to Cultural
Environmentalism and Information Policy,’ 49 Jurimetrics 203–43.
Ostrom, Elinor. 1990. Governing the Commons: The Evolution of Institutions for Collective Action. Cambridge:
Ostrom, Elinor. 2005. Understanding Institutional Diversity. Princeton, NJ: Princeton University Press.
Ostrom, Elinor. 2007. ‘A Diagnostic Approach for Going Beyond Panaceas,’ 104 Proceedings of the National
Academy of Sciences of the United States of America 15181–7.
Papandreou, Andreas A. 1994. Externality and Institutions. Oxford: Oxford University Press.
Peat, F.D. 2002. From Certainty to Uncertainty: The Story of Science and Ideas in the Twentieth Century.
Washington, D.C.: National Academy Press.
Pigou, Arthur Cecil. 1920. The Economics of Welfare. London: Macmillan and Co.
Pink, Daniel, 2010. Drive: The Surprising Truth About What Motivates Us. Riverhead Books.
Portes, Alejandro. 1998. ‘Social Capital: Its Origins and Applications in Modern Sociology,’ 24 Annual Review
of Sociology 1.
Putnam, Robert. 2000. Bowling Alone: The Collapse and Revival of American Community. New York, NY: Simon
and Schuster.
Rai, Arti Kaur. 1999. ‘Regulating Scientific Research: Intellectual Property Rights and the Norms of Science,’
94 Northwestern University Law Review 77–1547.
Ramsey, Frank P. 1927. ‘A Contribution to the Theory of Taxation,’ 37 The Economic Journal 47–61.
Reichman, J.H., and Paul F. Uhlir. 2003. ‘A Contractually Reconstructed Research Commons for Scientific
Data in a Highly Protectionist Intellectual Property Environment,’ 66 Law and Contemporary Problems
315–462.
Robinson, Joan. 1933. The Economics of Imperfect Competition. New York, NY: St. Martin’s Press.
Romer, Paul M. 1986. ‘Increasing Returns and Long Run Growth,’ 94 Journal of Political Economy 1002–37.
Romer, Paul M. 1996. ‘Why, Indeed, in America? Theory, History, and the Origins of Modern Economic
Growth,’ 86 American Economic Review 202–6.

Rose, Carol. 2003. ‘Romans, Roads, and Romantic Creators: Traditions of Public Property in the Information
Age,’ 66 Law and Contemporary Problems 89–110.
Rosenberg, Nathan. 1994. Exploring the Black Box: Technology, Economics, and History. Cambridge:
Samuelson, Pamela. 1990. ‘Benson Revisited: The Case against Patent Protection for Algorithms and Other
Computer Program-Related Inventions,’ 39 Emory Law Journal 1025–401.
Samuelson, Pamela. 2006. ‘Enriching Discourse on Public Domains,’ 55 Duke Law Journal 783–834.
Samuelson, Paul A. 1954. ‘A Pure Theory of Public Expenditure,’ 36 Review of Economics and Statistics
387–9.
Samuelson, Paul A. 1958. ‘Aspects of Public Expenditure Theories,’ 40 Review of Economics and Statistics 332–8.
Sarnoff, Joshua. 2013. ‘Government Choices in Innovation Funding (with Reference to Climate Change),’ 62
Emory Law Journal 1087–613.
Schatz, Thomas G. 1977. ‘New Directions in Film Genre Studies (a Response to Charles F. Altman),’ in Ben
Lawton and Janet Staiger, eds., Film Historical-Theoretical Speculations (The 1977 Film Studies Annual: Part
2). Pleasantville, NY: Redgrave Publishing.
Schatz, Thomas G. 1981. Hollywood Genres: Formulas, Filmmaking, and the Studio System. New York, NY:
McGraw-Hill.
Scherer, Frederic M. 1984. Innovation and Growth: Schumpeterian Perspectives. Cambridge, MA: MIT Press.
Scherer, Frederic M., and David Ross. 1990. Industrial Market Structure and Economic Performance. Boston,
MA: Houghton Mifflin.
Schmalensee, Richard. 1981. ‘Output and Welfare Implications of Monopolistic Third-Degree Price
Discrimination,’ 71 American Economic Review 242–7.
Schmidt, Gordon W. 2003. Dynamics of Endogenous Economic Growth: A Case Study for the Romer Model.
Amsterdam: Elsevier.
Schultz, Jason, and Pamela Samuelson. 2011. ‘Clues for Determining Whether Business and Service Innovations
are Unpatentable Abstract Ideas,’ 15 Lewis and Clark Law Review 109–1107.
Scotchmer, Suzanne. 1996. ‘Protecting Early Innovators: Should Second-Generation Products be Patentable?,’
27 RAND Journal of Economics 322–31.
Sen, Amartya. 1993. ‘Capability and Well-Being,’ in Martha C. Nussbaum and Amartya Sen, eds., The Quality
of Life. New York, NY: Oxford Clarendon Press.
Sen, Amartya. 2009. The Idea of Justice. Cambridge, MA: The Belknap Press of Harvard University Press.
Shapiro, Carl, and Hal R. Varian. 1998. Information Rules: A Strategic Guide to the Network Economy. Boston,
Smith, Henry E. 2000. ‘Semicommon Property Rights and Scattering in the Open Fields,’ 29 Journal of Legal
Studies 131–69.
Staff of House Comm. on Science. 1998. 105th Cong., 1st sess., ‘Unlocking Our Future: Toward a New
National Science Policy’ 8 (Comm. Print 1998), available at http://www.access.gpo.gov/congress/house/science
/Cp105-b/science105b.pdf (last visited Jan. 1, 2012).
Steinmueller, W. Edward. 1996. ‘Technological Infrastructure in Information Technology Industries,’ in Morris
Teubal, Dominique Foray, Moshe Justman, and Ehud Zuscovitch, eds., Technological Infrastructure Policy:
An International Perspective. Dordrecht: Springer Science+Business Media.
Stole, Lars A. 2007. ‘Price Discrimination and Competition,’ in Mark Armstrong and Robert K. Porter, eds.,
Handbook of Industrial Organization. Amsterdam: Elsevier.
Tassey, Gregory. 1996. ‘Infratechnologies and Economic Growth,’ in Morris Teubal, Dominique Foray,
Moshe Justman, and Ehud Zuscovitch, eds., Technological Infrastructure Policy: An International
Perspective. Dordrecht: Springer Science+Business Media.
Taylor, Charles. 1995. Philosophical Arguments. Cambridge, MA: Harvard University Press.
Taylor, John G., and Akila Weerapana. 2009. Principles of Economics (7th edn). Boston, MA: Cengage.
Teubal, Morris, Dominique Foray, Moshe Justman, and Ehud Zuscovitch. 1995. Technological Infrastructure
Policy: An International Perspective. Dordrecht: Springer Science+Business Media.
Tirole, Jean. 1988. The Theory of Industrial Organization. Cambridge, MA: MIT Press.
Van Schewick, Barbara. 2010. Internet Architecture and Innovation. Cambridge, MA: MIT Press.
Vaidhyanathan, Siva. 2001. Copyrights and Copywrongs: The Rise of Intellectual Property and How it Threatens
Creativity. New York: NYU Press.
Varian, Hal R. 1985. ‘Price Discrimination and Social Welfare,’ 75 American Economic Review 870–75.
Varian, Hal R. 1996. ‘Differential Pricing and Efficiency,’ 1 First Monday, available at http://ojphi.org/ojs/index.
php/fm/article/view/473/829 (last visited Jan. 1, 2012).

Vetter, Greg R. 2007. ‘Open Source Licensing and Scattering Opportunism in Software Standards,’ 47 Boston
College Law Review 225–891.
Von Hippel, Eric. 2005. Democratizing Innovation. Cambridge, MA: MIT Press.
Weiser, Phillip J. 2003. ‘The Internet, Innovation, and Intellectual Property Policy,’ 103 Columbia Law Review
534–613.
Williamson, Oliver E. 1971. ‘The Vertical Integration of Production: Market Failure Considerations,’ 61
Yen, Alfred C. 2003. ‘Eldred, the First Amendment, and Aggressive Copyright Claims,’ 40 Houston Law Review
673.
Yoshida, Yoshihiro. 2000. ‘Third-Degree Price Discrimination in Input Markets: Output and Welfare,’ 90
Yu, Peter K. 2005. ‘Intellectual Property and the Information Ecosystem,’ 2005 Michigan State Law Review
1–20.

PART V
IP, DEVELOPMENT, AND

INTERNATIONAL TRADE

22. Creative development: copyright and emerging
creative industries
Sean A. Pager* 13
Contents
I. Introduction
II. Emerging Creative Industries in the Global South
A. Home-Grown Media and Cultural Progress
B. Digital Technology Overthrows Cultural Hegemony
C. Nigeria
D. India
E. China
III. State Support for Creative Industries
A. Copyright
1. Production incentives
2. Allocative efficiency
3. Private ordering
4. Autonomy
5. Prior commitments and path dependencies
6. Summing up
B. State Patronage and Other Subsidy Mechanisms
1. Direct patronage
2. Collective licensing, rewards and indirect subsidies
C. Commons-Based Creativity Models
1. Non-commercial creativity
2. Alternative business models
3. Copyright futility
IV. Copyright and Creative Development: Reviewing the Record
A. Thriving or Surviving?
1. Pirate innovators
2. Depressed revenues
3. Copyright’s positive role
B. Pursuing ‘Copyright’ by Other Means
C. The Hidden Drawbacks of Copyright Alternatives
1. Lack of scalability
2. Abuse of power
3. Distorting effects of alternative revenue sources
* Professor of Law and Associate Director of the Intellectual Property, Information and
Communications Law Program, Michigan State University College of Law.
582

Copyright and emerging creative industries 583
D. Distribution of Benefits
E. Context-Specific Copyright
F. The Formalization Imperative
G. Taking Stock
V. Copyright and Cultural Diversity
A. Diversity Effects in Developed Markets
1. Commodity culture
2. Market structure
3. Anticompetitive abuse
4. Effects of digitization
5. Restrictions on downstream creativity
B. Recontextualizing to the Development Context: Evidence from Case
Studies
C. Global Diversity
1. Overview of diversity concerns
2. The role of copyright
3. Cross-cultural remakes, parallel imports and versioning
VI. Conclusion
References
I. INTRODUCTION
Western commentators celebrate the transformative potential of digital media in empow-

ering amateur creativity—blogs, remixes, mash-ups, and the like. Yet, the effects of digital
technologies on commercial creativity in the developing world have been just as profound.
Nigeria went from producing an average of three films per year in the 1980s and early
1990s to producing roughly a thousand movies annually in the years since 2000 (Pager,
2012a). China has seen similarly exponential growth in home-grown content (Priest,
2016). Indian content producers have exploited new media to capture a global following
(Athique, 2008). Commercial content industries have proliferated elsewhere on a smaller
scale across the developing world (Pager, 2012a).
Such burgeoning creative content industries hold powerful implications for debates
over global intellectual property (IP). The ‘IP and Development’ debate has tended to
focus on patent law to the exclusion of other IP rights and to privilege the perspective
of foreign investors over domestic innovators (Schultz and van Gelder, 2008). When
copyright does feature in development debates, it is generally seen as a negative influence,
an access barrier to knowledge rather than an engine of creativity (Pager, 2012a). The
content industries emerging in the Global South present a very different context in which
to frame debates over global copyright policy. But what exactly is the role that copyright
plays in the development of such industries?
When Western commentators acknowledge the existence of creative industries in the
developing world, they have a tendency to project upon them narratives shaped by ideologi-
cal debates at home. Copyright proponents emphasize how piracy harms such emerging
content industries, presenting them as windows into a dystopian future that may materialize
in developed markets next (Priest, 2014). Copyright skeptics unsurprisingly paint an inverse

picture: They argue that the ability of these industries to prosper despite high rates of piracy
shows that copyright incentives are no longer needed (Anderson, 2009).
Drawing on case studies of emerging content industries in China, India, and Nigeria,
this chapter offers a more nuanced account of the interaction between copyright law and
development: While copyright law may not be a sine qua non for creative development,
it offers important advantages over alternative funding models. Copyright’s benefits may
be realized in different ways and to varying degrees as content industries develop. In
this respect, the account here builds on earlier narratives that posited a ‘crossover point’
whereby countries reach a stage of development at which the net benefits of copyright
protection outweigh the costs (Yu, 2007). More recent commentary has suggested that
there may be multiple ‘crossover points’ as countries develop (Yu, 2009).1 This chapter
adds a further refinement: Rather than emphasizing specific inflection points at which
copyright protection takes off, it suggests copyright formalization functions as more of
an asymptotic process. The logic and external benefits of copyright formalization exert
a gradual pull over time, and embrace of formal copyright norms, when it comes, often
remains partial, selective, and contextually contingent.
The remainder of this chapter proceeds as follows: Section II provides an overview
of creative industries in the developing world and then a closer look at film and music
industries in three countries—Nigeria, India, and China.2 Section III surveys competing
models by which governments subsidize creative production: copyright, state patronage,
as well as commons-based alternatives. Section IV examines how de facto commons
models have functioned in the development context based on the case studies, and
highlights the comparative advantages of copyright. Section V considers how copyright
law affects cultural diversity both at the national and global levels. Section VI concludes.
II. EMERGING CREATIVE INDUSTRIES IN THE GLOBAL

SOUTH
A. Home-Grown Media and Cultural Progress
A wealth of scholarship and policy studies testify to the benefits of home-grown creative
industries. Such industries—including music, film, television, and publishing—contribute
to both economic and cultural development in a variety of ways. As pillars of the knowl-
edge economy, creative industries promise well-paying jobs, above-average economic
growth, sustainable development, and positive effects on the balance of trade (UNCTAD,
2010). Other indirect benefits include boosts to tourism, potential marketing tie-ins, and
a reversal of brain drain (Pager, 2011).
1
Yu describes how attitudes to IP rights can differ even within a single country due to
disparities in regional development. He also points to sectoral disparities that can arise as different
content-producing industries adopt diverging views of copyright’s value (2007, 2009). Indeed, the
case study of Nollywood in this chapter shows that such sharply diverging viewpoints can arise
even within a single industry.
2
This chapter will use the terms ‘content industries,’ ‘creative industries,’ and ‘cultural indus-
tries’ largely interchangeably and will focus primarily on music and film industries.

The cultural benefits of such home-grown industries are equally compelling. Beyond
the intrinsic value of cultural innovation in expanding horizons, provoking insight,
and enriching cultural heritage, such industries also makes vital contributions to public
discourse and democratic governance (Netanel, 1996; Baker, 2002), foster national cohe-
siveness and social inclusion (Voon, 2007) and nourish personal autonomy and identity
formation in ways that facilitate human flourishing (Sunder, 2012).
What role does copyright play in fostering such development? According to the
formulation enshrined in the US Constitution and influential elsewhere, copyright law
aims to further ‘progress’ in the creative arts. On its face, ‘progress’ sounds remarkably
consonant, if not synonymous with ‘development.’ Yet, just as the preceding paragraphs
demonstrate the multiple dimensions on which economic and cultural development can
be measured, so, too, ‘progress’ from a copyright standpoint is susceptible to competing
normative interpretations.
Some question whether copyright is the right vehicle to achieve developmental aims at
all. Skeptics have cast doubt on the extent to which extrinsic incentives for commercial
culture industries are justified in the digital age and argue that copyright blocks more
expression than it fosters.3 From the standpoint of developing countries, copyright
faces additional objections: Some worry that the benefits of copyright protection flow
primarily to foreign producers (Story, 2003).4 Others fret that copyright harms cultural
diversity, channeling production toward globally homogenized expression at the expense
of more authentic local voices.5 Here, the question may not be does copyright foster
development, but rather development of what?
Copyright doctrine itself provides little help in answering this question. Ever since the
Bleistein decision (Bleistein v. Donaldson Lithographing Co., 188 U.S. 239 (1903)), copy-
right law has espoused an ethos of non-discrimination. In contrast to patent law, where
the non-obviousness test explicitly probes inventive achievement, copyright doctrine
deliberately eschews assessment of artistic merit (Walker and Depoorter, 2015). Such
assessments are unavoidably subjective and fraught with value judgments. For example,
how do we decide which furthers ‘progress’ more: a commercial feature film or an amateur
video mash-up? Are we concerned only with the number of eyeballs impacted, or do we
care how deeply the works affect viewers? If the latter, should we measure short-term
effects or long-term? Do we only care about consumption, or is participation in produc-
tion an important value (Skladany, 2008)? Different answers to these questions dictate
diverging implications for copyright policy.
B. Digital Technology Overthrows Cultural Hegemony
Arguments over copyright policy are also embedded in a broader discourse of imperial-
ism. Copyright protection has long been seen as the handmaiden of Western hegemony
(Story, 2003). Critics denounce copyright law as an alien appendage that tramples on
3
See infra Section III-C.
4
Policy-makers may also see copyright protection as an expensive distraction from other
more pressing social needs. Such comparative balancing of social priorities, however, lies outside
the scope of this chapter.
5
See infra Section V.

local traditions (Gana, 1995) and unfairly privileges Western-centric forms of creativity
(Coombe, 1998). Others decry the Agreement on Trade-Related Aspects of Intellectual
Property Rights (TRIPS Agreement) and the harmonization of global IP law as a ‘royalty
extraction vehicle’ designed to enrich powerful multinational corporations at the expense
of poor people in the Global South (Pager, 2006). This ‘neocolonial’ narrative posits
innovation as almost exclusively the product of developed countries, with the developing
world relegated to the role of consumer, or, at best, the source of raw inputs from which
information goods would be fashioned elsewhere (Boyle, 1997).
In the realm of creative expression, the discourse of cultural imperialism reinforces
such neocolonial narratives. Dominant messages emanating from the ‘center’ through
the vehicle of American/Western popular culture are seen as passively consumed in the
‘periphery’ (Pager, 2012c). Marxist diagnoses ascribing such one-sided cultural flows to
capitalist hegemony were supplemented by empirical models by media economists which
showed that the country with the largest home market—which for most of the twentieth
century meant the United States—was destined to dominate global markets (Wildman
and Siwek, 1988). Either way, the implications were the same: the market was rigged.
Yet, further work by media scholars has shown that market size alone is not the only
relevant variable. Audiences everywhere prefer locally produced expression, but the
strength of this preference varies according to the audience’s ‘cultural distance’ from
the dominant global supplier. Accordingly, for developing countries with cultural tradi-
tions and present-day contexts radically different from the West, a substantial ‘cultural
discount’ applies to imports of global culture (Fu, 2012). In other words, the demand in
developing countries for local content is strong. All that was missing was supply (Schultz,
2012).
Digital technologies have changed that. While some feared digital communications
would merely supply a fatter set of pipes through which popular culture exports could
penetrate the developing world (Schiller, 1999), in fact, digital media have served to level
the playing field. The case studies that follow illustrate the transformative effects of digital
technologies in democratizing global creative content production. Nigeria, India, and
China are each home to burgeoning content industries whose successes potentially recast
the terms of global debate.6 Yet, the question remains: Are they thriving because of or
in spite of copyright law?
C. Nigeria
The origin of Nigeria’s video film industry—Nollywood—reflects a confluence of factors

specific to the Nigeria context: the macro-economic crisis of the late 1980s, the closing of
cinema houses due to urban violence, the collapse of Nigeria television production (Miller,
2016). Yet, there is also an element of technological determinism to the tale: Earlier efforts
to market filmed productions of Yoruba folk theater—perhaps Nollywood’s closest
6
Nigeria, India, and China are all countries with large domestic markets. Nigeria and India,
in particular, are global exporters of popular culture and regional hegemons. Accordingly, one
would expect them to have strong creative industries. Yet, as we will see, even with these advantages,
they have sometimes struggled to reach their potential.

cultural antecedent—had foundered on the high costs and logistical hurdles associated
with celluloid films. Nollywood’s shift to video media—initially analog VHS tapes then
later digital VCD discs—proved a game-changer. Perhaps not coincidentally, Kenneth
Nnebue, an electronics dealer who imported VCR equipment and blank videotapes, is
credited with discovering the medium’s potential.7 Nnebue bet that his tape stock would
sell better filled with content than empty. His 1992 hit film, Living in Bondage, galvanized
attention, and a direct-to-video film industry sprang up almost overnight (Miller, 2016).
Within a decade, Nollywood had grown to become Africa’s dominant film producer,
churning out hundreds of films each year, watched by millions daily across Africa (Pager,
2012c).8 By volume of production, it is widely hailed as the world’s second most prolific
film industry.9
As a world-leading fully digital film industry, Nollywood exemplifies the potential
for developing countries to leapfrog outdated technologies (Arewa, 2015). With annual
revenues numbering in the hundreds of millions (in US dollars), Nollywood has become
the country’s largest private employer (Economist, 2010).10 It serves as a ‘model of entre-
preneurial achievement’ in a country plagued by corruption and rent-seeking (McCall,
2002). Nollywood’s success has inspired similar film industries in other African countries
and among expatriate Nigerian communities (Pager, 2012a; Haynes, 2013).
The cultural significance of Nollywood is equally notable. Africa has a deeply ingrained
storytelling tradition, but has long lacked a mass media vehicle to harness its creative ener-
gies. For the first time, African stories told by Africans can be shared by audiences across
the continent (McCall, 2002). That Nigerian films regularly outsell Hollywood imports
made with far higher budgets and more sophisticated production values testifies to the
hunger of African consumers for a genuinely popular medium of expression (Schultz,
2012). African diasporal communities overseas have proven equally avid consumers for
whom watching Nollywood film provides a cultural connection to Africa (Miller, 2016).
Nollywood films are made by a decentralized network of producers who operate at
extremely low cost using rudimentary equipment. A budget of $50,000 and production
calendar of four weeks from script to final release are not uncommon. Over 90 percent
of revenues come from sales of physical media routed through four central market hubs
and then resold across Nigeria and beyond. Films are sold for roughly $2, and sales
average anywhere from 50,000 to 200,000 authorized copies per film, with the occasional
blockbuster surpassing one million copies (Miller, 2016).
7
While VCR technologies were widely available in Nigeria by the mid-late 1980s, the cost of
such technologies plummeted in the 1990s, and, in particular, massive quantities of blank tapes
became available at cut-rate prices as dealers in developed markets shifted to more advanced
technologies (Miller, 2016).
8
Nigerian video film production is as diverse as the country itself. Each of Nigeria’s main ethnic
groups produces films in their local language. However, Nollywood has come to signify the English-
language films produced in Southern Nigeria, which enjoy the widest distribution (Miller, 2016).
9
A widely cited 2009 UNESCO report ranked Nollywood second only to Bollywood;
however, this claim hinges on a statistical anomaly: Nollywood’s count is based on video film
production, whereas other countries count only films released for theatrical distribution and omit
direct-to-video films (Bud, 2014).
10
The industry also generates indirect benefits such as road construction by film crews in rural
villages (McCall, 2004).

Funding and distribution of Nollywood films is dominated by shadowy guilds of

‘marketers’ who operate through informal networks that originally served to smuggle
pirated copies of foreign movies. As a result, although it has grown into a billion dollar,
global industry, Nollywood still operates almost entirely through informal mechanisms.
Cash predominates over credit. Trust relationships replace contracts. Copyright formali-
ties are ignored. Instead, Nollywood guilds enforce order through informal disciplinary
measures, and actively discourage recourse to formal legal institutions. Accurate records
of sales and revenues are impossible to obtain. Nor it is easy to establish who holds the
rights to a given title; fraudulent sales agents abound (Miller, 2016).
While Nollywood’s reliance on erstwhile pirate networks gave it far greater reach than
conventional distribution channels could have achieved, piracy today is the industry’s
Achilles heel (Paulson, 2012). Unauthorized copies of Nollywood films usually appear
within a couple weeks and cannibalize sales. Anywhere from 60–80 percent of revenues
may be diverted in this fashion. Pirate sales likely account for an even greater percentage
of international revenues (Miller, 2016). Unauthorized distributions of Nollywood films
occur even in developed country markets that have functioning copyright regimes (Pager,
2012a).
Because filmmakers reap only a fraction of the total revenue that their movies gener-
ate, the industry suffers from chronic underinvestment. Lack of copyright protection
also introduces perverse incentives. Filmmakers are forced to mass-produce films at
a breakneck schedule to stay ahead of the pirates. Slap-dash productions featuring
formulaic plots, wooden acting, and crude production values are the predictable result
(Pager, 2012a).
In recent years, a group of successful filmmakers has sought to launch a ‘New Nollywood’
comprising more ambitious, higher budget films, with glossier production values, splashy
marketing, and international financing (Pratt, 2015; Miller, 2016). The opening of high-
end multiplex cinemas in Nigeria’s largest cities has allowed New Nollywood films to tap
theatrical exhibition revenues. These films also travel the international film festival circuit
and are increasingly shown on airline flights, satellite TV, Netflix, and even at London
cinemas. An online service, iROKOtv, backed by Western private equity funds, offers an
extensive film catalog to subscribers in over a dozen countries (Miller, 2016).
The reality remains, however, that these alternative revenue sources do not suffice to
cover the costs of production. New Nollywood, like old Nollywood, remains dependent
on revenues from sales of physical media which the marketers control. Attempts to
establish alternative distribution channels in Nigeria have thus far failed. As a result, New
Nollywood remains a largely marginal presence, and Nigerian film production remains
centered on the high volume, direct-to-video model, whose financing and distribution
remains firmly under the control of the marketers’ guilds (Miller, 2016).
Some have argued internet distribution offers a means to bypass the marketers’
stranglehold over the industry (Pratt, 2015). However, although internet penetration
rates in Nigeria have risen in recent years, low bandwidth speeds and high data costs limit
the domestic potential for video distribution; the online video market remains primarily
comprised of diasporal communities (Miller, 2016).
The attitude of Nigeria’s government has been ambivalent. Initially ashamed and
embarrassed by the industry’s vulgarity, Nigeria’s leaders gradually came to take pride in
Nollywood’s accomplishment and also eyed the industry as a juicy tax source. However,

government initiatives to formalize the distribution sector and curtail piracy have largely
failed (Bud, 2014, 2016; Lobato, 2012).
By contrast, Nigeria’s music industry has made a more successful transition to digital
distribution platforms, taking advantage of the lower bandwidth requirements for music.
Piracy has devastated sales of physical media, but mobile music offers a promising replace-
ment. Working through telecommunications companies facilitates online payment—an
otherwise difficult hurdle in a country where credit cards are rare, albeit at the cost of 70
percent of revenues.
Licensing of ring-back tones alone comprises a $150+ million market (Sanchez,
2014).11 Nigeria’s revitalized collecting society, Coson, has also stepped up enforcement
and licensing of music performance rights, and the extension of YouTube’s Partner
Program to Nigeria has unlocked additional revenues from digital streaming. However,
despite this rebound in revenues from music recordings, concerts and endorsement deals
remain the largest source of income for musicians (Rutschman, 2015).
Nigeria’s music industry has grown to become Africa’s largest and—in contrast to
Nollywood—has attracted investment interest and distribution deals from Western music
industry multinationals (Ingham, 2016). Hip-hop icon, Jay Z, has also reportedly explored
investment opportunities in Nigeria (Abulude, 2015). Meanwhile, i-Tunes’ debut of a
Nigerian version of its online store has been matched by a pair of Western-backed local
music platforms, Freeme Digital and iRoking, launched in recent years (Rutschman, 2015).
D. India
On its face, India’s film industry presents a very different context than Nollywood. India
has had a long, successful history producing celluloid films, and the industry remains
focused on theatrical exhibition, with video sales largely an afterthought. India’s annual
production of over a thousand films makes it the world’s most prolific industry, and over
three billion box office admissions each year give it a claim to the world’s largest audience
(Banerjee, 2016). Movies are central to public life in India. Indian films are also avidly
consumed not only across South Asia and among its diasporal communities, but also by
native populations in much of Asia and Africa, for whom the films’ wholesome family
values supply an attractive alternative to Hollywood (Rajadhyaksha, 2008; Liang and
Sundaram, 2011).
Bollywood’s global reach may long predate Nollywood’s, and its theatrical orientation
contrasts with Nollywood’s direct-to-video model, but on closer inspection, the two
industries have much in common as decentralized, low cost, high volume producers.
Bollywood properly refers only to the Hindi-speaking film industry based in Mumbai,
whose films circulate primarily in North India, just as Nollywood usually refers to
English-language films from Southern Nigeria. Both represent the global face of their
country’s ethnically and regionally fragmented film industries.12
11
Ring-back tones are personalized music clips played to an incoming caller in lieu of a ring-
ing sound.
12
Hindi films only command 40 percent of the domestic market, and several of India’s
regional film industries are commercially significant, global exporters as well (Pager, 2011).

The parallels between the two become even stronger if one compares Nollywood’s
current position to Bollywood three decades ago. In the 1980s, Bollywood was a largely
informal industry revolving around loosely organized, highly decentralized studios that
churned out an endless stream of musical melodramas. Fueled by ‘dubious money’ sup-
plied by gangsters and tax dodging money-launderers, the industry was characterized by
shambolic management and murky accounting (Athique, 2008; Moullier, 2007). Little
heed was paid to copyright norms. Story lines were widely recycled, often taken from
successful films produced elsewhere. Sheltered from foreign competition by protectionist
barriers and restricted in its ability to export, the industry relied on a captive domestic
audience to consume its often formulaic output (Pager, 2011).
Things began to change, however, with the spread of VCR technologies in the 1980s,
which suddenly exposed Indian film producers to competition from widely available
pirated videotapes. Concerned over mounting losses to piracy, the industry mobilized to
place copyright enforcement on the policy agenda. India’s national government had been
long indifferent to Bollywood, except to exploit it as a tax cow. IP law was also generally
suspect, seen as an imperialist imposition and viewed through a patent-centric prism
(Banerjee, 2016).
Local governments have proved more amenable to industry concerns, however, espe-
cially in the South. Police conduct sweeping anti-piracy raids to accompany big releases of
local films. The effective result is akin to the guild-created window in Nollywood: a short
period to recoup investment at the box office. Some state governments have also pressed
anti-gangster legislation into service against commercial pirates (Liang and Sundaram,
2011; Scaria, 2014).13
Despite initial skepticism toward IP rights, the national government eventually
responded to industry lobbying by passing more stringent laws, modernizing the
Copyright Act in 1994 and 1999, and stepping up enforcement. A crackdown on cable
piracy, in particular, bore fruit, and television became an important revenue source
(Telang and Waldfogel, 2014). The recent shift to digital distribution has made further
inroads against piracy, by allowing nationwide release of blockbuster movies, thus avoid-
ing problematic delays in rural distribution of popular movies that created openings for
pirated distribution as the default provider.14
Other developments around the turn of the millennium further enhanced the com-
mercial prospects of the film industry. The government’s 2001 decision to grant formal
industry status to the film industry enabled it for the first time to turn to conventional
sources of finance. ‘Corporatization’ became the watchword of the day, as industry
leaders worked to attract investors by putting their operations on a more professional
footing and taking strides toward horizontal and vertical integration (Moullier, 2007;
Ganti, 2012).
The development of modern shopping malls, encouraged by tax incentives, also led
to investment in high-end, multiplex theaters that catered to urban professionals. By
13
State enforcement initiatives are often organized along starkly parochial lines, targeting only
piracy of local films, while otherwise ignoring trade in illicit media (Liang and Sundaram, 2011).
14
Digital distribution also has allowed single-use copies to be sent to individual theaters,
enabling watermarking to trace the source of pirated copies and thereby break up camcording
rackets (Liang and Sundaram, 2011).

providing a superior theatrical experience, the multiplexes could charge much higher
admission fees, yielding far greater revenues (Pager, 2011). Relaxation of trade restrictions
also allowed the industry to develop profitable export markets tapping affluent Indian
diasporal communities. The combination of these two more sophisticated audiences led
Indian filmmakers to produce more ambitious, high budget productions that ventured
beyond the usual formulas (Pager, 2011). These developments have spurred a renaissance
in Indian filmmaking.15
The industry has grown to $2.8 billion annual revenues, with double-digit growth
in recent years forecast to continue (Deloitte, 2016). Online piracy rates are rising as
broadband speed and penetration increases (Liang and Sundaram, 2011). However,
the industry’s main revenue source—theatrical exhibition—remains relatively insulated
due to the distinctive nature of India’s movie-going culture based on active audience
participation (Hammer, 2014). As a result, pirated wares are not a direct substitute for
the theatrical experience.16
The Indian music industry has had a less illustrious past. Recorded music in India has
long been viewed primarily as an adjunct to film production (Booth, 2015). Most popular
films feature extended musical segments and rely on catchy new music to help market the
movie. Moreover, 70 percent of album sales are based on Indian movie soundtracks, and
Indian film music constitutes its own distinctive genre of popular music that dominates
industry production (Liang and Sundaram, 2011).
Toward the end of the twentieth century, the music industry had begun to exploit new
playback media by lowering prices and expanding its distribution (Athique, 2008). Sales
duly exploded and rapid growth ensued. As profits rose, newcomers joined the industry
and pioneered new markets. Producers diversified into neglected genres besides film
music, and began to record music in languages other than Hindi. For a little more than
a golden decade, the industry began to view recorded music as an independent revenue
source, worthy of investment in its own right (Booth, 2015; Liang and Sundaram, 2011).
The advent of digital piracy abruptly reversed these trends. As pirates and producers of
unauthorized ‘version’ recordings increasingly diverted revenues, the industry retrenched
sharply. Language and genre diversity suffered, as production of recorded music largely
reverted to its prior diminished status as a loss-leading investment designed to sell film
tickets (Booth, 2015).
E. China
Whereas the Indian and Nigerian content industries largely comprise private actors
functioning autonomously from the government, China’s modern content industries
operate in a very different context. Although China has undertaken a remarkable program
of privatization and liberalization in recent decades, the state retains considerable control
15
Indian filmmakers did not, however, break entirely with the formulas of the past. Rather,
the industry bifurcated between more ambitious ‘crossover’ films that catered to wealthy urban
and expatriate audiences and a reinvigoration of the old-style musical melodramas that had long
thrilled rural masses, now filmed with ever-more elaborate production values and exotic locales.
16
Because the overseas market is less theatrically based, it is more directly affected by piracy
(Scaria, 2014).

over the media. Key distribution channels remain state monopolies, and state censorship
continues to impose restraints on expressive content. On its face, the Chinese government
remains committed to fostering a ‘quality culture’ that will instill the correct moral values
in its citizenry. At the same time, the need to cater to popular demand and fend off the
competitive pressure from foreign media has led to a progressive liberalization of censor-
ship standards (Priest, 2015a).
As state subsidies were gradually withdrawn in the 1990s, China’s culture industries have
had to manage the transition from churning out state propaganda to courting audiences
with crowd-pleasing fare. Private investment has flown into a host of new enterprises, and
the content industries have enjoyed considerable success in recent years (Montgomery and
Priest, 2016). The film industry has benefited from extensive construction of state-of-
the-art cinemas in urban centers that have made going to the movies a fashionable leisure
activity for China’s newly affluent professionals, commanding box office ticket prices as
high as $26 (Priest, 2015a). The Chinese box office is now the second largest in the world,
and Chinese domestic films have claimed an ample share of the proceeds, in recent years
rivaling the take of Hollywood’s blockbusters. As explained below, the music industry has
had a tougher road, but even it has perked up recently (Song, 2016).
Censorship aside, however, the biggest challenge that China’s content industries face
remains extraordinarily high rates of domestic piracy. Chinese consumers have long
relied on informal distribution mechanisms to access popular media that may not have
been available through legitimate channels. With almost ubiquitous access to pirated
wares on the internet, they have grown accustomed to obtaining all manner of crea-
tive media instantly, free of charge. A ‘culture of unauthorized reuse’ pervades even
commercial enterprises, with amusement parks, media firms, and even state television
making liberal use of unlicensed creative content (Montgomery and Priest, 2016; Liu,
2015).
China’s content industries have therefore struggled to devise business models that allow
them to appropriate revenue in a climate of pervasive piracy. Unsurprisingly, performance
models—theatrical exhibition for films; concerts for music—have provided the main
source of income as these delivery models are subject to physical exclusion (admission
controls) and offer a marketable experience that is distinguishable from pirated copies
consumed at home. While profitable, these models are subject to capacity constraints, as
Section IV-C(1) will elaborate.
The digital content market has faced even graver challenges. The market for recorded
music all-but imploded in the first decade of the millennium, with sales and investment
in new music plummeting by as much as 90 percent (Liu, 2015). Until recently, the only
way that ‘Chinese musicians and music companies [could] actually make money from
music sales [came from] two narrow markets: ringback tones sales [for mobile phones]
and overseas sales’ (Liu, 2015).
Until recently, copyright law played little or no role in China’s content industries.
However, this has begun to shift. China has made considerable improvement in its copy-
right infrastructure in recent decades. Much of its efforts were undertaken in response to
treaty obligations and external pressure. Public enforcement campaigns against piracy
have often served as a form of kabuki theater performed for foreign consumption, with
showy, albeit ultimately ineffective raids to seize and destroy illicit material (Montgomery
and Priest, 2016).

Private enforcement of copyright law, however, tells a different story. China has become
the most IP-litigious society in the world, and almost 98 percent of the plaintiffs have been
Chinese. Chinese rightsholders have thus expressed an enthusiastic vote of confidence in
the benefits of IP law (Priest, 2014). Chinese policy-makers have also come to recognize
that copyright law has a role to play in building the strong creative content industries that
they see as underpinning China’s soft power (Priest, 2015a). As a result, copyright law
has increasingly become seen as a matter of domestic concern, rather than an unwanted
foreign imposition—the long-heralded ‘crossover point’ for IP development (Yu, 2007).
Furthermore, in what some commentators have hailed as a ‘watershed moment
for China’s cultural and creative industries,’ the last few years have witnessed a major
transformation in China’s digital content landscape (Montgomery and Priest, 2016).
China’s leading online music and video streaming platforms have begun to purge their
sites of pirated works.17 Having long attracted traffic by hosting a vast sea of unlicensed
content of variable quality, the websites have shifted strategies and are now focusing on
negotiating exclusive licenses for professionally produced content, as well as investing in
production in-house. Having purchased exclusive rights to such valuable content, Chinese
firms are also increasingly turning to litigation to enforce them, with all of the leading
sites vigorously prosecuting claims and counterclaims against one another (Montgomery
and Priest, 2016).
Copyright’s role throughout these case studies remains, at best, ambiguous. To resolve
this ambiguity, we should first clarify as a theoretical matter the economic function that
copyright is expected to perform and to highlight its advantages and disadvantages as
compared to alternative models.
III. STATE SUPPORT FOR CREATIVE INDUSTRIES
Creative markets are risky; entry often requires high capital investments that are subject
to uncertain returns. Because cultural mass media represent public goods likely to be
produced at suboptimal levels through market mechanisms, the conventional wisdom
deems that public subsidies are merited. Yet, questions remain: How should government
support artistic production? What kind? Who decides? Who pays? Who bears the risks?
In part, the answer to these questions is determined by our choice of subsidy mechanism.
A. Copyright
Copyright law gives authors exclusive rights in their expressively creative works. These rights
function to confer an implicit subsidy through decentralized, market-based exchanges.
Because the benefits of copyright are tied to commercialization, authors and publishers
bear the costs and risks of creative investment in advance of any market returns.
17
Several factors likely account for this turn to licensing. Government edicts, pressure from
advertisers, litigation by rightsholders, and even the prospect of stock market flotations in Western
bourses have all been cited as possible explanations; Priest credits advertisers as playing the decisive
role (Priest, 2015b).

1. Production incentives

Most copyright scholars justify copyright protection based on some variation of the
incentive-access theory (Bracha and Syed, 2014a). This theory begins by observing that
many creative works are expensive to make, but cheap to copy. Because copyists do not
bear the initial costs of creativity, they could drive the price of the work down to the mar-
ginal cost of producing copies. If so, the original creator would be unable to recoup her
initial investment in creating the work. Given such prospect, the theory holds that creators
would have inadequate incentive to produce the works in the first place.18 By providing
creators a legal entitlement to prevent a range of market-impairing copying, copyright
restores the incentives to invest in creative production to a socially optimal level.19
Copyright law protection is not cost-free. Its most immediate costs inhere in the higher
prices that copyright exclusivity facilitates, thereby enabling creators to recoup their
up-front investments. Such costs are primarily borne by end users who must pay prices
set above marginal cost. Copyright thus famously functions as ‘a tax on readers for the
purpose of giving a bounty to writers’ (Macaulay, [1841] 1915).
The costs of copyright are a particular concern to developing countries. The ‘copyright
tax’ may deter purchases of copyright works, leading to reduced access to information
goods that inhibits educational and scientific progress (Kapczynski, 2008) and impede
cultural literacy (Liang, 2009). There are also distributional justice concerns (Chon, 2006).
Developing countries are invariably net importers of copyrighted material. Copyright
benefits tend to flow disproportionately to a handful of global superstars, who are often
foreigners. The siphoning of royalty payments offshore can trigger balance of payment
concerns (Pager, 2012a).
Furthermore, copyright exclusivity is more than just a tax; it confers exclusionary rights
that creators can use to further restrict access, even where consumers are willing to pay
the price. High piracy rates and fear of reimportation often make rightsholders reluctant
to authorize distribution in developing countries. Copyright exclusivity can also deter
secondary creators from engaging in creative expression that builds upon the copyrighted
original, resulting in chilled speech and reduced rates of secondary innovation.20
Copyright’s costs can be mitigated through carefully tailored limitations and excep-
tions. However, even in the best scenario, some degree of restricted access remains
inevitable (Bracha and Syed, 2014a). Moreover, because developing countries often fail
to make effective use of the policy levers available to them, in practice, the results may fall
even further from the ideal (Kaminski, 2014).
18
Recently, copyright skeptics have challenged the utility of copyright incentives, citing evi-
dence that creativity is intrinsically motivated. These arguments are explored in Section III-C. It is
also worth noting that copyright also serves to protect more subjective interests of creators, includ-
ing the desire for attribution and control over the work’s integrity. Although the extent to which
such ‘moral rights’ are operationalized separately within copyright law varies, protection of such
authorial interests could further encourage creative investment (Hansmann and Santilli, 1997).
19
Note that copyright incentives govern the level of creative investment, which is a different
metric than the actual number of works created. As will be argued in Section V, increased copy-
right protection may actually result in a smaller number of high value works being produced, as
compared to the high volume, low investment equilibrium that prevails in the absence of copyright.
20
Some argue that these dynamic costs have grown particularly intolerable in the digital age as
new forms of follow-on creativity have proliferated. See Section V-A(5).

2. Allocative efficiency

By encouraging authors and publishers to monetize demand for creative works through
commercialization, copyright engenders a set of market feedbacks that, in turn, influ-
ence future investments. In this way, copyright markets respond to popular demand and
allocate creative investments toward socially desirable production through decentralized
exchanges.21 Some have sought to extend copyright’s allocation function broadly into
derivative markets to monetize all productive uses to which creative works are put,
thereby allowing the market to properly internalize the full spectrum of societal demand
(Goldstein, 2003). Such a broad vision of copyright, however, conflicts with countervail-
ing values that favor access to information and breathing space for transformative uses
(Netanel, 1996).
3. Private ordering

By allowing authors to transfer their copyrights to publishers and distributors, the
copyright system has facilitated the growth of sophisticated ‘culture industries.’ Such
intermediaries greatly increase the efficiency of cultural markets by exploiting economies
of scale to develop specialized editorial, production, distribution, marketing, and enforce-
ment capabilities. Intermediaries also furnish an important source of venture capital to
front the costs of commercialization. Investment across a portfolio of works functions
as a crucial form of risk management whereby a few commercial ‘hits’ subsidize many
‘misses’ (Anderson, 2006).22 Accordingly, some scholars have explained copyright incen-
tives as directed toward publishers, not authors (Barnett, 2013; Cohen, 2011), with a goal
of inducing commercialization rather than creativity per se.23
More broadly, copyright law helps to sustain concentrated creative infrastructures with
the capabilities to support a wide range of creative projects. Such ‘creative clusters’ could
doubtless function in the absence of copyright, relying on contracts and other private law
norms. However, in reducing the risks associated with capital-intensive creative invest-
ments, copyright law has encouraged their development.
Copyright’s risk-reduction functions go beyond its protection against copying.
Copyright law provides a set of building blocks around which to structure transactions.
Its statutorily defined cluster of rights and obligations reduces uncertainty (Pager, 2012a).
The divisibility of copyright’s propertarian ‘bundle of sticks’ further facilitates private
ordering. Indeed, pre-sales of foreign distribution rights allocated by national territory
provide a crucial source of financing for independent film production (Dale, 1997).
Copyright law also provides default ownership and evidentiary rules that further enhance
predictability and facilitate creative collaborations (Sawicki and Casey, 2013; Lichtman,
21
In theory, the global reciprocity provided by international copyright treaties further rein-
forces the decentralized nature of copyright markets, allowing them to aggregate niche demand
from around the world. As we will see, however, some question the extent to which copyright
markets cater to the global long-tail. See infra Section V.
22
Individual creators may lack the means to bear such up-front costs or face potentially
ruinous losses should their creative investments fail. By allowing copyright entitlements to be
transferred from authors to commercial intermediaries, copyright law facilitates risk-shifting.
23
Other commentators dismiss such intermediaries as obsolete ‘gatekeepers’ that the digital
age is destined to bypass. See infra Section III-C.

2003).24 Finally, by providing a robust set of remedies that transcends contractual privity,
copyright law provides an added measure of security against defection.
4. Autonomy
A key feature of copyright markets is their ability to operate in relative autonomy from
state control. Such autonomy reflects both the decentralized nature of copyright’s market-
driven incentives and the privately ordered infrastructure that copyright helps to sustain.
Several facets of the copyright system reinforce its decentralized nature. Copyright
accrues automatically upon fixation, with no formalities required and a minimal original-
ity threshold. This lack of formalities reduces entry barriers to authors and insulates the
copyright system from the meddling hand of the state. To further minimize the potential
for the state to preempt market allocations, copyright law has built-in firewalls against
state interference. In particular, an ethos of non-discrimination pervades copyright doc-
trine, discouraging subjective judgments as to the merits of copyrighted works (Walker
and Depoorter, 2015).25
By facilitating cultural discourse through decentralized market institutions that operate
largely outside state control, copyright serves a political, as well as economic, function.
Indeed, Neil Netanel (1996) has argued that a system for funding autonomous cultural
production constitutes an essential prerequisite of liberal democracy. Subsequent work by
Netanel (1998) has underscored the particular importance of copyright law in sustaining
independent media in emerging democracies. Robert Merges has similarly emphasized the
role that copyright plays in sustaining a class of creative professionals who work under
conditions of relative creative autonomy that enable them to refine and perfect their craft
(Merges, 2011; Pager, 2015).
5. Prior commitments and path dependencies

Finally, regardless of any theoretical advantage that copyright law affords a priori, copy-
right protection is mandated by international treaty. While countries have discretion as to
the extent to which they enforce such norms, deviations from compliance may come at a
geopolitical cost. Conversely, there are benefits to operating within the copyright system
that derive from preexisting understandings and institutional arrangements. Creative
industries are governed globally by a complex system of private ordering arrangements
premised on formalized understandings of copyright law. Accordingly, to the extent
that individual countries establish a functioning copyright system that is compatible
with these preexisting arrangements, they enable national rightsholders to plug into this
global system and thereby to derive a host of network benefits. Advantages range from
reciprocal benefit-sharing between collective rights organizations to standardized chain
of title protocols for international distribution. As explored further below, the structural
imperative exerted by such preexisting institutions and practices exert a powerful pull
toward copyright formalization.26
24
In jurisdictions such as Nigeria where such formalities are lacking, uncertainties as to
ownership can have a deleterious effect on licensing (Miller, 2016).
25
Of course, markets impose their own biases on cultural production. Such concerns are
explored further in Section V.
26
See infra Section IV-F.

6. Summing up
The preceding frameworks for theorizing copyright to some degree overlap and are
complementary. For present purposes, therefore, we need not commit to any particular
normative theory. Rather, our understanding of copyright’s various aims and functions
provides a basis to compare copyright against competing paradigms to sustain creative
production.
B. State Patronage and Other Subsidy Mechanisms
Copyright is hardly the only mechanism to support creative production. Governments

can subsidize artistic production through a variety of means both direct and indirect.
1. Direct patronage

Historically, the most prevalent means of support for the arts was aristocratic patronage:
Wealthy patrons bestowed funding or employment upon creators. Such direct patron-
age mechanisms persist. However, governments have largely supplanted aristocrats as
the primary sponsor of direct patronage mechanisms.27 The Nigerian government’s
announcement of a $200 million ‘Special Entertainment Fund’ to underwrite the produc-
tion of Nollywood films represents a recent iteration in this tradition (Abulude, 2016).
State patronage represents a very different model than copyright law. Where copyright
confers its benefits indirectly through exclusive rights, patronage funding is conferred
directly. Instead of relying on decentralized market mechanisms to determine demand for
the finished work, patronage is typically allocated up front through competitive selection
processes that operate in top-down fashion. Finally, instead of ‘tax[ing] readers’ directly,
patronage regimes typically draw their funding from other sources, including general tax
revenues.
In some ways, patronage models are undoubtedly more appealing than copyright:
They avoid the access restrictions engendered by exclusive rights, thereby facilitating dis-
semination and enabling secondary creators to build on existing expression without fear
of infringement.28 Awarding grants directly to creators may allow them to bypass industry
gatekeepers (or at least to negotiate more favorable terms). By committing to funding in
advance of production, patronage regimes afford creators greater license to experiment
and take risks without the need to worry about recouping their investment in the market.
Indeed, the ability of patronage systems to correct for market biases in the copyright
system—awarding funding based on artistic merit rather than commercial appeal—may
be one of its most salient advantages. In theory, patronage systems also allow for greater
precision, directing funding toward cultural works responsive to social needs and avoiding
the overinclusiveness of the copyright regime, which protects works that may not need
external incentives.
27
Private funding from corporations and foundations also accounts for a significant source of
support for the arts. However, our focus here is on public subsidies.
28
It is worth noting that, in practice, patronage awards do not normally preclude copyright
exclusivity. Several commentators have proposed that state subsidies be leveraged to curtail copy-
right, however (Lee, 2012; Baker 2003), and the National Institute of Health has recently taken a
small step in this direction (Contreras, 2013).

Patronage regimes suffer from several drawbacks, however, that hamper their effective-
ness. Substituting government bureaucracies for market-driven mechanisms is almost
always a recipe for inefficiency, not least because markets are better at uncovering the
private information necessary to match supply to demand (Demsetz, 1969). The com-
parative advantage of copyright markets over top-down processes may be particularly
salient in the context of creative works whose value is notoriously difficult to appraise at
the time of creation.29 There is a long history of cultural innovation being rejected by
contemporary experts only to win acceptance by subsequent generations (Pager, 2012b;
Cowen, 2006). The relatively long duration of copyright terms allows innovative works to
benefit from such esthetic revisionism.
Patronage schemes are also vulnerable to other distortions and abuses. First, although
both copyright and patronage confer subsidies to creators, the nature of the subsidy is far
more visible in the patronage context, and it comes directly out of the state treasury. This
makes patronage funding vulnerable to budget cuts prompted either by fiscal austerity
or artistic philistinism.30 Second, patronage’s theoretical advantages in encouraging
dissemination are often not realized in practice. Patronage schemes tend to focus on
underwriting the creation of new works without necessarily ensuring that such works
reach actual audiences; creators whose costs have been covered in advance may also be
less incentivized to commercialize their work (Pager, 2011).31 Third, instead of spurring
artistic breakthroughs, committing funding up front can encourage self-indulgent auteurs
to produce esoteric works of questionable social value. It is notable, for example, that
increased state subsidization of the French and Italian film industries resulted in films
that not only underperformed at the box office, but also garnered a diminished share of
prestigious film festival prizes compared to prior decades (Pager, 2011).32
Furthermore, selecting works based on artistic merit rather than commercial appeal
pushes cultural production toward esthetic standards dictates by the cultural elites who
populate selection committees. In the development context, such elitist tendencies have
sometimes meant favoring foreign audiences over domestic ones. For example, in the early
post-colonial period, African governments funded celluloid films ‘as prestigious cultural
trophies to impress European elites.’ Such ‘embassy films’ were rarely seen by African
audiences (Pager, 2012b).33 Chinese filmmakers in the 1980s followed a similar pattern,
29
This comparative advantage becomes greater still where patronage funding is committed
in advance of production. In such cases, review committees must assess the comparative merits of
competing works without seeing them in a fully realized form. It is, of course, possible to allocate
funding ex post, evaluating completed works through a competitive prize system. However, in
practice, competitive prizes do not account for a significant share of cultural funding.
30
Drastic cuts in Nigerian media funding in the economic crisis years of the early 1990s
offers an example of the former (Miller, 2016). Newt Gingrich’s savaging of public arts funding
in the United States offers an example of the latter (Cowen, 2006). Developing economies may be
particularly prone to such fiscal upheavals.
31
By contrast, the market-centric structure of copyright exclusivity inherently encourages
commercial dissemination.
32
This reduced haul of festival prizes is particularly striking because such prestigious prizes
epitomized the cultural currency that patronage funding was intended to generate.
33
In later years, European cultural funds supported African cinema even more directly, and
Europeans played an active role in the production process. Unsurprisingly, the films became even

producing films designed to win prizes at foreign film festivals, but largely unintelligible
to Chinese audiences (Priest, 2015a).
Where patronage regimes rely on expert committees drawn from the cultural establish-
ment, biases toward elite culture may be especially pronounced. At the same time, a prefer-
ence for works that conform to establishment canons means that such regimes may reject
truly revolutionary work (Cowen, 2006).34 Delegation to committees can also introduce
further selection biases due to agency constraints; problems can include ideological biases,
private rent-seeking, or outright corruption (Pager, 2011). Such problems are likely to be
particularly grievous in developing countries with weak rule of law norms and quasi-tribal
loyalties based on ethnicity (Nigeria), caste (India), or personal connections (China).
While political oversight can curb such problems to some extent, political pressure
can also lead to excessive conservatism in making awards. The firestorm over National
Endowment for the Arts funding in the US in the late 1990s illustrates the potential
for populist demagoguery to curtail artistic innovation. Accountability to democratic
watchdogs led funders to gravitate toward sterile ‘safe bets’ that could escape censure by
the self-appointed watchdogs of public morality (Cowen, 2006).35
Even more seriously, governments can use patronage funding to advance their own
agendas, rewarding favored speakers and viewpoints. Culture is a powerful tool to influ-
ence hearts and minds and mold public opinion. As such, patronage regimes can easily
become instruments of censorship, propaganda, or authoritarian control (Pager, 2011).36
Indeed, the patron’s influence can be felt even without the need to take overt steps to
impose an agenda. Artists relying on state funding are likely to self-censor or engage in
sycophantic projects to curry favor with those holding the purse strings.
Such concerns are especially stark in developing countries where authoritarian rulers
have few scruples about wielding power through propaganda. China’s Communist party
has a long history of funding cultural agitprop. While China’s decades-long liberalization
process has led artistic production to be dictated largely by markets rather than party
apparatchiks, in recent years, the government has increased pressure on creative artists
to toe the line ideologically and encouraged a revival in works exalting the Communist
regime and its current leader (Economist, 2016; 2014). Such political preferences also
express themselves on a more general level. For example, the Chinese government
disfavors musical genres such as jazz and rock that are deemed ideologically suspect (Liu,
2015).
While China represents an extreme case, its instrumental view of the arts is hardly
unique. The idea that popular culture should serve as an agent of moral perfection
that instills wholesome virtues in audiences through positive examples is widely held in
more oriented toward European sensibilities, purveying an esthetic fetishization of the exotic
(Pager, 2011).
34
Where funds are committed in advance of production, the difficulty in predicting the merits
of a proposal can further reinforce a bias toward established artists who have a proven track record.
35
Democratic regimes face popular outcries whenever state-subsidized creativity is deemed
blasphemous, offensive, or immoral. Cutting-edge works are more likely to provoke extreme reac-
tions. Because they speak in a language unintelligible to conventional esthetic codes, their meanings
may be misunderstood. Cf. Bleistein v. Donaldson, (1903).
36
The old saw ‘He who pays the piper, calls the tune’ has a ring of truth here.

developing countries (Netanel, 1998).37 Even in countries with established commercial

content industries such as Nigeria and India, government officials continue to evince a
belief that artists have a ‘pedagogical mission’ to present a ‘positive’ vision of society
(Rajadhyaksha, 2008; Miller, 2016). Of course, all countries enforce community decency
standards. However, there is a difference between policing such outer bounds and
affirmatively encouraging ‘correct’ or ‘moral’ values through competitive allocations of
funding. Patronage regimes offer a natural vehicle for such positive censorship, and it is
all too easy for subjective standards to be manipulated to stifle viewpoints threatening to
the ruling establishment.38
In contrast to the decentralized market for expression supported under a copyright
model, a patronage regime may therefore lead to less vibrant modes of cultural expression
covering a narrower range of viewpoints and styles. Moreover, where patronage regimes
dominate funding of cultural production, alternative funding sources can be crowded out
(Pager, 2011; Cowen, 2006). Allowing state domination of cultural production thus raises
serious democratic concerns (Netanel, 1996).
Conversely, market forces expressed through the popular demand of paying audiences
can themselves function as a counter to state control. Priest (2015a) argues that in the
Chinese context, commercial imperatives have, over time, served to soften censorial
impulses. Similarly, Nollywood’s contribution to public discourse in Africa has been
significant. In contrast to the tight control over state-funded media hitherto exercised
by Nigeria and other African states, Nigeria video films have enjoyed an unprecedented
license to poke fun at cultural taboos and establishment foibles in the guise of popular
entertainment (Pager, 2012b). While both the Nigerian and Indian film industries remain
formally subject to pre-release censorship, the commercial appeal of their offerings has
helped to insulate them from aggressive enforcement of decency standards (Miller, 2016;
Bose, 2006).
2. Collective licensing, rewards and indirect subsidies

Copyright and direct state patronage are hardly the only means by which governments
subsidize cultural production. While space does not permit a comprehensive survey, it is
worth noting a few common alternatives. First, hybrid models exist that combine elements
of both copyright and patronage, typically seeking to harness the market-based incentives
that the copyright system supplies, while reducing the costs of exclusivity. For example,
compulsory licenses, collective rights management, and market-based reward regimes all
reduce access barriers by effectively shifting from a property to a liability rule whereby
copyright holders lose veto rights, but still get paid compensation (Fisher, 2004).
The downside of these hybrid models is that they compromise the market mechanisms
used to value works individually. In the absence of price signals, alternative mechanisms
must be devised that rely on indirect proxies such as statistical sampling or administrative
processes. Moreover, the efforts required to render such alternative mechanisms into
37
By contrast, the Western liberal ideal of the artist as iconoclast who challenges the status
quo remains suspect.
38
Even established democracies are not immune to the temptation to engage in such ideologi-
cal meddling, as shown by recent scandals in South Korea and Canada (Sang-hun, 2017; Pager,
2011).

lausible substitutes for market transactions can introduce further complexities whose
p
design, operation, and validation require administrative resources and careful oversight.
Hybrid regimes thus forgo some of the virtues of decentralized copyright markets and
come to more closely resemble top-down patronage models. Indeed, the very complexity of
such regimes may result in reduced transparency even compared to traditional patronage,
making them vulnerable to favoritism, fraud, and other abuses such as antitrust concerns.
Indeed, the track record of existing collective rights organizations in many developing
countries hardly inspires confidence (Schlatter, 2005; Schultz and van Gelder, 2008).
Second, governments can employ mechanisms to support cultural production indirectly.
Tax incentives offer one such approach. For example, using tax credits to induce private
investment in creative content allows public subsidies to harness private information
(Hemel and Ouellette, 2013). While such an approach can be effective, it requires careful
policing to ensure that the investments are actually aimed at the desired target activities,
and that the tax credits are not misused.39 Investments in training or infrastructure that
facilitate the production of commercial content can similarly yield powerful payoffs where
well-targeted, or result in white elephant boondoggles where misplaced (Pager, 2012c). As
with the hybrid models, to be executed effectively, these strategies thus require competent
administration, a commodity that is in short supply in much of the developing world.
Ideally, governments would manipulate these different policy levers in a targeted fash-
ion to achieve optimal outcomes. In practice, government is not nearly nimble enough to
manage this feat. The best compromise may therefore be a belt-and-suspenders approach
that blends several overlapping support mechanisms in the hopes that different actors will
self-select into the options that work best for them.
C. Commons-Based Creativity Models
The notion that subsidies are needed to avoid suboptimal investment in creative produc-
tion has come under attack in recent decades. Commentators have questioned this logic
in light of the bottomless upwelling of creative expression on the internet (Moglen,
1999). Perhaps we can just sit back and reap the digital manna from commons-based
production. There are two versions of this ‘comedy of the commons.’ The first focuses on
non-commercial creativity; the second emphasizes alternative business models premised
on ‘open’ distribution. A third, related argument emphasizes the impracticality of enforc-
ing copyright in the digital age.
1. Non-commercial creativity

In questioning the need for copyright incentives in today’s digitally empowered, post-
scarcity world, critics point to empirical research by psychologists suggesting that creators
are primarily driven by intrinsic motivations (Lemley, 2015). While such intrinsic motiva-
tions are not new, skeptics argue that digital technologies have dramatically lowered the
costs of creating and distributing new works, thereby removing the resource constraints
that previously held creators back (Benkler, 2004). As such, the extrinsic incentives
39
Developing countries often have weak tax bases and erratic collection, which reduces the
value of tax write-offs.

supplied by copyright or other subsidies have become either irrelevant or downright

counterproductive (Zimmerman, 2011).
It is clear that digital technologies have unleashed many new forms of creativity. The
proliferation of creative works by non-commercial amateur creators generates no end of
gaudy statistics (e.g. over 400 hours of video uploaded to YouTube every minute; Von
Lohmann, 2016) that testify to the cornucopia that digital networks have unleashed.
Amateur work can be genuinely creative and worthy of admiration. Moreover, the inclu-
sion of voices from underrepresented communities and viewpoints must be applauded on
democratic grounds.
At the same time, the level of authorship invested in any given work is typically low.
Amateurs generally restrict themselves to short-form content involving modest produc-
tion values, often piggybacking on preexisting works through mash-ups and remixes. The
result is an ocean of ‘lol cats’ and dancing babies, but very little that can rival commercially
produced content in scope, sophistication, or production values (Pager, 2015). Indeed,
despite the ready availability of abundant free content, consumers continue to signal
the premium they place on commercial content by patronizing it in the marketplace and
paying substantial fees for the privilege.
The limited scope of amateur creativity reflects some enduring realities of content
production that the sheer volume of production sometimes obscures. Creating anything
of more than passing interest to people outside one’s immediate family requires talent and
skills that remain in short supply. Moreover, undertaking creative projects of ambition
and scope still requires investment of time and resources that exceed the enthusiasm of
most hobbyists.40 Beyond a certain level of investment, intrinsic motivations no longer
suffice, and creators need a viable prospect of recouping their investment to proceed
(Pager, 2015).
Such exigencies are felt especially strongly in developing countries where discretionary
incomes are much lower than the United States and many struggle to meet basic social
needs. Cyberspace may be borderless and frictionless, but in the real world authors need
to eat. Money buys practical necessities such as equipment, plus the luxury of time and
artistic freedom, enabling higher quality output (Liu, 2015). Time spent struggling for
survival means less time to refine and perfect one’s craft (Merges, 2011). If creativity is
obliged to pay its way, however, the question remains how.
2. Alternative business models

The need for extrinsic revenue sources to buttress intrinsic motivations does not necessarily
mean government subsidies are required. In many circumstances, market mechanisms can
underwrite investments in creativity. In an influential early article, then-professor (now
40
Peer production can sustain more ambitious commons-based production by integrating the
collective contributions of decentralized collaborators. However, only a narrow subset of creative
projects fit the template required for such collaborative models to function effectively. You need a
modular project that can be farmed out to individual contributors and a means by such granular
output can then be efficiently reviewed, assembled, and integrated into a coherent whole (Benkler,
2002). Peer production works well for software and Wikipedia, but not for many other types of
high value authorship; assembling music, film, or novels by committee, for example, is a recipe for
disaster.

Supreme Court Justice) Stephen Breyer (1970) cited the historical example of nineteenth-
century publishing practices whereby US publishers relied on lead-time to recoup their
investment, beating copyists to the market by securing exclusive advance-sheets from
British authors whose works were denied copyright protection in the United States. Such
arrangements eventually ripened into an informal cartel system based on ‘trade courtesy’
between leading publishing houses that led to de facto exclusivity.
Lead-time is harder to preserve in the digital age, and cartels are illegal. However,
another alternative mechanism that Breyer describes—advance purchase contracts—are
not only widely used to fund indie films, but also form a mainstay of crowdfunding
campaigns. There are many more possibilities. Commons enthusiasts have catalogued
a wide array of alternative business models based on reputational economies, sales of
ancillary products and services, or advertising/sponsorship that could facilitate creative
production under an ‘open’ distribution model (Anderson, 2009; Lemley, 2015).41 They
argue that copyright’s ‘closed’ model of proprietary production should give way to more
diverse, democratic, and communal modes of creativity that digital technologies facilitate.
Removing copyright’s access barriers could encourage more creative remixing of
culture and avoid the costs of exclusivity. Critics also argue that copyright’s benefits are
monopolized by a handful of moguls and superstars, but largely irrelevant to everyone
else (Zimmerman, 2011). Creative production under a commons model is seen as poten-
tially more egalitarian. However, such assumptions are difficult to test in countries where
copyright norms remain entrenched.42
The proliferation of vibrant creative industries across the developing world that
seemingly thrive in the absence of copyright protection have accordingly been hailed as
harbingers of a post-copyright future (Anderson, 2009; Montgomery, 2010). Because
such emerging content industries effectively function in a de facto copyright commons,
they serve as a laboratory in which the viability of alternative business models can be
tested and evaluated (Rizk, 2014). Yet, as Section IV reveals, the record is far from an
unvarnished endorsement.
3. Copyright futility

Even if a copyright business model premised on controlling and monetizing individual
copies were normatively desirable, skeptics further argue that enforcing copyright exclu-
sivity has become infeasible in a world where digital technologies make unauthorized
41
Some have questioned the extent to which the revenues such alternative models generate
are sufficiently robust to replace copyright (e.g. Lowery, 2012). Others question whether such
putatively ‘post-copyright’ alternatives could really function in the absence of copyright (Priest,
2016; Liu, 2015).
42
Commons-based production in the developed world is shaped by default norms of copyright
in several ways. First, when copyright is the default from which ‘open’ production must opt out,
comparisons between the kinds of creativity that emerges under open versus closed models are
tainted by selection bias. Second, copyright remains active as a background norm that shapes
market conditions and behavior. Third, copyright may actively hinder access to creative materi-
als that commons-based producers need. Finally, open source licenses often leverage copyright
exclusivity themselves as an enforcement strategy. The ‘open businesses’ that flourish under
these conditions are therefore a poor predictor of the outcome that would ensue in a truly ‘post-
copyright’ world.

copying and distribution virtually costless (Lemley, 2015). Stepped-up enforcement,

however draconian, will not put the copying genie back in the bottle, and propping
up outdated business models just delays the inevitable. Rather than investing further
resources in such a futile quest, skeptics argue that we should embrace post-copyright
paradigms and focus on making them economically viable (Anderson, 2009).
Such claims again hold special resonance in the developing world context where the
pervasive presence of piracy and its seeming invulnerability to copyright enforcement
appears to validate such claims (Karaganis, 2011). Doubts about copyright’s viability
in the developing world, in part, reflect disillusionment over overly optimistic promises
that strong IP protection would serve as a ‘power tool’ for development in the aftermath
of the 1995 TRIPS Agreement (Pager, 2012a).43 Yet, the skepticism prompted by such
oversold claims may have veered equally far off course in the opposite direction.44 The
challenges in developing a functional copyright regime do not necessarily make the quest
futile. Moreover, evidence from developed markets shows that a combination of effective
enforcement campaigns and proactive distribution strategies can make a meaningful dent
in piracy (Danaher, et al., 2017).
The question remains whether making such a sustained effort is merited. What specifi-
cally would be gained by investing resources to this end? Would creative development be
advanced? As the following part explores, the answer is more complicated than either
copyright proponents or skeptics acknowledge.
IV. COPYRIGHT AND CREATIVE DEVELOPMENT:

REVIEWING THE RECORD
A strong body of empirical evidence makes clear that piracy hurts content industry
revenues in developed country markets (Danaher, et al., 2017). Digital technologies have
certainly facilitated piracy; yet, digital technologies have also lowered production and
distribution costs, reducing the revenue necessary to sustain content production. In his
contribution to Volume II of this series, Joel Waldfogel presents evidence suggesting
that the net effect on commercial content creation has been positive.45 While backed by
an impressive array of data, Waldfogel’s conclusions are hardly beyond question.46 Yet,
43
For example, the World Bank embarked on a much publicized initiative to develop the rich
musical heritage in West Africa into a digitally empowered music industry. Proponents hailed this
plan as a visionary attempt to build ‘Nashville’ in Africa (Finger and Schuler, 2004). The idea that
a dollop of copyright capacity-building would prove magically transformative proved misguided,
however. Building a viable content industry required more than just passing laws and setting up
websites.
44
Many developing countries have come to view copyright law as a Trojan horse designed to
advance the interests of multinational corporations at the expense of indigenous creators (Pager,
2012a). After all, IP rights were the ‘quid’ exchanged for the promised ‘quo’ of market access for
textiles and agricultural exports. This ‘bargain narrative’ thus posited IP as an inherently losing
proposition (Yu, 2006).
45
This is not to speak of the massive proliferation of amateur content creation whose limita-
tions were addressed above.
46
For example, Stan Liebowitz suggests that a ‘generational cohort’ effect may account for

even if they are accepted on face value, the most they show is that the relative decline in
copyright enforcement experienced thus far in the developed country jurisdictions from
which his data is drawn has not yet had an injurious effect. Such findings are thus in no
way predictive of the outcomes in a world in which copyright law has ceased to function
entirely.47 For a closer approximation, we can look to developing country jurisdictions in
which copyright norms are substantially weaker.
A. Thriving or Surviving?
The claim that emerging content industries are thriving in the absence of copyright
deserves scrutiny. Eric Priest (2014) uses the metaphor of ‘extremophiles’ to describe the
hardy creative industry life-forms that survive in the exigent conditions of a copyright
desert. It does not follow, however, that the absence of copyright facilitates their survival
any more than camels flourish because they are denied water. Weak copyright norms may
encourage the pioneering of new markets, but such conditions arguably hinder creative
development later on.
1. Pirate innovators

Nollywood’s origin story presents an archetypal example of piracy’s generative potential.
The extensive transnational distribution networks that pirate operators had cultivated
were instrumental in popularizing Nigeria’s nascent video film industry (Santos, 2013).
Nigerian marketers were skilled at evading censors and customs officials alike, ensur-
ing a steady supply of Nollywood videos reached audiences across Africa. India’s
content industries have similarly benefited from the pioneering of new markets by pirate
entrepreneurs. Liang and Sundaram (2011) chronicle the successful development of a
new mass market for low cost audio cassettes filled with unauthorized recordings of
popular film music.48 The pioneer in this nascent market, T-Series, also ‘expanded the
music-consuming public by focusing on genres and languages that had been ignored by
the dominant Indian record labels.’ Adrian Athique (2008) tells a similar story of new
export markets for Indian film developed by selling pirated video cassettes to South Asian
emigres through a decentralized network of Indian grocery stores.49 By doing so, such
informal distribution cultivated a new generation of Indian film devotees. Chinese search
Waldfogel’s finding that music quality has been harmed by filesharing. The idea is that people
generally favor the music they grew up with; thus, the larger the generational cohort, the more
popular the music. Millennials, the generation that has come of age in the filesharing era, have
recently passed Baby Boomers as the single largest tranche of the US population. Thus, ‘their
music’ would be expected to measure favorably in surveys (Liebowitz, 2016).
47
Indeed, as discussed below, Waldfogel co-authored a separate case study of the effects of
piracy on Bollywood, and came up with a diametrically opposite result: showing clear harm to both
the number and quality films being produced (Telang and Waldfogel, 2014).
48
Many of T-series’ recordings were technically not pirated, as they exploited a loophole
in Indian copyright law that allowed the release of new versions of existing songs (Liang and
Sundaram, 2011).
49
As with Nollywood, the informal, decentralized nature of such distribution networks served
to evade formal barriers to media exports, of particular value in Pakistan where Indian films had
long been banned (Athique, 2008).

engines and content hosting sites similarly lured online traffic with a limitless supply of
pirated media (Priest, 2015a).
Initially, the producers of the creative works in question received little benefit from
such entrepreneurial efforts. In most cases, however, the leading players in these new
distribution networks eventually legitimized their operations by entering into licensing
arrangements with producers—or became producers themselves—in part, to assure a
more regular supply of new works and to preclude competition from upstarts (Miller,
2016; Athique, 2008; Priest, 2016). Such entrenched incumbents would often be chal-
lenged in turn by a new generation of pirates exploiting newer technology to once again
disrupt existing distribution (Athique, 2008).
2. Depressed revenues

The absence of effective copyright enforcement in these markets, however, means that
piracy never really disappears. Such unauthorized distribution not only diverts revenue
from content producers, it also exerts a downward pressure on pricing both in the (legal)
market and on the licensing fees that can be extracted from intermediaries.
The effects of such diminished revenues are perhaps most visible in Nollywood. Despite
the massive global audience that its films enjoy, the industry has struggled to effectively
monetize its customer base. While Nollywood films remain profitable, the industry
operates on a breakneck schedule, perpetually starved of funds. Filmmakers are forced
to pursue a churn strategy that rushes new videos to market weekly to beat the pirates
(Barrot, 2009). Pricing of discs is kept close to marginal cost, in a further effort to deter
piracy (Miller, 2016).
Such high volume, low margin production restricts the creative ambition that can be
invested in developing any single project. Moreover, without enforceable copyrights in
their work, filmmakers cannot offer collateral to obtain financing. Instead, they must
either surrender control to marketers or tap informal short-term lenders at punitive inter-
est rates—reinforcing the ‘rush to market’ mentality that fosters slap-dash productions
(Pager, 2012b). And while the industry has expanded the reach of authorized distribution
channels both domestically and abroad, the specter of rampant piracy drives down the
prices such licensing arrangements garner (Miller, 2016). The ‘New Nollywood’ film-
makers whose higher quality productions compete for screening in domestic cinemas and
global festivals offer a glimpse of the industry’s broader creative and commercial ambi-
tions. Yet, it is far from clear that the industry’s existing revenue base can support such
lavish productions. Without a more effective means to monetize consumption, Nollywood
remains a shadow of its potential.
Filmmakers in China and India are doing better, thanks in part to robust and expand-
ing revenues from theatrical distribution. Such reliance on theatrical revenues has its
downsides, however. Bollywood remains far more dependent on theatrical exhibition:
reaping over 70 percent of its total revenue from box office sales (Singh, 2013). By
comparison, Hollywood’s theatrical take represents less than 20 percent of total revenue
(Young, et al., 2010). Relying so heavily on this single income stream makes Bollywood
vulnerable to shifts in consumption patterns.
Showing this threat is more than theoretical, Telang and Waldfogel (2014) document
the effects in the late 1980s and early 1990s of the diffusion of home video technologies
(the VCR) and independent cable television operators. These emerging channels provided

conduits for pirate distribution that allowed consumers to watch home movies for free,
including films currently in the cinema.50 Such illicit competition led Bollywood’s per
movie revenues to decline by as much as 50 percent during the period 1985–2000.51
Reduced revenues in turn led to fewer movies being produced. In notable contrast to
Waldfogel’s findings in the US context (see Volume II), Telang and Waldfogel (2014) also
analyze IMDb ratings data and conclude that the quality of Indian films produced in this
period also declined.
While piracy remains rampant in China, China’s audiovisual producers have enjoyed
comparatively healthy growth in recent years due to expanding markets for both theatrical
distribution and online streaming platforms. However, the turnaround in online revenue is
a recent phenomenon, and some question its sustainability (Priest, 2016). Moreover, the
ready availability of pirate media elsewhere continues to inhibit consumer willingness to
pay for creative content, restricting the revenue available to fund new productions.
The situation in the music industry broadly parallels that of audiovisuals. The dystopian
effects of piracy are perhaps most visible in China. Priest describes the massive drop in
revenues experienced in China from roughly 2003 onward as online distribution of pirated
music rapidly displaced legitimate sales. At a time when Chinese spending on entertain-
ment and leisure was steadily rising, revenue from recorded music dropped by more than
half (Priest, 2014). The revenue drop-off was not for lack of consumer demand. Accessing
music remained one of the most popular activities on the internet. However, Chinese con-
sumers became accustomed to not paying for something that was abundantly available for
free, and alternative revenue sources could not make up the shortfall. Liu’s (2015) survey
evidence shows that declining revenues led to reduced investment in music. Arguably as a
result, the Chinese music market became increasingly dominated by imported music from
neighboring countries (Liu, 2010). Industry revenues have rebounded in the past few years
as online intermediaries have entered into licensing deals and largely weaned themselves
off pirated content. However, China’s music industry remains comparatively undeveloped
for a country of its size and affluence (Liu, 2015, 2010).52
Digital piracy has had similarly harmful effects on the Indian music industry, decimat-
ing legitimate sales channels and causing industry-wide retrenchment (Booth, 2015).
50
Geographically sequenced theatrical releases meant that rural audiences particularly proved
receptive to such pirate distribution. Because the high cost of celluloid limited the number of
film copies that could circulate at a time, Indian theatrical exhibition followed a staggered release
schedule with new films shown first in big cities, then regional towns, only reaching village theaters
at the end of the sequence. Film copies were often worn out by this time leading to an impaired
viewing experience, and many rural residents were, in any case, impatient to wait (Liang and
Sundaram, 2011).
51
A separate study of the effects of ‘camcording’ on film revenues in seven developing
countries reached similar conclusions, showing that the availability of pirated movies online
significantly reduced box office earning (Koch, et al., 2011).
52
Music industry revenues are also paltry compared to other Chinese content industries such
as books and films, a result which both Liu (2015) and Priest (2014) attribute to music’s greater
exposure to piracy. For example, the market for digital music in China is comparable to the total
domestic box office take. However, whereas the film industry receives roughly 40 percent of box
office sales, almost none of the money generated from digital music actually goes to the music
industry (Priest, 2014).

Toward the end of the twentieth century, the music industry had begun to expand
beyond its historic role as an adjunct to the film industry and to diversify into other
genres. However, the advent of digital piracy meant that music reverted to its focus on
selling movie soundtracks (Booth, 2015). As in China, music industry revenues remain
significantly lower than film (Liang and Sundaram, 2011).53
The story in Nigeria is somewhat less dismal. Digital piracy certainly caused revenues
to plummet, leading labels to close and investment to fall (Santos, 2013). However, the
growing popularity of Nigerian music across Africa has helped to cushion the losses and
open up alternative revenue sources. In recent years, the industry has grown steadily and
attracted outside investment (Rhodes, 2014).
3. Copyright’s positive role

In each of these countries, a revival of industry fortune may be partly attributable to
copyright. This turnaround is most dramatic in the Chinese online market for streamed
content, where the leading websites have executed a remarkable turnaround from havens
of piracy to platforms populated almost entirely by licensed content. While several factors
appear to have contributed to this industry shift, all of them reflect direct or indirect
pressure to comply with copyright law (Priest, 2014). As the leading streaming platforms
purged their sites of pirated content and embraced licensing norms, they have competed
for exclusive rights to prime content. Per episode costs of leading popular dramas have
risen from $1,500 in 2009 to as much as $290,000 by 2011 (Montgomery and Priest,
2016). Soaring licensing fees have in turn spurred investment in better quality films and
television shows and provided much needed revenue for indie films that are denied access
to theatrical distribution (Montgomery and Priest, 2016).
The revival of India’s film industry after 2000 also partly reflects stepped-up
copyright enforcement. Telang and Waldfogel (2014) describe how more effective
enforcement measures after 2000 curbed the threat from pirate cable channels. Other
factors also contributed to the industry’s improving outlook, including the growth of
multiplex theaters, television licensing, and overseas revenues. While these factors are
not directly tied to copyright enforcement, it should be noted that all of these revenue
streams depend, to some degree, on copyright exclusivity to function effectively. Most
people do not watch a movie more than once, and it is hard to get customers to pay
for something that is available for free, underscoring the importance of suppressing
unauthorized distribution.
A positive role for copyright is harder to trace in Nigeria. Government initiatives
to combat piracy and formalize distribution have achieved meager results. However, a
few bright spots are apparent. COSON, a music rights management organization, has
launched a vigorous campaign to license public performances and filed a series of lawsuits
against holdouts, yielding tangible payoffs to musicians (Akpotaire, 2016). IROKOtv con-
tinues to expand its catalogue of Nollywood movies, and airline in-flight entertainment
systems have contributed additional licensing revenues (Miller, 2016). Overall, copyright
norms remain shaky, however, and benefits slim.
53
By comparison, music industry revenues in the US are comparable in size to that of film.

B. Pursuing ‘Copyright’ by Other Means
Just because creative industries can operate in developing countries where copyright
norms are weak or non-existent does not prove the viability of open business models. In
the cases examined here, it is notable the extent to which industries’ revenues depend on
exclusionary strategies. Such strategies restrict access to copyrighted works in ways that
sharply diverge from the unconstrained flows of ‘free’ content that commons enthusiasts
contemplate.
Most obviously, the theatrical film exhibition model that has been the mainstay of
Chinese and Indian film industry revenues relies on physical exclusion: Patrons must
purchase tickets to enter the cinema.54 Here, physical control over access substitutes for
copyright exclusivity. Sale of concert tickets for musical performances also relies on physi-
cal access control. However, there is a difference: Music concerts are less threatened by
open distribution of copyrighted media. Indeed, many musicians promote their concerts
by releasing free music recordings (Santos, 2013).55 By contrast, cinema revenues are much
more vulnerable to cannibalization by distribution of recorded media.56 As a result, strict
access controls across the entire theatrical distribution chain are essential to preserve
content exclusivity.
Most Nollywood films do not enjoy theatrical release, and the bulk of their earnings
come from distribution of physical media, which are even more vulnerable to piracy. In
the absence of viable copyright protection, Nollywood producers have again developed
workarounds based on a combination of lead-time and physical exclusion. The marketers’
guilds control the principal urban markets in which film copies circulate by employing
brawny enforcers to prevent pirates from setting up rival stalls; here, physical muscle
substitutes for legal writ.57 The marketers’ guild has also developed informal distribu-
tion norms that further extend exclusivity based on mutual interest (Miller, 2016). Such
informal norms recall the ‘trade courtesies’ that Breyer (1970) touted as regulating the
nineteenth-century trans-Atlantic book trade in the absence of copyright.
54
As noted above, theatrical exhibition models also depend on backdrop norms of copyright
law to function effectively—i.e. to prevent camcording and subsequent distribution or exhibition
of pirated versions of the films. However, physical exclusion backed by real property law provides
the first line of defense.
55
The music business is better suited to a commons-based business model because studio
production costs are significantly lower than film, and live concert performances offer experiential
value whose spontaneity and emotional resonance cannot be readily duplicated by recordings
(Pager, 2012a).
56
The social experience of going out to the movies makes home viewing an imperfect sub-
stitute; some customers will still pay for the experience. The vocal participation of Indian cinema
audiences, in particular, makes the theatrical experience a distinctive communal event (Hammer,
2014). For its part, Hollywood has turned to technology to replicate a similar premium on the
theatrical experience through a combination of lavish production values and special effects that
showcase best on giant screens, with 3-D images and high quality sound. Yet, the investment levels
required to achieve production qualities sufficient to make the theatrical experience significantly
more appealing than home viewing keep rising and may be out of reach of most developing
countries.
57
These practices are not unique to Nollywood. Spike Lee apparently fell back on such
‘muscular enforcement’ tactics to combat piracy of his films in New York City (Lobato, 2012).

In the case of digital music sales in China, technologically enabled access controls do
copyright’s work. Priest (2014) notes that 90 percent of recorded music revenues in 2010
derived from an extremely narrow source: cell-phone ring-back tones. Ring-back tones
are essentially music that a caller hears while awaiting an answer on the other end. Mobile
phone users pay a small monthly fee to the phone company to select personalized music.
Ring-back tones are extremely popular in China; they generate US $4 billion in annual
revenue—an amount comparable to the entire gross revenues for recorded music in the
United States (Priest, 2014). Because the tones are stored on the centralized architecture
of the phone company rather than individual users’ phones, they are insulated from piracy
(Liu, 2015).58 As such, they offer a rare context in which Chinese consumers are willing
to pay for recorded music. In theory, such technological precautionary measures (TPM)
could be used more widely to prevent unauthorized copying. However, TPM uptake in
the developing world has been slow.59
Such replication of copyright exclusivity by other means hardly seems like proof of
copyright’s irrelevance. It is not as if the content industries profiled here are blind to the
potential offered by non-exclusionary revenue models. Sponsorship deals, merchandise,
product placement, social media engagement, and the like are widely employed and
furnish a welcome source of ancillary revenues. However, it is no accident that these
industries rely on exclusionary models for the lion’s share of their revenues; the revenue
potential from alternatives sources is too limited.
Some might suggest, however, that use of these exclusionary alternatives is still prefer-
able to copyright. As examples of ‘order without law’ (Ellickson, 1991) or ‘IP without IP’
(Raustiala and Sprigman, this volume), these regimes arguably embody the virtues of pri-
vate ordering, providing bespoke solutions that serve the needs of particular communities
and contexts. Could it be that these ‘non-legal regulatory tools . . . enable more granular,
and potentially more effective, management of creative incentives’? (Perzanowski and
Darling, 2017). Such commons-based management systems potentially minimize much of
the dead-weight losses of the copyright system, in that access controls tend to be limited
and short-lived in practice. Despite their revenue limitations, such ‘synthetic copyright’
models (Liebowitz, 2017) therefore merit further exploration. Yet, closer examination
reveals that copyright’s alternatives—both exclusionary and non-exclusionary—come
with substantial drawbacks of their own.
C. The Hidden Drawbacks of Copyright Alternatives
1. Lack of scalability

Copyright law, at least in theory, functions as a seamless, global system. International
agreements such as Berne and TRIPS ensure a minimum standard of protection across
national boundaries. By contrast, the alternative revenue models described above often
rely on local arrangements that do not readily scale to broader horizons.
58
Ring-back tones are also popular in Nigeria, accounting for an estimated $150 million market.
59
Content industries appear to have prioritized other issues, perhaps out of recognition
that pirated content is already so widely available that it makes no sense to penalize purchasers
of legitimate copies by encumbering them with digital rights management or other such controls
(Scaria, 2013).

Such limitations are most apparent in the Nigerian context, where the marketing guilds’
control is highly localized: confined to specific urban street markets in a handful of cities.
Although the guilds’ internal self-regulation has helped to extend this zone of exclusivity,
such informal norms are themselves limited in scope; they function best in close-knit com-
munities where trust is based on personal relationships or kinship and monitoring against
defections is feasible.60 Indeed, Barnett argues more generally that commons-based
regulation of information goods is inherently limited in scale (Barnett, 2010).
The Chinese ring-back-tone architecture, while national in scope, is subject to its own
scale limitations. First, it only applies to music—and only specific types of music at
that. Second, it only works for music consumed in this one very narrow context. Third,
the market could eventually become saturated, if Chinese consumers tire of continually
updating their tone files with newer offerings.
Finally, while cinema exhibition is certainly scaleable in theory; there are practical
limitations in this regard. China still lacks sufficient screens to meet domestic demand for
movies (although it is building more), and state distributors further limit the number of
films authorized for theatrical distribution (Priest, 2014). Movie theaters in developing
countries are typically concentrated in urban areas, whereas potential audiences can
be widely dispersed across the countryside (Pager, 2012a); diasporic audiences can be
similarly scattered. High crime rates make evening cinema-going unsafe in some areas.
Indeed, most cinemas closed in Nigeria in the 1980s for this reason. Moreover, in many
Muslim regions, including Northern Nigeria, women are denied access to public cinemas
due to prohibitions against gender mixing (Pager, 2012a).
These limitations apply to music concerts as well. In China, in particular, there is a
shortage of suitable venues. The Communist party’s distrust of public gatherings outside
their control further limits the number of concerts and festivals that can be staged, as well
as the types of musical acts that are allowed to perform (Montgomery, 2010; Liu, 2015).
Moreover, whereas multiple copies of a movie can be screened simultaneously, concert
performers can do at most one show per night (and even that quickly becomes too much;
Schultz, 2009).
By contrast, copies of recorded media can reach dispersed audiences wherever they are
located. Such decentralized distribution affords the convenience of on-demand consump-
tion in the home and can easily be scaled to meet global demand. Yet, monetizing such
consumption is difficult without copyright.
2. Abuse of power

Another drawback of alternatives to copyright is their vulnerability to power asymmetries
and abuses. Copyright law, in principle, is administered in an evenhanded fashion subject
to principles of fairness, due process, and non-discrimination (see, e.g. TRIPS, Arts. 3,
4, 41–2). Under the TRIPS Agreement, compliance with such rule of law safeguards is
enforceable through binding dispute resolution at the World Trade Organization backed
60
Jennifer Davis makes a similar point about the trade courtesy regime governing nineteenth-
century trans-Atlantic book publishing whose membership was bounded by class and ethnicity
(2014). Similar kinship ties also characterize (and delimit) Bollywood’s film families (Liang and
Sundaram, 2011).

by trade sanctions. This is not to say that deviations never occur in practice. In many coun-
tries, for example, enforcement practices tend to favor the products of local industries over
imports (Karaganis, 2011). In South India, such favoritism is amplified by the close ties
between local film industries and provincial political parties.61
However, informal, norm-based, alternative regimes are often marred by lack of
transparency and discriminatory treatment. This is certainly the case with the Nigerian
marketers’ guilds (Miller, 2016). When informal norms are combined with private deploy-
ment of violence, the potential for abuse becomes all the more grievous: Nigerian ‘big
men’ in the guild have a tendency to flout rules with impunity (Paulson, 2012).62
A different sort of power asymmetries governs the Chinese ring-back-tone market.
This market is controlled by mobile phone companies, and in China two firms dominate.
Because the Chinese music industry lacks other viable sources of income (due to piracy),
it is obliged to license music to this duopoly on starkly unequal terms: The Chinese mobile
operators keep over 97 percent of the proceeds, with less than 3 percent split between the
record labels and musicians (Priest, 2014).
Stanley Liebowitz (2017) documents how the trans-Atlantic trade courtesy regime
operated by nineteenth-century publishers was similarly marred by exploitative behavior.
American authors (whose works were subject to copyright) received much less generous
terms compared to British authors (whose works were not). Moreover, the publisher’s
regime treated the latter as permanently bound to whichever publisher first published his
or her work, accentuating the potential for monopsonistic abuse.
Different, but equally objectionable, proprietary attitudes toward artists are displayed
by corporate sponsors today. In China and Nigeria, such funders often seek to exert
creative control or impose other conditions that inhibit artistic autonomy (Liu, 2015;
Pager, 2012b). In this regard, private patronage proves no less objectionable than its public
analogue (Netanel, 1996).
Of course, power asymmetries between authors and intermediaries exist in the
copyright system as well. Different jurisdictions provide doctrinal safeguards against the
resulting inequities to varying degrees (Pager, 2015). Moreover, copyright’s market struc-
ture helps to enable successful authors to work under conditions of creative autonomy
(Merges, 2011).
61
The starkly provincial bent of enforcement practices in South India reflects the intimate
connection between film and politics: Politicians fund movies that burnish the leading stars’ heroic
image, and the stars, in turn, campaign for the politicians (and often become politicians later in
their own right). As a result, vendors of pirated discs who dare to sell pirated versions of local films
face punishing police crackdowns. Such highly selective enforcement remains locally bounded:
Tamil police protect the Tamil film industry; Kannada police protect Kannada films, etc. (Liang
and Sundaram, 2011). Accordingly, despite their formal use of police power and ostensible basis in
copyright law, such practices are not that different from Nigeria’s marketers’ guild, and suffer from
some of the same limitations: vulnerability to abuse, distorting effects in production signals, and
potentially inequitable distribution of benefits.
62
While not directly related to copyright, it should also be noted that Bollywood’s flirtation
with mafia financing in the 1980s led to similarly extralegal uses of force to exert control over film
production; the sometimes fatal consequences for industry participants represent a particularly
sinister form of dead-weight losses (Athique, 2008).

3. Distorting effects of alternative revenue sources

A further critique of alternative business models and other copyright substitutes is that
they tend to monetize only a limited segment of total consumption as their revenue
source. Because producers only internalize benefits from one part of the market, they
receive distorted signals as to societal demand. This in turn can channel future production
in suboptimal directions. For example, Chinese musicians and record labels who rely on
ring-back tones for the bulk of their income have an incentive to create new music that
is suitable for use in the ring-tone market and to ignore competing sources of audience
demand that they are unable to monetize. This means composing short, catchy melodies
suitable for the low quality acoustics of the ring-back context at the expense of more
sophisticated works with higher production values. Jiarui Liu (2015) describes how
Chinese music companies are doing exactly that. The Indian music industry has similarly
reverted to its traditional subsidiary role as a supplier of Bollywood film music, forsaking
the linguistic and genre diversity that had emerged when music could be sold directly to
consumers as a standalone good (Booth, 2015).
This distorted incentive critique can be applied more broadly to other commons-based
alternative business markets.63 For example, Nagla Rizk (2014) describes the ability of
Egyptian singers to earn a living performing at weddings and other social gatherings. One
suspects that such singers will develop a repertoire of sentimental ballads rather than, say,
protest songs or devotional hymns in order to cater to the narrow niche of Egyptian music
demand that happens to be monetizable.64 Indeed, Rizk suggests that Arab authorities
deliberately favor pop stars over more socially conscious ‘underground music’ as a ‘means
to distract the masses’ (2014).
Katherine Strandburg (2013) predicts similar distortions will occur in US content
markets funded through online behavioral advertising, invoking an analogous theory
of price signal mismatches. Advertising supported creative markets are already subject
to a host of undesirable biases, from the undue influence of sponsors (Baker, 1997)
to the prevalence of click-bait (Gardiner, 2015) and the distorting effects of two-sided
markets (Reidel, 2011).65 More generally, any reliance on ancillary revenues to subsidize
the production of creative content is vulnerable to distortions where the proxy market
represents an imperfect substitute for the social value of the primary good.66 For
example, the ability to sell happy meals figurines only imperfectly captures the value
of a film qua film. When merchandising potential and product placements provides
63
As noted, Chinese and African filmmakers under state patronage regimes were similarly
incentivized to produce films that catered to European film festivals at the expense of domestic
audiences. One can speculate as to the extent to which such distortions occur in other markets. For
example, do South Indian filmmakers overinvest in heroic star-vehicles in order that their political
funders can reap the electoral benefits?
64
Capacity constraints on performance models impose further selection biases, favoring mass
market over niche content. Such biases are particularly grievous in China as elaborated further in
Section V, infra.
65
This is not to mention the privacy concerns that invasive online advertising implicates
(Kapczynski, 2012).
66
One can also raise questions of fairness: Why should consumers in one market subsidize
products consumed in a different market. And what happens if demand for the proxy market
disappears?

the primary value proposition underlying movies, the quality of the film itself becomes
secondary.67
Such distorted signals can be contrasted with the allocative efficiency that a prop-
erly functioning copyright system ensures. Copyright’s broad bundle of rights allows
producers to internalize benefits across the full range of demand for a given creative work,
thereby capturing market signals that orient future production accordingly (Goldstein,
2003).68 The result is a mix of creative works that serves a broad range of societal tastes
and interests.69
D. Distribution of Benefits
Furthermore, the claim that copyright only benefits a narrow few must be evaluated
against the alternatives. Much of the skewed distribution of benefits arguably flows from
superstar economics and other inherent biases of media markets that have very little to
do with copyright (Shur-Ofry, 2012). The same skews can result under alternative business
models based on sponsorship, concert performances, and sale of merchandise (Liu, 2015).
Alternative business models also favor particular types of creators with particular traits
that may have little to do with the quality of their work. For example, sponsorship models
place a premium on attractive, charismatic pitchmen (Liu, 2015).70 Similarly, a concert
performance model works best for established bands whose fan base is concentrated in
major urban areas. Performers with a fan base dispersed across a wide geographic area
may struggle to fill venues. Touring on the road also imposes hardships that not everyone
can readily endure (e.g. single parents). Some artists’ talents and certain musical genres
simply lend themselves to studio projects better than live concerts (Schultz, 2009). Such
groups could be systematically disadvantaged were concert revenues the only source of
remuneration. At minimum, such biases call into question claims that commons regimes
yield a fairer distribution of benefits than copyright.
Copyright models may exacerbate distributional skews to the extent that they
generate higher levels of revenue overall, making skews seem more extreme. However,
it is important to keep in mind that a portion of the revenues that superstars and Big
Media conglomerates generate trickles down to subsidiary personnel (session musicians,
67
The Pixar movie Up provides a real-life example of such dystopian incentives: although the
film was hailed as a critical success and made a profit at the box office, Disney stock still fell due to
the movie’s lack of merchandisable characters.
68
We can debate how far along the chain of derivative revenue streams we want to allow
authors to internalize benefits. This point is explored further in the discussion of cultural diversity,
infra Section V. The primary point, however, is that copyright does a better job of capturing a
broad spectrum of market signals than many alternatives.
69
Of course, intellectual property markets are not without their own distortions. As Amy
Kapczynski points out, only consumers who have the ability to pay get counted (2012). Moreover,
pricing of media goods in some developing markets can be disproportionately oriented toward the
wealthy (Karaganis, 2011)—although query whether this trend would persist if piracy were better
controlled, allowing for more effective price discrimination strategies.
70
Physical attractiveness and charisma provide advantages in the copyright system as well.
However, the effect of this benefit is likely to be particularly emphasized under a sponsorship
model.

composers, camera-men, editors, and the like). In this way, copyright helps to subsidize
up-and-coming artists, many of whom hone their skills through such industry day jobs
while they await their turn in the limelight (Hazucha, 2014).71
E. Context-Specific Copyright
What about the claim that eliminating piracy is futile—akin to King Canute’s struggle
to hold back the tide? Should creators in the digital age just learn to live with it? There’s
undeniably some truth to this claim. Digital technologies reduce the marginal costs of
reproduction and distribution to almost zero, whereas production costs remain high,
allowing pirates an opportunity for profitable arbitrage.
However, copyright enforcement does not have to be an all-or-nothing proposition.
Reducing or delaying piracy by even a small amount can make huge differences to an
industry’s bottom line, as the Indian and Chinese cases demonstrate. Furthermore, even
if piracy remains rampant at the retail distribution level, this does not obviate the benefits
of copyright entirely. Although our usual image of ‘piracy’ conjures up shadowy figures
in trench coats peddling misbegotten wares, content industries face a range of potential
appropriators which include competitors, intermediaries, broadcasters, cable networks,
airlines, hotels, and other established enterprises. Enforcing copyrights against such
entities may be more feasible than preventing retail piracy as high profile enterprises
can hardly hide in the shadows. As noted, India has made tremendous headway toward
enforcing copyright norms in this regard: bringing pirate cable operators into the fold;
ensuring that public performances are licensed; and clearing rights for film inputs. Nigeria
has made recent progress in this regard as well; the main holdup is a lack of reliable proof
of ownership. China too has moved toward a licensure model on many fronts, although
state broadcasters and other government entities remain largely untouchable. Such
incremental gains belie claims of copyright’s inevitable doom.
Enforceable copyrights can also promote greater trust and collaboration internally
within content industries. Small creators who furnish inputs for content production
such as songs or scripts are particularly vulnerable to predatory corporations in the
absence of copyright (Hughes, 1999). Lack of copyright protection also introduces
perverse incentives in the way that creative industries operate. For example, fear of
script piracy has even led some Nollywood directors to withhold scripts from their
actors; instead, actors are only given their lines for individual scenes as they are shot
(Pager, 2012a).72 However, recent developments, including a favorable Nigerian high
court ruling on script piracy as well as the Nigerian Copyright Commission’s online
platform for electronic registration, offer Nigerian creators the prospect of greater
protection against such internal misappropriation (Opara, 2017; Nigerian Copyright
Commission, 2017).
71
Copyright arguably exacerbates distributional inequalities in another respect: by helping
well-capitalized media firms to fund marketing campaigns that accentuate the advantage of
blockbuster culture over indie alternatives. See infra Section V-A.
72
Admittedly, trade secret law can counter this threat to some degree. But copyright’s reach is
broader, allowing remedies against third parties who exploit the purloined property, even without
showing culpability or knowledge of the misappropriation.

As these accounts demonstrate, copyright norms operate according to a variable

geometry whose contours remain highly contextualized. Even if it is true that digital
piracy can never be eradicated, achieving more modest enforcement and licensing goals
can still yield palpable benefits to creators’ bottom lines in the form of increased revenues
and reduced uncertainty.
F. The Formalization Imperative
Furthermore, the value of adhering to copyright norms grows as industries develop over
time and move—or are nudged—toward more formal modes of operations. Expanded
commercial ambitions bring with them the pressure to conform to established structures
of the global copyright. For example, a filmmaker that seeks a bank loan for an upcoming
production will usually be required to offer the copyright as collateral. Completing such
transaction will require the filmmaker to demonstrate a host of formal copyright instru-
ments that collectively document ownership of the work to establish the chain of title.
Similarly, if the filmmaker seeks an international distributor or a film festival exhibitor,
they will likely have to document clearance of all the film’s copyrightable inputs—music,
screenplay, background art, and so on (Miller, 2016). It may not take too many iterations
of this exercise before such formalities become routinized as a business practice.
Potential sponsors of the work can also apply pressure toward formalization. Advertisers
of international brands do not want to be associated with illegal practice, as Nollywood
filmmakers discovered when their penchant for using unauthorized music tripped them
up (Pager, 2012a). Similarly, pressure from foreign advertisers is said by Priest to have
played an instrumental role in persuading Chinese web video sites to embrace licensure
norms (Priest, 2014).
Even without such overt pressure to embrace copyright, other structural forces exert
their own pull, tugging emerging content industries toward global copyright norms.
For example, participating in the reciprocal arrangements that govern international col-
lecting societies brings with it the prospect of royalty payments from wealthy diasporal
populations, but requires investment in formal record-keeping and local monitoring/
enforcement. Similarly, engaging in joint productions with Western content industries can
expose local partners to global copyright ‘best practices’ and bring with it the concomitant
expertise required for compliance.73 Such structural forces can push creative content
industries toward embracing copyright formalization even in the absence of credible
enforcement threats. Indeed, exposure to global copyright norms can potentially exert a
ripple effect that shifts industry behavior merely by making such norms a focal point of
attention (McAdams, 2000).
73
Such collaborations are especially common in the music industry where world music is a
growing market niche and Western stars such as Paul Simon and Ry Cooder have made a habit
of collaborating with artists across the developing world. Joint ventures in film are also becoming
more common. India has produced a steady diet of crossover films combining Indian talent with
Western funding, direction, and distribution: e.g. Monsoon Wedding, Slumdog Millionaire, Best
Exotic Marigold Hotel, and so on. A forthcoming Chinese film by Zhang Yimou is to be made in
English starring Matt Damon (McClendon, 2016). New Nollywood has also begun to move in this
direction, drawing backing and distribution from Nigerian expatriates (Miller, 2016).

A shift toward embracing copyright norms over time can also arise for competitive
reasons. As noted, there is a tendency for content distributors exploiting new technologies
or business models to build their platform and attract users on the back of authorized
content. However, once they have achieved critical mass, the now-established incumbents
then seek to entrench their position through exclusive content licensing deals that give
them a leg up over competitors. China’s internet video industry, and, more recently, its
music streaming services have followed this path, and indeed are now busy suing each
other for copyright infringement (Priest, 2016).74 A similar pattern emerged in the Indian
music industry (Athique, 2008). Piracy certainly has not gone away, but at least some
powerful incumbents have taken a stake in upholding the copyright system.75
Even consumers can sometimes play an informal role in policing copyright norms,
filling a void left by ineffectual state institutions. For example, Chinese and Indian fan
groups monitor distribution of works by their favored artists and can mobilize to exert
sanctions against unauthorized distributors (Hammer, 2014). Over time, such collective
mobilizations could conceivably pave the way for a transition toward an emerging norm
of copyright compliance.
G. Taking Stock
This chapter has offered a nuanced account of the interaction between copyright law and
development, and the extent to which commons-based models offer a viable alternative.
The case studies examined show that copyright law is not a sine qua non for creative
development. Yet, they also suggest that the ability of creative industries to grow beyond
a certain level in its absence is limited and that there are structural imperatives that push
toward copyright formalization. Copyright protection benefits creative industries by
allowing them to diversify their revenue streams and pursue a variety of possible business
models. Importantly, copyright law does not prevent rightsholders from releasing works
under open licenses or pursuing commons-based business models, if they choose. It
merely gives them the option not to (Barnett, 2013).76
Copyright is not cost-free. Consumers have to put up with reduced access and higher
costs, thus some amount of consumer surplus is forfeited (Karaganis, 2011).77 Yet, if
74
Google arguably has played a similar game with YouTube and GoogleBooks (pre-settlement
collapse); although its licenses are non-exclusive, it arguably enjoys market power via network
effects that allow it to achieve more favorable terms than an upstart competitor.
75
This pattern can be cyclical as it progresses through iterative evolutions toward new technol-
ogy and new platforms. Thus, the shift toward formalization is not necessarily continuous (Lobato,
2012). However, as industries develop, the long-term trend seems be toward embrace of copyright
norms. The biggest hold-out remains Nigeria, but even here the emergence of new licensing
platforms backed by Western investment, such as Iroko in audiovisuals and Feeme in music, offers
hope that formalization may yet arrive as an industry norm.
76
Some might argue that widespread instantiation of copyright business models crowd out
space for innovative ‘open’ alternatives. By preventing such alternatives from licensing a critical
mass of content, the incumbent content industries keep them from getting off the ground, or deter
investors by threatening infringement suits.
77
Unauthorized distribution can sometimes also add value to the consumption experience,
e.g. where fans of a foreign television show distribute it for a domestic audience with high quality

they get higher quality productions to enjoy as a result, arguably the ledger looks more
even.78 What about claims of blocked innovation and censorship? While commentators in
the developed world obsess over user generated content by non-commercial artists, such
issues hardly register in the developing world context. Content industries in the Global
South are waging an existential struggle against commercial enterprises whose unauthor-
ized distribution of verbatim copies directly supplants legitimate sales in their primary
market.79 Such industrial scale piracy has nothing to do with internet mash-ups, parodies,
and the like. Enforcement actions against such secondary creators is the last priority of
rightsholders, and appropriately tailored exceptions can easily be made to accommodate
them, if need be.80
Arguably, the biggest difference between copyright models and alternatives may be the
extent to which content industries develop into highly concentrated, integrated structures
and acquire more capital-intensive capabilities. This begs the questions posed at the start
of this chapter: What kind of cultural development do we want? Do we have to choose
between blockbusters and viral web videos? Arguably, what matters most is having a
diversity of offerings. Is copyright compatible with cultural diversity? These issues are
addressed in the next part.
V. COPYRIGHT AND CULTURAL DIVERSITY
Thus far, this chapter has discussed the development of creative content industries largely
in terms of the economic revenue they generate. Cultural development, however, is argu-
ably about more than cranking out cultural widgets, however profitable. As Big Media
conglomerates have grown more concentrated and powerful, concern has mounted over
cultural diversity. Copyright and media theorists argue that access to a diverse supply of
cultural media is essential from a variety of normative standpoints, including democratic
discourse (Netanel, 1996; Baker, 2002), human flourishing (Fisher, 1988; Sunder, 2012),
personal autonomy (Benkler, 2001), and multiculturalism (Shur-Ofry, 2012; Burri, 2012).
However, the manner in which diversity is defined varies greatly by context. In the United
subtitles added through crowd-sourced translations, often within days or even hours of the original
broadcast.
78
Concerns over access to knowledge (e.g. scientific publishing and educational materials)
present a more serious worry (Chon, 2007). The Berne Annex has provisions to help developing
countries overcome such hurdles, but there are still real costs to be borne here. Such costs, however,
are largely extrinsic to the ‘cultural development’ focus of this chapter. Concerns over access to
cultural resources are more relevant as media literacy arguably provides a foundation for creative
innovation. Appropriate tailoring of exceptions and limitations, and public funding for libraries
offer partial solutions. Further exploration, however, lies outside the present scope.
79
Peer-to-peer filesharing can be non-commercial, but the leading platforms are almost always
supported by for-profit entities that earn money from advertising.
80
A rare case of a parody remix in China that provoked claims of copyright infringement
offers an instructive example: When Hu Ge’s video remix parodying a blockbuster film by
acclaimed director Chen Kaige went viral online, attracting Chen’s ire, China’s netizens mobilized
in support of the remix, forcing Chen to back off from his threatened litigation (Gong and Yang,
2010). More generally, Chinese courts have shown themselves willing, in principle, to construe
Chinese copyright law flexibly to accommodate fair use interests (Song, 2011).

States, for example, diversity can be discussed in terms of media ownership, racial/ethnic
representation, or political viewpoint (Napoli, 2012). Internationally, diversity is often
measured in terms of language or national origin (Burri, 2012). Some discuss diversity in
terms of ‘shelf space’ in cultural markets; others focus on production process, contrasting
artisanal with ‘industrialized’ creativity; still others are preoccupied with tracking actual
consumption patterns (Napoli, 2012; Arsenault and Castells, 2008).
By lowering the threshold costs of creating and distributing content, digital technologies
have democratized cultural expression. As such, one might expect that they have proven
to be an unmitigated boon for cultural diversity. The explosion of Nigerian filmmaking
in the early 1990s offers perhaps the best example of this. Not only did Nollywood grow
from almost nothing to become a major global industry within a decade, but each of
Nigeria’s major ethnic groups has its own regional film industry that produces movies
in a different language (Miller, 2016). Nollywood movies are diverse in other ways as
well: by genre, subject matter, viewpoint, and also the backgrounds of those who make
them: multi-ethnic casts and crews are common, and a significant number of Nollywood
directors are female (McCall, 2002).
However, one could also argue that by hewing rigidly to genre conventions, Nigerian
movies are themselves commodity products, mass-produced according to the logic
of global capitalism (Lobato, 2012). Critics contrast such formulaic, commercial
output with the artistic masterpieces of Africa’s celluloid film tradition (Pager, 2012b).
Nollywood movies also freely borrow from Hollywood tropes, techniques, and sometimes
specific plots, while Hausa films, made in Northern Nigeria, have drawn heavily from the
Bollywood film tradition, further undermining their authenticity in the eyes of detractors
(Larkin, 2004). Moreover, the widespread circulation of Nollywood videos and music
across Africa and beyond has sometimes come at the expense of local media, prompting
charges of cultural imperialism (Krings and Okome, 2013). These contrasting viewpoints
point to the deep ambiguities lurking within the concept of cultural diversity.
For present purposes, we need not commit to any single definition of diversity and can
consider the concept from multiple standpoints. The salient question to explore, however,
is what role do copyright and other cultural funding models play in shaping diversity?
Subsection A examines evidence from and debates surrounding developed markets.
Subsection B recontextualizes these debates in the developing world context. Subsection
C then briefly considers copyright’s transnational effects on diversity in the global market.
A. Diversity Effects in Developed Markets
Copyright provides an incentive to invest in cultural production. More copyright means

more investment, albeit subject to diminishing returns. Note that this does not mean that
strengthening copyright will lead to more creative works being produced and thus more
diversity. It may have the opposite effect: Producers may choose to invest bigger sums in a
smaller number of works or focus on a narrow range of genres (Nadel, 2004). Therefore,
the effect of copyright on diversity a priori is ambiguous.
As noted, copyright law is largely agnostic as to output. It leaves it to the market to
determine the appropriate allocation of investment and discourages subjective judgments
about cultural value within copyright doctrine. The connection between copyright and
cultural diversity has been explored by commentators, however, along several dimensions,

including: (1) the threat from commoditized culture; (2) structural biases; (3) anticompeti-
tive abuses; (4) effects of digitization; and (5) restrictions on downstream creativity. These
dimensions are explored below in turn.81
1. Commodity culture

Ever since the Frankfurter School, commentators have reviled commercial ‘culture indus-
tries’ for producing a bland ‘commodity culture’ the same way that McDonalds makes
hamburgers—mass-produced to appeal to the lowest common denominator of taste. Such
cultural junk food allegedly manipulates its viewers into over-consumption at the expense
of more authentic, wholesome fare (Skladany, 2008). Modern-day heirs to this tradition
view copyright as a tool for entrenching the dominance of transnational corporate media
(e.g. MacMillan, 2014; Story, 2003). Their critiques are echoed by cyberlibertarian schol-
ars who view copyright law as ‘resistance’ to be routed around by the digital networked
currents of authentic folk culture (Moglen, 1999; Benkler, 2006).
Yet, yearnings for some imagined folkloric past or cyber-utopian future do not alter the
reality that the production of cultural goods remains, in many cases, a resource-intensive
enterprise requiring extrinsic funding. Such funding can come either from the govern-
ment or from the market. The tradeoffs between copyright and state patronage explored
in Section III pointed to characteristic biases that each regime displays in incentivizing
particular forms of cultural production: State patronage tends to be oriented toward
commercially marginal niches, for example, funding avant-garde works or subsidizing
minority expression. By contrast, copyright directs investment toward commercially
viable forms of expression.
Setting aside Marxist distaste for commodification on principle, what exactly is wrong
with commercial culture? After all, to succeed, such works require the validation of paying
audiences. What is wrong with giving consumers what they want? (Baker, 1997).
The cultural junk food metaphor referenced above offers some indication of possible
responses. Critical media scholars have identified a long list of societal ills that com-
mercial mass culture allegedly exacerbates, including harmful stereotypes, valorization
of celebrity, hyper-sexuality, consumerist materialism, stultifying passivity, and the need
for instant gratification. To the extent that copyright law subsidizes such addictive and
unhealthy consumption, it arguably produces a public bad rather than good (Skladany,
2008).
More generally, critics such as Edwin Baker (2002) and Amy Kapczynski (2012)
note that cultural media are associated with a wide range of social externalities not
captured by market pricing. Included among them are normative values associated
with culture diversity. Baker also emphasizes that audiences’ preferences are not
endogamous: What people want is partly shaped by the menu of options with which
81
Given the obvious connection between diversity and product differentiation, the emerging
literature on copyright and product differentiation (Bracha and Syed, 2014b; Abramowicz, 2011;
Yoo, 2004) deserves mention here. Such scholarship recognizes that the sheer number of works
produced matters less in terms of satisfying societal demand than the spacing of such works across
the demand spectrum and seeks to apply such insights to optimize copyright doctrine. Further
theoretical work is arguably required, however, to develop this literature before its normative
payoffs can be usefully connected to the broader discourse on copyright and diversity.

they are presented.82 Accordingly, copyright markets respond to consumption patterns

that may be a poor proxy for social welfare.
2. Market structure

A related diversity critique emphasizes not so much the commercial nature of copyright
markets as the particular structure such markets assume. In theory, copyright markets
respond to decentralized demand signals from the consuming public to satisfy a wide
range of tastes, both mainstream and niche. Yet, markets for mass media characterized
by high production costs and low distribution costs have a tendency to tip toward winner-
take-all outcomes because bestselling works deliver outsized returns. The winning strategy
is therefore to invest in lavishly produced, heavily promoted works of blockbuster culture
that maximize market returns by crowding out competition from more diverse works
(Nadel, 2004).83 The resultant ‘tyranny of the market’ (Waldfogel, 2007) means that the
profit maximizing solution is not necessarily the best outcome in terms of satisfying
consumer preferences.
Many commentators link such market failures to copyright law. Mark Nadel (2004)
argues that copyright increases the winner-take-all biases of media markets by generating
inflated profits that Big Content uses to fund its marketing juggernaut. Guy Pessach
(2003) attributes such harms to the expansive derivative works right that copyright
provides, leading to overinvestment in blockbuster works that drive merchandise sales
and other ancillary revenues. Neil Netanel (2000) describes how copyright facilitates price
discrimination and bundling strategies that can be best implemented on a large scale,
further tipping copyright markets toward concentration.
However, Pager (2011) notes that revenues from blockbuster works often cross-subsidize
more experimental forms of creativity. Moreover, Michal Shur-Ofry (2012) argues that
copyright is, at most, a marginal player in this story. She sees the scale economies of
superstar economics as driving outcomes and suggests there is not much we can do about
network-induced solidarity preferences that are an integral part of being human. Shur-
Ofry (2014) further argues that media markets are far too complex a system to fine-tune
through copyright doctrine.
It should be also noted that relying on alternative revenue models such as sponsorship
or merchandising can push just as strongly toward winner-take-all outcomes (Liu, 2015).84
More generally, Guy Pessach (2016) describes informational capitalism under the com-
mons model as dominated by winner-take-all platforms online that perpetuate forms of
cultural commoditization that are just as troubling as the criticisms laid at copyright’s door.
82
Baker describes a feedback loop between the media goods that consumers encounter in the
marketplace and the molding of consumer preferences that determine future consumption choices
(1997, 2002).
83
Large media conglomerates can also exploit economies of scale and scope through growth
and industrial integration, not to mention market power (addressed in subsection A(3) infra).
Moreover, as noted in subsection A(3) infra, large content firms can more easily negotiate copy-
right permissions to use cultural media as inputs.
84
As noted below in subsection A(4), Elberse (2013) views media blockbusterization, in part,
as a strategy that responds to weakening of copyright exclusivity, relying instead on branding
power to generate ancillary revenues.

3. Anticompetitive abuse

Commentators have also expressed concerns over the market power that large content
conglomerates wield. Detractors allege that Big Content abuses its clout, disadvantaging
smaller players through a variety of anticompetitive practices, including product tying
(e.g. block booking; Burri, 2012) and preferential dealing (Netanel, 2000; Nadel, 2004).
Copyright plays a role in this story as well. In so far as copyright encourages media
concentration, it exacerbates the potential for harm. Big Content leverages its exclusive
rights over large catalogues of valuable content to negotiate more favorable terms in the
marketplace than smaller competitors can obtain (Garcia, 2014). Such firms can also
more easily obtain permission to reuse commercial content owned by others through
implicit cross-licensing deals (Pager, 2015). Yet, it is far from clear the extent to which
copyright itself is inherently culpable in this narrative, as opposed to, for example, regula-
tory failures in antitrust and communications law.
Moreover, some commentators have also argued that too little copyright can have
equally pernicious effects on cultural diversity. Guy Pessach (2013, 2016) suggests that
copyright plays an important structural role in mediating between authors and interme-
diaries. In conditions of weak copyright enforcement, he worries about dominant content
hosting platforms, such as YouTube, abusing their market power at authors’ expense in
ways that harm diversity. Justin Hughes (1999) likewise argues that copyright promotes
diversity in part by protecting small creators from abuses of corporate power. Netanel
(1996) and Merges’ (2011) points about the autonomy enhancing function of copyright
are apropos here as well.85
4. Effects of digitization

A further facet of this debate concerns the effects of digitization on creative content
markets. Chris Anderson (2006) captured a lot of attention with his ‘longtail economics’
theory holding that the advent of digital media and, in particular, the vastly expanded
scope for distribution in cyberspace would lead to a decisive shift in both supply and
consumption away from the mass market.
The supply of online content has unquestionably grown by orders of magnitude
(Lemley, 2015). While scholarly attention has focused on amateur creators, a new breed of
commercial creators have braved the new world of digital disintermediation (Pager, 2015)
to contribute to output diversity as well. Less clear is the extent to which consumption
patterns have shifted. Some suggest that digital markets can advance diversity by supply-
ing more personalized media offerings through automated recommendation engines and
innovative data mining techniques (Smith and Telang, 2016).
Yet, Anita Elberse’s book, Blockbusters (2013), shows that the Big Content producers
have moved the other way: focusing ever-more myopically on promoting and hyping a
narrow range of mass market offerings. In a way, these narratives are complementary:
The massive proliferation of free amateur and indie content has made it necessary for the
big players to launch their marquee content with a supersized splash that cuts through
the noise and grabs public attention. The real loser may be content in the mid-market.
85
See Section III-A(4), above. Netanel (2000) also argues that a degree of concentrated power
is necessary to give the media the clout to stand up to the government and other power centers.

The big studios have downsized their art-house divisions, and a similar trend can be seen
in the music industry.
Where is copyright in this story? It is arguably felt by its absence: By facilitating both
commercial piracy and peer-to-peer filesharing, digital media have clearly hurt creative
industry revenues (Danaher, et al., 2017). This may exert a negative effect on diversity.
Elberse (2013) draws a connection between piracy and blockbusterization, suggesting
that ‘tentpole’ strategies, in part, respond to loss of control over content distribution.
Big Content can still make money from sponsorship, merchandise, tie-ins, and all the
other brand-driven benefits that flow from a blockbuster franchise. By contrast, smaller
competitors are less able to tap such ancillary revenues. Large distributors can also better
cope with the challenges of digital enforcement compared to upstart creators (Pager,
2015). Some, however, argue piracy-induced revenue hits have weakened Big Content’s
hegemony, creating openings for upstart rivals (Burri, 2012; Waldfogel, vol. II this
Handbook).86
5. Restrictions on downstream creativity

Another place where copyright is seen as having a negative impact on diversity concerns
its effect on downstream creativity and, in particular, on derivative art forms such as
remixes, mash-ups, and fan fiction that are popular with amateurs. The digital age
makes it easier for rightsholders to target forms of amateur creativity that previously
flew beneath the radar (Litman, 2006). Moreover, the proliferation of user generated
content online means there is a lot more to target. Fans of such grass roots creativity hail
its liberating potential. Proponents of remix culture cite its potential to affirm minority
identities, challenge orthodoxies, and air alternative perspectives (Sunder, 2012; Lessig,
2008). Much scholarly ink has been spilled exploring the threat to such follow-on creativ-
ity posed by expansive interpretations of copyright and the need to carve out a safe zone
to protect it (e.g. Menell, 2016).
Responding to these views, some cast doubt on the extent to which copyright actually
exerts such chilling effects in practice (Joo, 2011; LaPolt, et al., 2015). Others contest the
diversity value that such derivative expression brings. For example, Joo (2011) argues that
remixes just reinforce the hegemony of the dominant media. And Fishman (2015) argues
that by constraining copying and thereby forcing authors to ‘create around’ copyrighted
roadblocks, copyright law actually promotes diversity, forcing us to be more original.
Such disputes highlight the inherent subjectivity of cultural diversity as a normative goal.
B. Recontextualizing to the Development Context: Evidence from Case Studies
The preceding diversity debate has largely played out among scholars in the developed
world, based on conditions which may not necessarily extrapolate to developing countries.
To the extent such discourse focuses on the US context, it may be particularly misleading
given the United States’ unique position as a global hegemon of popular culture.
86
Economists have also attempted to measure the effects of piracy on diversity more directly.
For example, Hammond (2014) suggests that piracy hurts sales of indie artists more than main-
stream artists.

Accordingly, insights from the developing world context may offer a useful corrective
to the existing debate. It should be noted, however, that the case studies in this chapter—
Nigeria, India, and China—focus on large countries whose creative markets may not be
representative of the developing world as a whole. Nigeria and India, in particular, are
global exporters of popular culture and occupy a role as regional hegemons comparable
to the United States, albeit on a smaller scale.
That said, it is worth noting that the content industries in India and Nollywood are far
more decentralized than their US counterparts. Their domestic markets are fragmented
along regional/linguistic/ethnic lines.87 Moreover, there are many more producers in
the market, and they release far more movies than Hollywood. One might be tempted
to attribute such decentralization to the weaker copyright norms that prevail in these
countries. On closer inspection, however, the picture becomes more ambiguous.
At first glance, China perhaps best illustrates Nadel’s (2004) argument about copy-
right reducing diversity through winner-take-all markets. The copyright enforcement
climate has generally been improving in recent years. At the same time, there has been
a trend toward consolidation in China’s creative industries and an increasing focus on
blockbuster/superstar models (Frater, 2015; Liu, 2015). Yet, it is premature to say that
copyright either caused consolidation or impaired diversity. Indeed, in the music realm,
the emphasis on superstars is arguably better explained as a response to the lack of
copyright enforcement that has forced a reliance on alternative business models (Liu,
2015; Priest, 2014).88
Moreover, China’s content industries are undergoing rapid transition in the wake of
market liberalization (Montgomery and Priest, 2016). Many of China’s big companies
are looking for opportunities to diversify into the fast-growing entertainment market
and have consciously emulated Hollywood’s blockbuster model (Sun, 2016; Frater, 2015).
And China’s government has increasingly looked to its culture industries as a source of
global ‘soft power’ (Montgomery, 2010). All of these factors contribute to the ongoing
industry restructuring, irrespective of copyright, underscoring Shur-Ofry’s (2012) point
that creative markets are responsive to many extraneous factors that can swamp the
diversity effects of copyright.
As for copyright’s supposed anticompetitive effects, it is true that Chinese intermediar-
ies are busy purchasing exclusive rights to popular content, leveraging copyright as the
means to build market share and shut out rivals. Recent mergers between such firms do
raise concerns over anticompetitive injuries (Priest, 2014; Economist, 2017). Independent
artists have been unfairly exploited due to their lack of negotiating clout (Economist,
2017). Yet, such anticompetitive abuses arguably pale compared to the extortionate licens-
ing terms imposed by China’s mobile phone duopoly for ring-back tones. As noted, the
exploitation there results from copyright failures rather than its excesses.
87
By contrast, China’s population is 90 percent Han, and use of Mandarin language almost
ubiquitous. China does have ethnic minorities, with whom it has politically fraught relationships.
Yet, such minority media as exists tends to be state subsidized (and controlled), rather than market-
driven (Na Yi, 2016).
88
Capacity constraints on film exhibition and music concerts further constrain content
diversity. With limited slots to allocate, distributors and promoters naturally favor content with
mass market appeal over more experimental fare.

At the other extreme, Nollywood seemingly illustrates the inverse implications of

Nadel’s argument. In the absence of enforceable copyrights, filmmakers make minimal
investments in any one single project but rather have adopted a ‘spawn’ instead of ‘nurture’
strategy that releases a steady flood of creative progeny into the market in the hope of
evading pirate predation through sheer numbers. Yet, while the resultant output is diverse
in many respects, the diversity of content is notably shallow: Nollywood movies recycle
familiar elements within well-defined genres based on established formulas. Without the
resources to invest in product differentiation strategies, a standardized product offers the
best way to secure customer patronage (Lobato, 2012). In other words, the absence of
copyright has encouraged commoditized culture.
Bollywood experienced a similar phenomenon in the early 1980s. As new technologies
enabled pirates to divert revenues, Bollywood’s creative ambitions stagnated. Its staple
output of masala (‘mix’) films had become largely fungible reworkings of the same stale
formula: the romantic musical melodrama (Pager, 2011).89 It is notable, however, that
copyright failures did not lead to an increase in production in the way that Nollywood’s
example might predict. Rather, Telang and Waldfogel (2014) found that increased piracy
led to lower production by Bollywood.
Conversely, when copyright enforcement improved after 2000, Indian film production
increased, and larger budgets allowed filmmakers to tackle more ambitious creative
projects that enhanced the overall diversity of movies on offer. Bollywood continues
to turn out masala films targeting consumption by rural masses, while simultaneously
producing more sophisticated offerings catering to urban elites and expatriates. Moreover,
India’s regional film industries catering to linguistic minorities have largely maintained
their market share despite the higher budget Hindi films on offer (Pager, 2011). India’s
example suggests that the causal relationships between copyright, industry concentration,
and cultural diversity are more complex than critics have allowed. One might, for example,
postulate a sweet spot at which the optimal level of copyright maximizes diversity.90 An
alternative explanation would focus on path dependencies that prevent industries from
easily restructuring and switching strategies as they move from one equilibrium level to
another.
Indeed, some of the decentralization in Nigeria and India today reflects the informal
basis on which their film industries operated in decades past, denied the benefits of gov-
ernment recognition and access to financial markets. Bollywood has made considerable
progress toward ‘corporatization’ and professionalization since obtaining official industry
status in 2001 (Pager, 2011). However, the industry remains remarkably decentralized.
89
Floyd Whaley’s (2012) account of the Filipino film industry also traces a causal trajectory
between increased piracy and less distinctive films. Booth (2015) similarly attributes retrenchment
in the Indian music industry and a resulting reduction in diversity to the advent of digital piracy.
90
One explanation in the film context may be that both big-budget blockbuster films made
under a strong copyright regime (because they involve bet-the-company investments) and low-
budget, high volume straight-to-video production models favored in a weak copyright climate (due
to lack of a marketing budget) permit little scope for risk-taking innovation. In both cases, the
incentive is to play it safe and work within existing genres/formulas to deliver a product that appeals
predictably to proven sources of demand. By contrast, a moderately effective copyright system
could allow space for diversity and experimentation.

Moreover, significant investment continues to be driven by ulterior motives includ-

ing money laundering and political influence (Rajadhyaksha, 2014), which acts as a
centrifugal force impeding industry concentration. Meanwhile Nollywood’s marketers
remain stubbornly resistant to change (Bud, 2014, 2016); the hegemonic grip exercised by
its guilds provides a further cautionary tale as to the anticompetitive outcomes that can
result in copyright’s absence.
Finally, one clear difference between emerging and developed markets is that copyright
lawsuits against follow-on creators are much more rare in the former. As noted, enforce-
ment priorities lie elsewhere. India represents a partial exception in that music versioning
and remixes have commercial significance there and have attracted some enforcement
efforts (Greene, 2013). However, India’s version of a cover tune license has been
interpreted broadly, and amendments to the Copyright Act in 2012 provided additional
protection. As such, concerns over copyright’s chilling effects remain more theoretical
than actual in the development context.
C. Global Diversity
1. Overview of diversity concerns

Cultural diversity concerns transcend the boundaries of individual national states, and
the advent of globalization and global communications has only increased them. In
general, a bigger global market should support a greater diversity of products. Yet, the
scale economies of media goods tend to tip toward domination by big-market producers
(Wildman and Siwek, 1988). For the better part of a century, the United States has had
the largest home market, and most countries have been on the receiving end of American
popular culture hegemony.
UNESCO’s 2005 Convention on the Protection and Promotion of the Diversity of
Cultural Expressions attracted almost the entire world as signatories (with the notable,
albeit predictable exception of the United States). Yet, while countries pay lip service
to the Convention’s ideal of cultural diversity as a global public good, they are mostly
concerned about preserving their own domestic content industries. Many employ overtly
protectionist policies, propping up domestic content industries through subsidies and
restricting foreign competition (Pager, 2011). China, for example, strictly limits imports
of foreign movies (Priest, 2014).91
In fact, the threat of US domination is arguably less dire for developing countries than
it is for developed ones due to greater differences in cultural and social contexts. Such dif-
ferences create a translational barrier that makes it harder for American cultural exports
to dominate in Lagos than in London (Schultz, 2012). This ‘cultural discount’ applied
to imported works creates a corresponding advantage for locally produced content that
is steeped in the everyday realities of domestic audiences. Moreover, now that digital
technologies have leveled the playing field, local content industries can finally exploit this
built-in advantage and satisfy such hitherto unmet demand (Pager, 2012c).
91
India used to do the same until the 1990s (Pager, 2011). Nigeria, for its part, heavily sub-
sidized domestic television production in the 1980s, which indirectly benefited Nollywood (Pager,
2012c).

Market size still matters to a degree. Regional superpowers Nollywood and Bollywood
exert their own cultural hegemony over their neighbors. Hong Kong and South Korea
played a similar role in East Asia (Pager, 2011), a mantle that China looks poised to
inherit. Such ‘cultural colonization’ can also extend to diasporic populations: Bollywood
increasingly targets affluent South Asian migrant communities who can be tapped for
higher-priced ticket sales; in recent years, it has produced a bumper crop of films centered
on the expatriate experience (the so-called ‘Manhattan-in-Mumbai’ genre; Pager, 2011).
Nollywood too has begun to follow this trend (Santos, 2013). But who is colonizing
whom? Even the US is now arguably a prisoner of its own hegemony. Because Hollywood
increasingly relies on the global box office, it makes films whose appeal travels as widely
as possible. The result has been a shift away from dialogue-intensive romantic comedies,
which American audiences appreciate but which are hard to translate, in favor of violent
action films built around lavish special effects and easily recognizable cultural icons such
as comic book characters (Pager, 2011).
2. The role of copyright

Once again, a key question is where does copyright figure in the global diversity equation?
Delegates negotiating the UNESCO Convention could not agree on an answer. As a
result, copyright was omitted from the final text, save for a brief mention in the preamble
(Hazucha, 2014).
Some see global copyright law as a threat. By creating a seamless web of globally
reciprocal protection, the Berne Convention and TRIPS Agreement have effectively
expanded the market for global content. The globalization of copyright law thus increases
the stakes of winner-takes-all contests. On this view, enforcing copyrights only serves to
enrich Western multinational conglomerates and entrench their hegemony (Story, 2003).
Yet, allowing piracy of foreign media arguably makes matters worse for local artists.
Piracy effectively acts as a zero-price subsidy to purchase foreign media thereby undercut-
ting the demand for locally produced content (Pager, 2012a).92 Fickle audiences spoiled
for choice by a massive repertoire of pirated content drawn from the world’s best films and
television no longer need to bother with sampling local unknowns. Moreover, local tastes
may become warped by exposure to esthetic norms forged in alien contexts (Montgomery
and Priest, 2016).
Furthermore, the worldwide protection afforded by global copyright law serves as
much opportunity as threat. The relationship between Nigerian and Indian content
industries and their diasporic communities in the West illustrates this potential. More
effective copyright norms could encourage expatriates to invest in content industries back
home and also ensure that such industries internalize the benefits from overseas markets
currently served by pirate distribution. Growing interest among Western audiences in
sampling ‘exotic’ cultures represents a further market for content industries to develop
(Pager, 2012a).
92
Complaints along these lines by US authors were instrumental in persuading Congress to
extend copyright to foreign works (Pager, 2012a). For the same reason, countries such as France
that are seriously concerned about cultural diversity are often strong proponents of copyright
enforcement (Hazucha, 2014).

To realize such a multipolar future of global content flows will require investments in
capacity-building. As noted, the copyright system functions through a complex set of
institutions and practices from collecting societies to clearance protocols that often require
specialized expertise to navigate (Pager, 2015). Emerging content industries often struggle
to realize the benefits copyright affords due to lack of information and capabilities (Pager,
2012a). Reducing such obstacles would arguably yield a more diverse copyright system.
3. Cross-cultural remakes, parallel imports and versioning

Copyright affects cultural diversity in other ways. Diversity concerns are also implicated
by forms of copying that fall short of verbatim piracy. For example, Bollywood has a pen-
chant for remakes of successful American films—everything from borrowing general plot
ideas to more extensive copying that goes scene by scene, replicating dialogue verbatim
and even camera angles (Desai, 2005). Such reworkings are common in other emerging
film industries as well.93 Producers like such remakes because borrowing market-tested
stories reduces risk. However, local writers and audiences arguably lose out on opportuni-
ties to have wholly indigenous stories appear on the silver screen.
That such remakes are rarely licensed raises the issue of infringement (Desai, 2005).
Where do we set the line between permissible borrowing of ideas and potentially unlawful
copying of expression? How much license should we give for transformative ‘cultural
translations’ of copyrighted originals? Some of this just recaps the debate, supra, regard-
ing constraints on downstream creativity. However, the trans-cultural dimension of the
copying implicates broader questions regarding diversity.
Taking global cultural forms and indigenizing them through local settings and idioms
can be seen as a mixed blessing. It assimilates dominant foreign media and messages,
which could be viewed as a threat to diversity. Yet, such hybrid forms also serve to recon-
textualize aspects of global modernity and reconcile them with local meanings and tradi-
tions in ways that could be viewed as positive (Pager, 2012b). After all, culture does not
exist in a pure form, frozen in time and immutable to outside influence; rather—especially
in a global, digital age—it must react to and engage with changes around it to survive.
Trans-cultural remakes arguably contribute to this adaptive process. Such trans-cultural
adaptations therefore do more than merely dissipate rents from the original. That said,
the extent to which copyright doctrine should encourage this practice remains unclear.
Such ‘glocalization’ practices—adapting global commodity culture to local contexts—
are connected to copyright doctrine in another way. The Kirtsaeng decision (Kirtsaeng v.
John Wiley & Sons, Inc., 568 U.S. 519 (2013)) construed US copyright law as governed by
an international exhaustion rule.94 Such a ruling makes it harder for US content producers
to price discriminate by offering lower priced copies for sale in developing markets.
93
Nigerian filmmakers have reworked foreign story lines from a variety of sources, although
the degree of copying is typically less extensive (Santos, 2013). Meanwhile, filmmakers in other
African countries have freely adapted Nollywood film plots to local contexts (Krings and Okome,
2013). Chinese copying, by contrast, seems to focus more on cloning American TV shows
(Generalpaperpress, 2012), although charges of copycat films are not unknown (Lu, et. al, 2015).
94
Rules on international exhaustion (the global version of first sale doctrine) were left unhar-
monized under TRIPS Article 6. Many developed countries apply a rule of national or regional
exhaustion, which segments the global market and precludes parallel imports. By contrast, coun-

Technological controls offer a partial solution. However, these can be circumvented. A

more effective solution is to create distinct versions of a work, tailored to the language and
customs of each market, thereby restoring the potential for differential pricing without
concerns over parallel importation. Kirtsaeng will thus encourage greater uptake of
versioning strategies.
Whether such an outcome is normatively desirable remains a difficult question. That
reasonable minds can disagree on the answer only underscores the ambiguities that sur-
round discourse on cultural diversity and transnational culture flows.
VI. CONCLUSION
This chapter has explored the relationship between copyright and cultural development.
It shows copyright is neither a magical elixir of growth, nor an outdated paradigm of
dubious value. Rather, copyright law provides a pragmatic tool to encourage investments
in particular kinds of cultural production oriented toward market demand. Having exam-
ined case studies based on the Nigerian, Indian, and Chinese music and film industries,
several conclusions emerge. First, we can see that copyright offers important advantages
over alternative models such as state patronage or commons-based production. In par-
ticular, the choice between ‘closed’ copyright and an ‘open’ commons may offer a false
dichotomy. The case studies reveal that creative industries employ alternative exclusionary
strategies in the absence of copyright. Such pursuit of ‘copyright by other means’ has
negative ramifications that arguably outweigh the downsides of formal copyright. At the
very least, such drawbacks must be reckoned with in any comparative assessment.
Furthermore, the benefits of copyright law can be realized even without committing
to a full-scale war on piracy. Moreover, the logic and external benefits of copyright for-
malization exert their own gradual pull over time. Finally, while the relationship between
copyright and cultural diversity is complex and ambiguous, there is reason to think that
copyright markets are not intrinsically hostile to diversity and that other factors play a
larger role in driving outcomes. Further research is needed both to clarify the underlying
dynamics at play and to unpack the values encoded under the rubric of diversity.
REFERENCES
Abramowicz, Michael. 2011. ‘A New Uneasy Case for Copyright,’ 79 George Washington Law Review 1644–91
Abulude, Samuel. 2015. ‘Nigeria: Jay Z to Invest in Nigerian Music Industry,’ AllAfrica.com (April 29), accessed
4 April 2019 at http://allafrica.com/stories/201504290086.html.
Abulude, Samuel. 2016. ‘Nigeria: Tackling Funding Challenge in Entertainment Industry,’ AllAfrica.com (April
10), accessed 4 April 2019 at http://allafrica.com/stories/201604110103.html.
Akpotaire, Ufuoma. 2016. ‘Ten Intellectual Property Cases that Made Headlines in 2016,’ Nigerian Law IP
Watch, Inc., accessed 4 April 2019 at https://nlipw.com/10-intellectual-property-cases-that-made-headlines-
in-2016/.
Anderson, Chris. 2006. The Long Tail: Why the Future of Business is Selling Less of More. New York: Hyperion.
tries applying an international exhaustion rule treat a lawful sale of a copyrighted work anywhere
in the world as enabling that particular copy to be freely resold (Gautam, 2014).

Anderson, Chris. 2009. Free: The Future of a Radical Price. New York, NY: Hyperion Books.
Arewa, Olufunmilayo B. 2015. ‘Nollywood and African Cinema: Cultural Diversity and the Global
Entertainment Industry,’ in Irene Calboli and Srividhya Ragavan, eds., Diversity in Intellectual Property. New
York, NY: Cambridge University Press.
Arsenault, Amelia H., and Manuel Castells. 2008. ‘The Structure and Dynamics of Multi-Media Business
Networks,’ 2 International Journal of Communication 707–48.
Athique, Adrian. 2008. ‘The Global Dynamics of Indian Media Piracy: Export Markets, Playback Media and
the Informal Economy,’ 30 Media, Culture & Society 699–717.
Baker, C. Edwin. 1997. ‘Giving the Audience What It Wants,’ 58 Ohio State Law Journal 311–417.
Baker, C. Edwin. 2002. Media, Markets, and Democracy. Cambridge: Cambridge University Press.
Baker, Dean, 2003. ‘The Artistic Freedom Voucher: An Internet Age Alternative to Copyrights,’ Center for
Economic and Policy Research (November 5), accessed 4 April 2019 at http://cepr.net/documents/publications/
ip_2003_11.pdf.
Banerjee, Arpan. 2016. ‘Copyright Piracy and the Indian Film Industry: A “Realist” Assessment,’ 34 Cardozo
Arts & Entertainment Law Journal 609–98.
Barnett, Jonathan M. 2010. ‘The Illusion of the Commons,’ 25 Berkeley Technology Law Journal 1751–816.
Barnett, Jonathan M. 2013. ‘Copyright Without Creators,’ 9 Review of Law & Economics 389–438.
Barrot, Pierre. 2009. Nollywood: The Video Phenomenon in Nigeria. Bloomington, IN: Indiana University
Press.
Benkler, Yochai. 2001. ‘Siren Songs and Amish Children: Autonomy, Information, and Law,’ 76 New York
Benkler, Yochai. 2004. ‘Sharing Nicely: On Shareable Goods and the Emergence of Sharing as a Modality of
Economic Production,’ 114 Yale Law Journal 273–358.
Booth, Gregory. 2015. ‘Copyright Law and the Changing Economic Value of Popular Music in India,’ 59
Ethnomusicology 262–87.
Bose, Derek. 2006. Brand Bollywood: A New Global Entertainment Order. New Delhi, Thousand Oaks, London:
Sage Publications.
Boyle, James. 1997. Shamans, Software, and Spleens. Cambridge, MA: Harvard University Press.
Bracha, Oren, and Talha Syed. 2014a. ‘Beyond Efficiency: Consequence-Sensitive Theories of Copyright,’ 29
Bracha, Oren, and Talha Syed. 2014b. ‘Beyond the Incentive-Access Paradigm? Product Differentiation &
Bud, Alexander. 2014. ‘The End of Nollywood’s Guilded Age? Marketers, the State and the Struggle for
Distribution,’ 6 Critical African Studies 91–121.
Bud, Alexander. 2016. Email correspondence with author, June 20.
Burri, Mira. 2012. ‘Cultural Protectionism 2.0: Updating Cultural Policy Tools for the Digital Age,’ in Sean A.
Pager and Adam Candeub, eds., Transnational Culture in a Digital Age. Cheltenham, UK and Northampton,
Chon, Margaret. 2006. ‘Intellectual Property and the Development Divide,’ 27 Cardozo Law Review 2821–912.
Chon, Margaret. 2007. ‘Intellectual Property from Below: Copyright and Capability for Education,’ 40 UC
Davis Law Review 803.
Cohen, Julie E. 2011. ‘Copyright as Property in the Post-Industrial Economy: A Research Agenda,’ 2011
Contreras, Jorge L. 2013. ‘Confronting the Crisis in Scientific Publishing: Latency, Licensing and Access,’ 53
Santa Clara Law Review 491–575.
Coombe, Rosemary. 1998. The Cultural Life of Intellectual Properties: Authorship, Appropriation, and the Law.
Cowen, Tyler. 2006. The Good and the Plenty: The Creative Successes of American Arts Funding. Princeton, NJ:
Princeton University Press.
Dale, Martin. 1997. The Movie Game: The Film Business in Britain, Europe and America. London: Cassell &
Company.
Danaher, Brett, Michael D. Smith, and Rahul Telang. 2017. ‘Copyright Enforcement in the Digital Age:
Empirical Evidence and Policy Implications,’ 60 Communications of the ACM 68–75.
Davis, Jennifer. 2014. Review of Robert Spoo, ‘Without Copyrights: Piracy, Publishing, and the Public Domain,’
119 American Historical Review 917–18.
Deloitte. 2016. ‘Indywood: The Indian Film Industry’ (September), accessed April 16, 2019 at https://www2.

deloitte.com/content/dam/Deloitte/in/Documents/technology-media-telecommunications/in-tmt-indywood-
film-festival-noexp.pdf.
Demsetz, Harold. 1969. ‘Information and Efficiency: Another Viewpoint,’ 12 Journal of Law and Economics 1–22.
Desai, Rachana. 2005. ‘Copyright Infringement in the Indian Film Industry,’ 7 Vanderbilt Journal of
Entertainment & Technology Law 259–78.
Economist. 2010. ‘Lights, Camera, Africa,’ The Economist (December 16), accessed 4 April 2019 at http://www.
economist.com/node/17723124.
Economist. 2014. ‘The Art Is Red: Propaganda Art is Enjoying a New Lease of Life,’ The Economist (December
20), 65–6.
Economist. 2016. ‘The Song Dynasty: Melodious Love-Offerings for Xi Jinping,’ The Economist (March 12),
32.
Economist. 2017. ‘A Pirate’s Life No More,’ The Economist (July 29), 56.
Elberse, Anita. 2013. Blockbusters: Hit-Making, Risk-Taking, and the Big Business of Entertainment. New York,
NY: Henry Holt & Co.
Ellickson, Robert C. 1991. Order Without Law: How Neighbors Settle Disputes. Cambridge, MA: Harvard
University Press.
Finger, J. Michael, and Philip Schuler. 2004. Poor Peoples’ Knowledge: Promoting Intellectual Property in
Developing Countries (Trade and Development). Washington, D.C.: World Bank Publications.
Fisher III, William W. 2004. Promises to Keep: Technology, Law, and the Future of Entertainment. Stanford, CA:
Fishman, Joseph P. 2015. ‘Creating Around Copyright,’ 128 Harvard Law Review 1333–404.
Frater, Patrick. 2015. ‘China Rising: How Four Giants Are Revolutionizing the Film Industry,’ Variety (February
3), accessed 4 April 2019 at http://variety.com/2015/film/news/china-rising-quartet-of-middle-kingdom-conglo
merates-revolutionizing-chinese-film-industry-1201421685/.
Fu, W. Wayne. 2012. ‘Cross-Country Comparison of Audience Tastes in Hollywood Movies: Cultural Distance
and Genre Preferences,’ in Sean A. Pager and Adam Candeub, eds., Transnational Culture in the Internet Age.
Gana, Ruth L. 1995. ‘Has Creativity Died in the Third World? Some Implications of the Internationalization
of Intellectual Property,’ 24 Denver Journal of International Law and Policy 109–44.
Ganti, Tejaswini. 2012. Producing Bollywood. Durham, NC: Duke University Press.
Garcia, Kristelia A. 2014. ‘Penalty Default Licenses: A Case for Uncertainty,’ 89 New York University Law
Review 1117–83.
Gardiner, Bryan. 2015. ‘You’ll Be Outraged at How Easy It Was to Get You to Click on This Headline,’ Wired
(December 18) accessed 4 April 2019 at https://www.wired.com/2015/12/psychology-of-clickbait/.
Gautam, Zubin S. 2014. ‘The Murky Waters of First Sale: Price Discrimination and Downstream Control in the
Wake of Kirtsaeng v. John Wiley & Sons, Inc.,’ 29 Berkeley Technology Law Journal 717–58.
Generalpaperpress. 2012. ‘Chinese Clones of US Shows a Big Turn-Off,’ Generalpaperpress (May 26), accessed
4 April 2019 at https://generalpaperpress.wordpress.com/2012/05/26/chinese-clones-of-us-shows-a-big-turn-
off/.
Goldstein, Paul. 2003. Copyright’s Highway: From Gutenberg to the Celestial Jukebox. Stanford, CA: Stanford
Law & Politics (Revised Edition).
Gong, Haomin, and Xin Yang. 2010. ‘Digitized Parody: The Politics of egao in Contemporary China,’ 24 China
Information 3–26.
Greene, Paul D. 2013. ‘Bollywood in the Era of Film Song Avatars: DJing, Remixing, and Change in the Film
Music Industry of North India,’ in Gregory D. Booth, ed., More Than Bollywood: Studies in Indian Popular
Music. New York, NY: Oxford University Press.
Guha, Jishnu. 2005. ‘Time for India’s Intellectual Property Regime to Grow Up,’ 13 Cardozo Journal of
International and Comparative Law 225–60.
Hammer, Brandon. 2014. ‘Smooth Sailing: Why the Indian Film Industry Remains Extremely Successful in the
Face of Massive Piracy,’ 5 Harvard Journal of Sports & Entertainment Law 147–87.
Hammond, Robert. 2014. ‘Profit Leak? Pre-Release File Sharing and the Music Industry,’ 81 Southern Economic
Journal 387–408.
Hansmann, Henry, and Marina Santilli. 1997. ‘Authors’ and Artists’ Moral Rights: A Comparative Legal and
Economic Analysis,’ 26 The Journal of Legal Studies 95–143.
Haynes, Jonathan. 2013. ‘The Nigerian Diaspora: A Nigerian Video Genre,’ in Matthias Krings and
Onookome Okome, eds., Global Nollywood: The Transnational Dimensions of an African Video Film Industry.
Bloomington, IN: Indiana University Press.
Hazucha, Branislav. 2014. ‘Cultural Diversity and Intellectual Property Rights: Friends or Foes?,’ in Lilian
Richieri Hanania, ed., Cultural Diversity in International Law: The Effectiveness of the UNESCO Convention
on the Protection and Promotion of the Diversity of Cultural Expressions. London: Routledge.

Hemel, Daniel, and Lisa Larrimore Ouellette. 2013. ‘Beyond the Patents-Prizes Debate,’ 92 Texas Law Review
303–82.
Hughes, Justin. 1999. ‘“Recoding” Intellectual Property and Overlooked Audience Interests,’ 77 Texas Law
Review 923–1010.
Ingham, Tim. 2016. ‘Sony Music Open Nigeria Office as It Expands across Africa’ (September 23), accessed 4
April 2019 at http://www.musicbusinessworldwide.com/sony-music-opens-nigeria-office-as-it-expands-across-
africa/.
Jedlowski, Alessandro. 2010. ‘Beyond the Video Boom: New Tendencies in the Nigerian Video Industry,’ ORBI,
accessed 4 April 2019 at https://orbi.uliege.be/handle/2268/177492.
Jedlowski, Alessandro. 2012. ‘Small Screen Cinema: Informality and Remediation in Nollywood,’ 13 Television
& New Media 431–46.
Joo, Thomas W. 2011. ‘Remix Without Romance,’ 44 Connecticut Law Review 415–79.
Kaminski, Margot E. 2014. ‘The Capture of International Intellectual Property Law through the U.S. Trade
Regime,’ 87 Southern California Law Review 977–1052.
Kapczynski, Amy. 2008. ‘The Access to Knowledge Mobilization and the New Politics of Intellectual Property,’
117 Yale Law Journal 804–85.
Karaganis, Joe, ed. 2011. Media Piracy in Emerging Economies. New York, NY: Social Science Research Council.
Koch, Jason, Mike D. Smith, and Rahul Telang. 2011. ‘Camcording and Film Piracy in Asia-Pacific Economic
Cooperation Economies,’ International Intellectual Property Institute, Washington, Carnegie Mellon
University, Study, accessed 14 April 2019, at https://iipi.org/wp-content/. . ./Camcording-and-Film-Piracy-in-
APEC-Economies.pdf.
Krings, Matthias, and Onookome Okome. 2013. ‘Nollywood and Its Diaspora: An Introduction,’ in Matthias
Krings and Onookome Okome, eds., Global Nollywood. Bloomington: Indiana University Press.
LaPolt, Dina, Jay Rosenthal, and John Meller. 2015. ‘A Response to Professor Menell: A Remix Compulsory
License is Not Justified,’ 38 Columbia Journal of Law & the Arts 365–74.
Larkin, Brian. 2004. ‘Itineraries of Indian Cinema: African Videos, Bollywood, and Global Media,’ in Ella
Shohat and Robert Stam, eds., Multiculturalisms, Postcoloniality, and Transnational Media. New Brunswick,
NJ: Rutgers University Press.
Lee, Edward. 2012. ‘Copyright, Death, and Taxes,’ 47 Wake Forest Law Review 1–43.
Lemley, Mark. 2015. ‘IP in a World Without Scarcity,’ 90 NYU Law Review 460–515.
Lessig, Lawrence. 2008. Remix: Making Art and Commerce Thrive in the Hybrid Economy. New York, NY: The
Penguin Press.
Liang, Lawrence. 2009. ‘Piracy, Creativity and Infrastructure: Rethinking Access to Culture,’ Alternative Law
Forum, accessed 4 April 2019 at https://papers.ssrn.com/sol3/papers.cfm?abstract_id=1436229.
Liang, Lawrence, and Ravi Sundaram. 2011. ‘Chapter 8: India,’ in Joe Karaganis, ed., Media Piracy in Emerging
Economies. New York: Social Science Research Council.
Lichtman, Douglas. 2003. ‘Copyright as a Rule of Evidence,’ 52 Duke Law Journal 683–743.
Liebowitz, Stan. 2016. Email correspondence with author, September 18.
Liebowitz, Stan. 2016. ‘Paradise Lost or Fantasy Island? Voluntary Payments by American Publishers to
Authors Not Protected by Copyright,’ 59 Journal of Law & Economics 549–567.
Litman, Jessica. 2006. Digital Copyright. Amherst, NY: Prometheus Books.
Liu, Jiarui. 2010. ‘The Tough Reality of Copyright Piracy: A Case Study of the Music Industry in China,’ 27
Cardozo Arts & Entertainment Law Journal 621–61.
Liu, Jiarui. 2015. ‘Copyright for Blockheads: An Empirical Study of Market Incentive and Intrinsic Motivation,’
38 Columbia Journal of Law & Arts 467–548.
Lobato, Ramon. 2010. ‘Creative Industries and Informal Economies: Lessons from Nollywood,’ 13 International
Journal of Cultural Studies 337–54.
Lobato, Ramon. 2012. Shadow Economies of Cinema: Mapping Informal Film Distribution. UK: British Film
Institute.
Lowery, David. 2012. ‘Meet the New Boss: Even Worse than the Old Boss,’ The Trichordist (April 15), accessed
4 April 2019 at https://thetrichordist.com/2012/04/15/meet-the-new-boss-worse-than-the-old-boss-full-post/.
Lu, Shenand Anna Hsieh, 2015. ‘Copycat Movie? Filmmaker Denies Knocking off Pixar’s “Cars,”’ CNN (July
7), accessed 4 April 2019 at http://www.cnn.com/2015/07/07/china/china-movie-disney-cars/.
Macaulay, Thomas Babington. [1841] 1915. ‘The First Speech on Copyright, February 5, 1841,’ in James
Fleming Hosic, ed., Macaulay’s Speeches on Copyright and Lincoln’s Address at Cooper Institute. New York,
NY: Henry Holt and Company.
Macmillan, Fiona. 2014. ‘Cultural Diversity, Copyright, and International Trade,’ in V.A. Ginsburgh and David
Throsby, eds., Handbook of the Economics of Art and Culture. New York, NY: Elsevier.
McAdams, Richard H. 2000. ‘A Focal Point Theory of Expressive Law,’ 86 Virginia Law Review 1649–729.

McCall, John C. 2002. ‘Madness, Money, and Movies: Watching a Nigerian Popular Video with the Guidance
of a Native Doctor,’ 49 Africa Today 79–94.
McCall, John C. 2004. ‘Nollywood Confidential: The Unlikely Rise of Nigerian Video Film,’ 13 Transition
98–109.
McClendon, Lamarco. 2016. ‘“The Great Wall” Director Addresses “Racist” Matt Damon Casting Controversy,’
Variety (August 4), accessed 4 April 2019 at https://variety.com/2016/film/news/the-great-wall-director-res
ponds-matt-damon-casting-controversy-1201830581/.
Menell, Peter S. 2016. ‘Adapting Copyright for the Mashup Generation,’ 164 University of Pennsylvania Law
Review 441.
Merges, Robert. 2011. Justifying Intellectual Property. Cambridge, MA: Harvard University Press.
Miller, Jade L. 2016. Nollywood Central. London: BFI Palgrave.
Moglen, Eben. 1999. ‘Anarchism Triumphant,’ First Monday (August 2), accessed 4 April 2019 at https://
firstmonday.org/ojs/index.php/fm/article/view/684/594.
Montgomery, Lucy. 2010. China’s Creative Industries: Copyright, Social Network Markets and the Business
of Culture in a Digital Age. Cheltenham, UK and Northampton, MA, USA: Edward Elgar Publishing
Limited.
Montgomery, Lucy, and Brian Fitzgerald. 2006. ‘Copyright and the Creative Industries in China,’ 9 International
Journal of Cultural Studies 407–18.
Montgomery, Lucy, and Eric Priest. 2016. ‘Copyright and China’s Digital Cultural Industries,’ in Michael
Keane, ed., Handbook of Cultural and Creative Industries in China. Cheltenham, UK and Northampton, MA,
Morarjee, Rachel. 2015. ‘China’s Music Dream Is Out of Tune with Reality,’ Reuters Breakingviews (December 4),
accessed 4 April 2019 at http://blogs.reuters.com/breakingviews/2015/12/04/chinas-music-dream-is-out-of-tune
-with-reality/.
Moullier, Bertrand. 2007. ‘Whither Bollywood? IP Rights, Innovation, and Economic Growth in India’s
Film Industries,’ Creative and Innovative Economy Center, George Washington University Law School,
Study.
Nadel, Mark S. 2004. ‘How Current Copyright Law Discourages Creative Output: The Overlooked Impact of
Marketing,’ 19 Berkeley Technology Law Journal 785–856.
Napoli, Philip M. 2012. ‘Diminished, Enduring, and Emergent Diversity Policy Concerns in an Evolving
Media Environment,’ in Sean A. Pager and Adam Candeub, eds., Transnational Culture in a Digital Age.
Netanel, Neil Weinstock. 1996. ‘Copyright and a Democratic Civil Society,’ 106 Yale Law Journal 283–387.
Netanel, Neil Weinstock. 1998. ‘Asserting Copyright’s Democratic Principles in the Global Arena,’ 51 Vanderbilt
Netanel, Neil Weinstock. 2000. ‘Market Hierarchy and Copyright in Our System of Free Expression,’ 53
Nigerian Copyright Commission. 2017. ‘About NceRS,’ Nigerian Copyright Commission (August 17), accessed
4 April 2019 at http://www.eregistration.copyright.gov.ng/ncc/about.
Opara, Jackie. 2017. ‘Court Ruling on IP Struggle between Movie Producers Shows Level of Copyright
Awareness in Nigeria,’ IP Watch.org (April 24), accessed 4 April 2019 at https://www.ip-watch.org/2017/04/24/
court-ruling-ip-struggle-movie-producers-shows-level-copyright-awareness-nigeria/.
Pager, Sean. 2006. ‘TRIPS: A Link Too Far? A Proposal for Procedural Constraints on the Use of Cross-Issue
Linkages to Advance Regulatory Harmonization in the WTO,’ 10 Marquette Intellectual Property Law Review
215–537.
Pager, Sean. 2011. ‘Beyond Culture vs. Commerce: Decentralizing Cultural Protection to Promote Diversity
Through Trade,’ 31 Northwestern Journal of International Law and Business 63–135.
Pager, Sean. 2012a. ‘Accentuating the Positive: Building Capacity for Creative Industries into the Development
Agenda for Global Intellectual Property Law,’ 28 American University International Law Review 223–94.
Pager, Sean. 2012b. ‘Folklore 2.0: Preservation Through Innovation,’ 2012 Utah Law Review 1835–95.
Pager, Sean. 2012c. ‘Digital Content Production in Nigeria and Brazil: A Case for Cultural Optimism?,’ in
Sean A. Pager and Adam Candeub, eds., Transnational Culture in the Internet Age. Cheltenham, UK and
Pager, Sean. 2015. ‘Making Copyright Work for Creative Upstarts,’ 22 George Mason Law Review 1021–55.
Paulson, Cole. 2012. ‘Marketers and Pirates, Businessmen and Villains: The Blurred Lines of Nollywood
Distribution Networks,’ 7 St Antony’s International Review 51–68.
Perzanowski, Aaron, and Kate Darling. 2017. ‘Introduction,’ in Aaron Perzanowski and Kate Darling, eds.,
Creativity without Law: Challenging the Assumptions of Intellectual Property. New York, NY: New York
University Press.
Pessach, Guy. 2003. ‘Copyright Law as a Silencing Restriction on Noninfringing Materials: Unveiling the Scope
of Copyright’s Diversity Externalities,’ 76 Southern California Law Review 1067–104.

Pessach, Guy. 2013. ‘Deconstructing Disintermediation: A Skeptical Copyright Perspective,’ 31 Cardozo Arts
and Entertainment Law Journal 833–73.
Pessach, Guy. 2016. ‘Beyond IP: The Cost of Free – Informational Capitalism in a Post IP Era,’ 54 Osgoode
Hall Law Journal 225–51.
Pratt, Lande N. 2015. ‘Good for “New Hollywood”: The Impact of New Online Distribution and Licensing
Strategies,’ 3 International Journal of Cultural and Creative Industries 70–84.
Priest, Eric. 2014. ‘Copyright Extremophiles: Do Creative Industries Thrive or Just Survive in China’s High-
Piracy Environment?,’ 27 Harvard Journal of Law & Technology 467–541.
Priest, Eric. 2015a. ‘Copyright and Free Expression in China’s Film Industry,’ 26 Fordham Intellectual Property,
Media and Entertainment Law Journal 1–69.
Priest, Eric. 2015b. ‘Acupressure: The Emerging Role of Market Ordering in Global Copyright Enforcement,’
68 SMU Law Review 169–242.
Priest, Eric. 2016. ‘Meet the New Media, Same as the Old Media: The Real Lessons from China’s Digital
Copyright Industries,’ 23 George Mason Law Review 1079–92.
Rajadhyaksha, Ashish. 2008. ‘The “Bollywoodization” of the Indian Cinema: Cultural Nationalism in a Global
Arena,’ in Anandam P. Kavoori and Aswin Punathambekar, eds., Global Bollywood. New York, NY: New
York University Press.
Rajadhyaksha, Ashish. 2014. ‘The Guilty Secret: The Latter Career of the Bollywood’s Illegitimacy,’ Asia
Research Institute, Working Paper Series No. 230, December.
Raustiala, Kal, and Christopher Sprigman. 2019. ‘Negative Space,’ in Peter S. Menell and David L. Schwartz,
eds., Research Handbook on the Economics of Intellectual Property, vol. I. Cheltenham, UK and Northampton,
Reidel, Ivan. 2011. ‘The Taylor Swift Paradox: Superstardom, Excessive Advertising and Blanket Licenses,’ 7
NYU Journal of Law & Business 731–809.
Rhodes, Dara. 2014. ‘Nigeria’s Music Industry Set to Hit a High Note’ (May 28), accessed 4 April 2019 at https://
www.cnbcafrica.com/news/west-africa/2014/05/28/nigerias-music-industry-set-to-hit-a-high-note/.
Rizk, Nagla. 2014. ‘From De Facto Commons to Digital Commons? The Case of Egypt’s Independent Music
Industry,’ in Jeremy de Beer, Chris Armstrong, Chidi Oguamanam, and Tobias Schonwetter, eds., Innovation
& Intellectual Property: Collaborative Dynamics in Africa. South Africa: UCT Press.
Rutschman, Ana Santos. 2015. ‘Weapons of Mass Construction: The Role of Intellectual Property in Nigeria’s
Film and Music Industries,’ 29 Emory International Law Review 673–704.
Sanchez, Dana. 2014. ‘How Big Is Africa’s Music Industry? Ringtone Sales in Nigeria: $150M’ (August 8), accessed
4 April 2019 at http://afkinsider.com/68131/big-africas-music-export-industry/#sthash.FbP9y8Iv.dpuf.
Sang-hun, Choe. 2017. ‘South Korea’s Blacklist of Artists Adds to Outrage over a Scandal,’ New York Times,
January 13.
Santos, Ana. 2013. ‘Nurturing Creative Industries in the Developing World: The Case of the Alternative
Systems of Music Productions and Distribution,’ 21 Michigan State International Law Review 601–779.
Sawicki, Andres, and Anthony J. Casey. 2013. ‘Copyright in Teams,’ 80 University of Chicago Law Review
1683–741.
Scaria, Arul George. 2013. ‘Online Piracy of Indian Movies: Is the Film Industry Firing at the Wrong Target?,’
21 Michigan State International Law Review 647–63.
Scaria, Arul George. 2014. Piracy in the Indian Film Industry. Delhi: Cambridge University Press.
Schiller, Dan. 1999. Digital Capitalism: Networking the Global Market System. Cambridge, MA: The MIT Press.
Schlatter, Sibylle E. 2005. ‘Copyright Collecting Societies in Developing Countries: Possibilities and Dangers,’ in
Christopher Heath and Anselm Kamperman Sanders, eds., New Frontiers of Intellectual Property Law: IP and
Cultural Heritage – Geographical Indications – Enforcement – Overprotection. Portland, OR: Hart Publishing.
Schultz, Mark F. 2009. ‘Live Performance, Copyright, and the Future of the Music Business,’ 43 University of
Richmond Law Review 685–765.
Schultz, Mark F. 2012. ‘The Nigerian Film Industry and Lessons Regarding Cultural Diversity from the Home-
Market Effects Model of International Trade in Films,’ in Sean A. Pager and Adam Candeub, eds., Transnational
Culture in a Digital Age. Cheltenham, UK and Northampton, MA, USA: Edward Elgar Publishing.
Schultz, Mark, and Alec van Gelder. 2008. ‘Creative Development: Helping Poor Countries by Building
Creative Industries,’ 97 Kentucky Law Journal 79–148.
Shur-Ofry, Michal. 2012. ‘Copyright, Complexity and Cultural Diversity: A Skeptic’s View,’ in Sean A. Pager
and Adam Candeub, eds., Transnational Culture in a Digital Age. Cheltenham, UK and Northampton, MA,
Shur-Ofry, Michal. 2014. ‘IP and the Lens of Complexity,’ 54 IDEA: The Intellectual Property Law Review
55–102.
Singh, Jyotsna. 2013. ‘Booming Bollywood Turns 100,’ Financial Times, May 5.
Skladany, Martin. 2008. ‘Alienation by Copyright: Abolishing Copyright to Spur Individual Creativity,’ 55
Journal of the Copyright Society of the U.S.A. 361–92.

Smith, Michael, and Rahul Telang. 2016. Streaming, Sharing, Stealing: Big Data and the Future of Entertainment.
Song, Mengxing. 2016. ‘China Shifts Toward Paid-Service Model for Audio Services,’ China Daily, COM.CN
(April 20), accessed 5 April 2019 at http://www.chinadaily.com.cn/cndy/2016-04/20/content_24686138.htm.
Song, Seagull Haiyan. 2011. ‘Reevaluating Fair Use in China: A Comparative Copyright Analysis of Chinese
Fair Use Legislation, the US Fair Use Doctrine, and the European Fair Dealing Model,’ 51 IDEA 453–89.
Story, Alan. 2003. ‘Burn Berne: Why the Leading International Copyright Convention Must be Repealed,’ 40
Stout, Kristie Lu. 2016. ‘Blockbuster Battle: Is China’s Movie Industry the New Hollywood?,’ CNN (June 23),
accessed 4 April 2019 at http://www.cnn.com/2016/06/23/asia/china-entertainment-movie-industry/.
Strandburg, Katherine J. 2013. ‘Free Fall: The Online Market’s Consumer Preference Disconnect,’ 2013
University of Chicago Legal Forum 95–172.
Sun, Rebecca. 2016. ‘Can China Make a Movie the Whole World Will Love?,’ The Hollywood Reporter (November
2), accessed 4 April 2019 at http://www.hollywoodreporter.com/news/can-china-make-a-movie-world-will-love
-943345.
Sunder, Madhavi. 2012. From Goods to a Good Life: Intellectual Property and Global Justice. New Haven: Yale
University Press.
Telang, Rahul, and Joel Waldfogel. 2014. ‘Piracy and New Product Creation: A Bollywood Story,’ Carnegie
Mellon University, Working Paper (August 6), accessed 4 April 2019 at https://papers.ssrn.com/sol3/papers.
cfm?abstract_id=2478755.
Thussu, Daya Kishan. 2008. ‘The Globalization of “Bollywood”: The Hype and Hope,’ in Anandam P. Kavoori
and Aswin Punathambekar, eds., Global Bollywood. New York: New York University Press.
UNCTAD (United Nations Conference on Trade and Development and United Nations Development
Programme). 2010. Creative Economy Report 2010, Creative Economy: A Feasible Development Option.
UNCTAD (December 14).
von Lohmann, Fred. 2016. ‘Google on What is Driving Creativity and Innovation in the Digital Economy,’
WIPO Magazine, April.
Voon, Tania. 2007. Cultural Products and the World Trade Organization (Cambridge Studies in International and
Comparative Law Series). Cambridge: Cambridge University Press.
Waldfogel, Joel. 2007. The Tyranny of the Market: Why You Can’t Always Get What You Want. Cambridge, MA:
Waldfogel, Joel. 2019. ‘Copyright and Technological Change in Music, Movies, and Books,’ in Peter S.
Menell and David L. Schwartz, eds., Research Handbook on the Economics of Intellectual Property, vol. I.
Walker, Robert Kirk, and Ben Depoorter. 2015. ‘Unavoidable Aesthetic Judgments in Copyright Law: A
Community of Practice Standard,’ 109 Northwestern University Law Review 343–81.
Whaley, Floyd. 2012. ‘New Ambitions in Philippine Film Business,’ The New York Times (May 7), accessed 4
April 2019 at http://www.nytimes.com/2012/05/08/business/global/new-ambitions-in-philippine-film-business.
html?_r=0.
Wildman, Steven S., and Stephen E. Siwek. 1988. International Trade in Films and Television Programs
(American Enterprise Institute Trade in Services Series). Pensacola, FL: Ballinger Publishing Company.
Yi, Na. 2016. ‘Ethnic Cultural Injuries and the “One Place, One Product” Strategy,’ in Michael Keane, ed.,
Handbook of Cultural and Creative Industries in China. Cheltenham, UK and Northampton, MA, USA:
Young, S. Mark, James J. Gong, and Wim A. Van der Stede 2010. ‘The Business of Making Money with Movies,’
February 2010 Strategic Finance, 35–40, accessed 17 April 2019 at https://sfmagazine.com/wp-content/
uploads/sfarchive/2010/02/The-Business-of-Making-Money-with-Movies.pdf.
Yu, Peter K. 2006. ‘TRIPs and Its Discontents,’ Marquette Intellectual Property Law Review 369‒410.
Yu, Peter K. 2007. ‘Intellectual Property, Economic Development, and the China Puzzle,’ in Daniel J. Gervais,
ed., Intellectual Property, Trade and Development: Strategies to Optimize Economic Development in a TRIPS-
Plus Era. New York: Oxford University Press.
Yu, Peter K. 2009. ‘The Global Intellectual Property Order and Its Undetermined Future,’ 1 The WIPO Journal
1–15.
Zimmerman, Diane. 2011. ‘Copyright Incentives: Did We just Imagine That?,’ 12 Theoretical Inquiries in Law
29–58.

23. Intellectual property and economic development:
a guide for scholarly and policy research
Shubha Ghosh* 95
Contents
I. Introduction
II. Models and Frames for Understanding Economic Development
III. Intellectual Property: Law and Institutions
IV. Mapping the Contours of What We Know About Law and Economic
Development
A. Macroeconomic Studies
B. Microeconomic Studies
V. Directions for Future Research
References
I. INTRODUCTION
State creation of intellectual property rights (IPRs) can be motivated by many goals.
Under the US Constitution, the primary goal is promotion of progress in the development
of knowledge and applied fields, referred to as Science and the Useful Arts. Under the
Agreement on Trade-Related Aspects of Intellectual Property Rights (TRIPS Agreement),
the ostensible goal is to promote technology transfer and innovation. Historically, differ-
ent jurisdictions have adopted intellectual property laws to pursue goals of industrial
development and domestic wealth creation (see, e.g., Sybaris, Florence, Venice, England).
Conceptually, IPRs promote economic development through incentivizing the invention
of technologies and creation of works that can spur knowledge creation. These new
technologies and works in turn contribute to production of new industries that increase
the growth of economic metrics such as national income (Cooter and Schäfer, 2012;
Deaton, 2013; Neal and Williamson, 2014). Although the connection between economic
development and patent law in the United States has become tenuous as a matter of doc-
trine and policy starting in the nineteenth century, the debate over how patents promote
industries continues among scholars and has implications for global intellectual property
law (Hovenkamp, 2016).
This chapter presents research methodologies and questions for the study of economic
development and IPRs. A separate chapter deals with the relationship between IPRs and
* Crandall Melvin Professor of Law and Director, Syracuse Intellectual Property Law Institute
(SIPLI) and IP & Tech Commercialization Curricular Law Program, Syracuse University College
of Law.
636

A guide for scholarly and policy research 637
cultural development. By economic development, we mean broadly metrics of material

success such as national income, industry profitability, and measures of well-being, such as
literacy, life expectancy, and physical well-being. Cultural development, by contrast, refers
to contributions to symbolic and social meanings, in the form of artifacts such as film,
literature, music, and other forms of communication. Cultural and economic develop-
ment are interrelated (Mokyr, 2017). Studies of both economic development and cultural
development focus on the national level. So economic development refers to the material
well-being of people within specific geographic boundaries, determined by history and a
shared sense of culture. Cultural development refers to the growth of a national identity
as defined through the particular artifacts defined earlier in this paragraph. Whether
there is a meaningful notion of global economic and cultural development we leave for
later discussion. The tension between national and global economic development will be
touched upon later in this chapter. Such a tension has implications for transnational legal
institutions as well as for disputes within specific nations.
Economic development, at an initial glance, has a linear relationship with IPRs. The
cadre of IPRs—copyrights, patents, trademarks, trade secrets, and related areas—creates
a legal and social environment for creation, invention, and the distribution of new works
and technologies. This environment can work at the micro level of incentives and at the
macro level of defining institutions that support innovation. As innovation occurs, new
industries emerge and economic growth follows. Economic growth in turn spurs economic
development.
While the linear model connecting IPRs to economic development through inven-
tion, creation, and economic growth provides a starting point for analysis, it also can
be misleading. Whether IPRs actually incentivize invention is a contentious question.
More likely, IPRs direct the commercialization of invention rather than spark creation.
Furthermore, economic growth is the result of variables other than technological change,
such as social institutions that promote human capabilities and effective governance.
Finally, IPRs may be the result of economic growth as interest groups seek to use the law
to either extend intellectual property protection to other regions or to expand the scope of
existing rights. The linear model serves as a frame for assessing complex economic, legal,
and social relationships synchronically and diachronically. Basberg (1987) and Griliches
(1990) both speak to the use of patent statistics to gauge the impact of patents on growth
and other economic variables.
Adding to the debate is determining how to measure economic growth. National
income measures often disguise distributional disparities and social metrics, such as
literacy, health, political participation, and cultural progress (Fields, 2001; Nussbaum and
Sen, 1993). Even if the linear model is accurate, questions would still arise as to how IPRs
affect different dimensions of development, quantitative and qualitative.
This chapter will untangle these complex issues by considering the question of the role
of law in economic development, broadly, and the specific question of IPRs. These two
questions are addressed first in terms of theory and second in terms of examples, drawn
from the extensive literature. Together, these two parts of the inquiry will aid in framing
a rich policy and research agenda for assessing and reforming IPRs in its promotion of
economic development.

II. MODELS AND FRAMES FOR UNDERSTANDING

ECONOMIC DEVELOPMENT
Classical theories of economic development typically ignore law. The theoretical over-
view by Joan Robinson (1979) illustrates how economic development was more firmly
grounded in economic investments of capital and labor rather than in legal regimes of
rights and duties. Neoclassical theories of economic growth, starting with the work of
Robert Solow (1956, 1988), reduce the growth rate of national income to the population
growth rate, the growth rate of capital, and rate of technological change, referred to as the
Solow residual (Spiegler, 2015, pp. 28–31). The rate of technological change was viewed
as exogenous and outside the control of economic policy. Population growth rate was
also considered exogenous as well, determined by sociological factors such as marriage,
fertility, and mortality.
As Schmookler points out, however, the problem with neoclassical theories of economic
growth is that they ‘may leave scientific and technological progress unexplained’ (1972,
p. 84). Refinements of the model incorporate factors that can affect technological change
and population growth such as human capital as measured by literacy, life expectancy,
and measures of health and well-being. These refinements potentially identify policy tools
to reach certain goals for economic growth and development. The qualification in the
word ‘potentially’ reflects certain limitations implicit in the models of the role of policy
intervention in altering predicted growth rates.
One specific limitation within the neoclassical models of economic growth is the
prediction of a long run convergence across countries in the rate of economic growth.
This prediction stems from an assumption that countries will converge in the underlying
growth rate of the population and technological change. These two factors are assumed
to converge globally as improvements in technology spread across countries resulting
in improvements of literacy, reductions in fertility and mortality, and diffusion of
knowledge. However, impediments to global diffusion, and the predicted convergence,
such as failures of family planning programs, difficulties in realizing improvements in
mortality due to failures in health care delivery, and household and family dynamics that
impede improvements in education across generations can result in gross disparities both
across countries and within countries in the levels of economic growth and development.
Professor Jeffrey Sachs (Sachs and Warner, 1997) refers to these disparities as the result of
poverty traps that require government intervention in the creation of educational, family
planning, and health care infrastructure to put developing countries back on track to
reach the economic heights achieved by developed countries.
Despite variations in neoclassical growth models that identify poverty traps and
tendencies to global disparities, the underlying prediction is one of global convergence.
Government intervention, largely targeting variables that influence population growth,
would be a necessity to attain the predicted global convergence. However, little attention
is placed within these models on technological change. This variable is viewed largely as
exogenous, the product of social forces that may be largely outside the control of adjust-
ing economic variables such as prices and incomes. At the same time, the neoclassical
model assumes that technology will diffuse, resulting in improvements in technological
change across countries. Needless to say, the rate of technological change does vary across
countries. Invention varies both in quantity and quality between the developing and

developed worlds. While the assumption of free trade, namely the unimpeded movement
of products and factors of production, such as capital and labor across borders, may lead
to technological diffusion, free trade is neither a reality nor a norm. Furthermore, even
if goods and factors did move freely across borders, there may be reasons for owners of
technology to protect their know-how from moving across borders.
Neoclassical growth theorists measure cross-national disparities in technology through
the Solow residual. A statistical concept, the residual measures changes in economic
growth that cannot be explained quantitatively by growth rates of capital and labor.
Measures of the residual demonstrate disparities across countries in their respective
growth rates and serve to counter theoretical predictions of global convergence. The
challenging questions are how to account for differences in the Solow residual and how
policy can affect differences in the rate of technological change. One possible explanation
is offered by the New Growth Theory, which emerged in the late 1980s and early 1990s.
The New Growth Theory, attributed to Paul Romer (1994) and to Robert Barro (Barro
and Sala-i-Martin, 2003), examines technological change as an endogenous variable, the
product itself of economic choices. Technological change produces increasing returns
to scale with past change having a magnifying feedback effect on growth. Technological
change creates externalities, or spillovers, within an economy, and government policy can
shape how these spillovers can be managed to influence economic growth. In contrast
with the neoclassical theory of growth, which focused on government policy to improve
rates of population growth, policy interventions under the New Growth Theory are
oriented to the ‘supply side’ of the economy, namely firms and industries that generate
new technologies and improvements to methods of producing and diffusing new products
and services. Government policy through taxes and subsidies can stimulate the supply side
to promote improvements in the rate of technological change. Furthermore, government
programs to develop infrastructure, such as roads and bridges, can help realize returns to
scale and spillovers for the benefit of potentially innovating firms.
While the New Growth Theory does not explain the Solow residual, it does aid
in unpacking it by shedding some light into the black box of technological change.
The models are needless to say abstract but do address several simplifying features of
the neoclassical growth model. Foremost, while the neoclassical model assumed that the
underlying production function does not change, the New Growth Theory recognized
that variables can change the shape of the function relating inputs such as capital and
labor to outputs such as goods and services. This shift in assumptions raises questions
about how the production function can change and how such changes can explain
ongoing disparities across countries. While the production function is an abstraction,
conceptually it captures the range of social institutions and practices that determine how
an economy functions, grows, and develops. Consequently, what the neoclassical theory
took as a black box, whose contours can be gauged through the Solow residual, the New
Growth Theory takes as something to be explained and possibly altered through a range
of government policies, beyond those that affect population growth.
Although not expressly focused on legal institutions, the New Growth Theory offers a
potential avenue for law to influence economic growth. As legal institutions can serve to
control externalities, law arguably would be relevant to the direction and level of economic
growth. The Legal Origins literature makes the case, for example, that common law
systems are more amenable to economic growth than civil law systems. While the evidence

for this claim is contested, the argument shows that law might have some influence on
economic variables, contrary to traditional theories of economic growth. However, the
contested empirical literature demonstrates the complex dynamics that belie that simple
linear model with which this chapter began. On the one hand, opening the black box of
technological change creates new possibilities for understanding growth and development
but also geometrically expands the set of potential policy tools for shaping the trajectory
of changing technology. Perhaps the initial light allowed into the black box is simply
blinding. More realistically, the light reveals a vast array of possibilities that may be
paralyzing for policy makers and governmental actors.
Several examples illustrate the cavernous layers within the black box of technological
change. The New Growth Theory harkens back to the work of Nelson and Winter (1983)
on economic theories of technological change that adopt an evolutionary and transaction
costs approach to peer inside the black box of the firm. The work of Nelson and Winter
in turn parallels transaction cost economics going back to Ronald Coase’s theories of
the firm and social costs (1937, 1960), to Oliver Williamson’s transaction cost approach
to contracts and corporate governance (2013), and to work in legal sociology by Stewart
McCauley (1963) and Ian MacNeil (1978, 1980). Through McCauley’s work on relational
contracting, we can further go back to the work of legal historian Willard Hurst (1956)
on the role of law in creating the conditions for freedom and business development. What
these seemingly disparate strands have in common is the development of contextualized
understanding of how a firm operates and how business practices foster growth and devel-
opment. Law plays varied roles in these scholarly examinations from creating conditions
for freedom to providing the background against which social and economic relations take
shape. Currently, these various strands arguably coalesce in the work of New Law and
Development scholars which draws on these extensive socio-legal and economic insights
to understand the relationship between law and development.
The New Law and Development literature, spearheaded by Trubek and Santos (2006),
promotes the claim that legal institutions, combined with social institutions, are needed
to direct economic development. This literature offers an alternative to purely invest-
ment based policies for stimulating development and to dogmatic rule of law policies as
providing the sole institutional prerequisite for economic growth. According to Trubek
and Santos, fashioning law to promote development is a contextual inquiry, dependent
on the particular historical, cultural, and social background within which development
policy will operate. The New Law and Development, the product of the scholarly and
policy activities of legal and sociological scholars, presents the multifaceted connections
between law and development. Unlike the New Growth Theory that presents important
refinements to a theoretical model of growth, the New Law and Development literature
focuses on country specific case studies that demonstrate how actual governments and
actual firms function to promote technological change through contracting, finance,
collaborative research and development, and partnerships between industry and govern-
ment. While the New Growth Theory raises questions, the New Law and Development
literature proposes to find specific answers that demonstrate what types of government
policies and business practices help to promote development.
There is a family resemblance and close intellectual ties between the New Law and
Development scholarship and the work of Professor Charles Sabel, a legal sociologist, on
the theory and mechanics of innovation (2005). Through independent scholarship and

intellectual collaborations with Professors Ronald Gilson and Robert Scott, Sabel dem-
onstrates how contracting practices can promote innovation (Gilson et al., 2013). While
the work of Gilson, Scott, and Sabel does not directly address questions of intellectual
property policy, their work is crucial to understanding the business practices that support
innovation and are instrumental in further unpacking the black box of technological
change. As a result there are potentially synergies and connections between Charles
Sabel’s scholarship and the New Law and Development literature.
Three distinct but overlapping lines of scholarship present models and frames for
delineating the relationship between economic development and legal institutions.
Neoclassical growth theory posits economic development as the product of technological
change, largely unknowable and purely exogenous, and population change, more malle-
able through social and economic policy. Light is cast on the black box of technological
change through the New Growth Theory that emphasizes externalities, spillovers, and
variations in the relationship between economic inputs and outputs, cast abstractly as a
production function. This framework provides a role of government policy and law to
shape the supply side of an economy beyond, and perhaps instead of, shifts in sociological
behavior affecting population growth. Finally, the New Law and Development breaks
open the black box of technology by exposing light to its nooks and crannies through
detailed contextualized studies of actual policies and practices.
These methodologies provide useful frames, but also leave many questions. Foremost
for the purposes of this chapter is the role of intellectual property as a set of institutions
grounded in law for economic development. None of the three approaches directly
address this question, but each can provide a background for engaging the policy
debates over intellectual property institutions. Debates over intellectual property law and
economic development center on how contextualized IPRs should be in order to reach
goals of development. The 2002 Commission on Intellectual Property Rights study, for
example, reflects such a contextualized approach in the broader context of the World
Trade Organization and TRIPS. The study identifies the various avenues through which
IPRs can affect economic development. It also highlights the complex set of objectives
captured by the word development, such as access to health care, education, and income
growth. Although not couched within the three theoretical frameworks presented in this
section, assessments of the 2002 Report and other contributions to the policy debates over
intellectual property and economic development can best occur against the theoretical
landscape of this section. Such assessments are the subject of the next section, which
addresses the debates over the design of intellectual property as a set of legal institutions
for the promotion of economic development.
III. INTELLECTUAL PROPERTY: LAW AND INSTITUTIONS
Neoclassical growth theory, the New Growth Theory, and the New Law and Development
scholarship frame the discussion of how IPRs shape economic development. While
neoclassical growth theory would turn a policy maker’s attention to population growth
and technological change as determinants of growth and development, the New Growth
Theory emphasizes returns to scale and spillovers that stimulate technological change.
The New Law and Development approach, by contrast, would frame questions of

development contextually and the role of policy makers in shaping laws and institutions
to reach national economic goals.
The 2002 Commission Report on intellectual property and development reflects each of
these perspectives to different degrees. Technology transfer serves to increase the capital
stock and the technological transformation of economies in developing countries, according
to one aspect of the Report. Consistent with the neoclassical growth theory, capital expan-
sion and the adoption of new technologies promotes economic growth through investments
in a country’s production function. In addition, technology transfer can promote spillovers
as indicated by the varying effects of intellectual property protection. Weaker IPRs tend to
spark economic growth and development for low income countries, and intellectual property
protection becomes stronger as national income increases. This finding is consistent with the
predictions of knowledge spillovers in the New Growth Theory as the spillover effects may
be greater for countries in the early stages of development. Finally, the 2002 Report extols
a country specific and contextualized perspective that takes into consideration history,
natural resources, and political institutions in the tailoring of IPRs to the goals of a specific
country. This last view is consistent with the new development economics.
While the 2002 Report, and other scholarship discussed here, may not have consciously
drawn on these three theoretical approaches, the theories can aid in gauging and under-
standing the policy reports and scholarly analyses. The big problem for policy analysis
is navigating between abstract models that might shed general insights on intellectual
property and development and disparate case studies that may point to successes and
failures of policies without the benefit of generalizable lessons across countries. Drawing
on a rich sociological literature, Trebilcock and Prado (2014, p. 42) state the problem as
follows:
The scholarly community has reached a relatively broad consensus around the idea that institu-
tions matter for development, but we still do not know how to reform dysfunctional institutions.
Even if certain ends of development were thought to be universal (which is contestable), the
different starting points or initial conditions of the large universe of developing countries may
suggest that there will be very few universally effective blueprints with respect to substantive
economic and social policies and the processes and institutions appropriate for implementing,
administering, and fostering such policies.
Although universal blueprints may not exist, scholarly study of development and IPRs
across countries drawing on econometric methods, historical analyses, studies grounded
in political science and sociology, and integrative case studies can shed some light on
whether and how IPRs affect economic development. Examples of such successful
approaches are collected in the edited volume by Aoki et al. (2012), which examines
institutions and economic development from a comparative economics perspective.
That institutions matter is a foundational idea for this chapter, pointing towards
contextualized analyses that built upon more abstract models. But as with any framework
that attempts to identify explanatory variables, the difficult question is identifying which
institutions matter. At a general level, law matters for development, but that generalization
invites identification of which legal institutions are critical. The rest of this section will
focus on this question in light of the scholarly literature and lay the foundation for the
discussion of more detailed scholarly work in the remaining two sections.
The three theories of development start with the nation-state as the unit of analysis.

Decision making is aggregated at the national level, and growth is determined by aggregate
levels of capital, labor, and technological change. The frameworks, except for the studies
within the New Law and Development school, abstract from particular industries or mar-
kets. Even if one accepts this gross level of description, the question still exists whether
the nation is the appropriate unit of analysis. Regional, or subnational, development has
been the focus of attention as some explore ways to replicate the success of Silicon Valley
(Casper, 2007). But the implicit assumption of the nation-state as the appropriate unit
of analysis is striking given the international treaty environment, whether the multilateral
harmonization of intellectual property laws through TRIPS or the bilateral negotiations
for the establishment of rule of law institutions, such as courts and democratically elected
legislatures. Mandel (2014) has commented on the nation-state focus of innovation in the
debate over TRIPS implementation. He develops an interest group analysis of nation-
states as a means of predicting what type of intellectual property regime a country would
prefer based on its pattern of exports and imports. While it is not completely clear how
these predictions would be implemented in policy, Mandel does highlight the nation-state
emphasis of most analyses of innovation.
The literature on global public goods (see Maskus and Reichman, 2005) provides a
framework for conceptualizing innovation in global as opposed to solely national terms.
Practical inventions that are the subject matter of patent law and the expressive works
protected by copyright are viewed as public goods that benefit all individuals. However,
these benefits are not limited to individuals within a nation-state. Scientific inventions,
entertainment products, information goods—all travel worldwide beyond the points
of their creation. Analyses of intellectual property policy need to consider the concept
of global public goods, which may be one way of formulating the global commons of
knowledge that all individuals, regardless of nationality, share.
What the policy implications of global public goods are is a complex question. Global
public goods are a form of transnational spillover, suggesting parallels with the New
Growth Theory that attempts to explain technological change and innovation through
regional spillovers and industry formation. This parallel suggests that the policy implica-
tions from the New Growth Theory, such as government investment in infrastructure and
the creation of legal protections for property and contract rights would have implications
for intellectual property policy making. International agreements are ways to internalize
the spillovers from global public goods and to institutionalize through state by state
negotiation legal infrastructure for the creation and dissemination of technology. Such
interactions between states serve to facilitate technology transfer across countries as
participants in the international regime innovate.
Multilateral organizations such as the World Trade Organization and supporting
treaties such as the TRIPS Agreement can be seen as a global government that aids in the
provision of global public goods, much like national and local governments provide public
goods for their citizens. The global government serves to provide harmonized rules for the
provision of technology by nation-states and the diffusion of technology through trade.
Political and legal rules define the terms of governance for members of the organization.
These rules serve to structure the global marketplace for technology, goods, and services.
A problem with the global public goods model is the diversity of benefits provided
by the wide range of technologies that are traded in the contemporary world. While
citizens of all countries benefit from innovations in medicine, lower income countries

need support for preventing basic, life threatening diseases such as cholera, diarrhea,
and malaria. Similarly, nutritional needs vary across countries as do needs for basic
transportation and communication infrastructure. These differences in needs shape
differing meanings of innovation. In developed countries, innovation flows through new
consumer goods and forms of entertainment and processing of digital information. In
developing countries, innovation may appear more early stage in the form of basic health
care, nutritional, and infrastructure needs. These differences belie the conception of a
homogeneous, undifferentiated set of global public goods.
Intellectual property institutions more effectively serve to provide a wide of range of
heterogeneous global public goods. The failure to account for these differences is illus-
trated, for example, by Mansfield (1994) and its criticism by Heald (2003). Mansfield pre-
pared his report on the need for strong IPRs in developing countries for the International
Finance Corporation. His argument for strong rights to promote technology transfer
has served as strong rhetoric for an expansionist view of TRIPS and for the promotion
of Western style IPRs in developing countries. Heald’s criticism of the article reveals
many veiled assumptions in the Mansfield survey of multinational businesses, especially
the failure to distinguish between different types of intellectual property. Specifically,
Heald points out that the multinational businesses surveyed were interested not so much
in technology transfer as in opening up new markets for their products. Strong IPRs
prevented the development of local industries that would copy, improve, and compete
with the importing companies.
The Doha Declaration from developing countries advocates for a development
agenda under the TRIPS Agreement to promote intellectual property institutions that
meet the development needs of low income countries. The World Intellectual Property
Organization has, from its inception, focused on the needs of developing countries. With
the Doha Declaration, the TRIPS Agreement has been refocused on the needs of develop-
ing countries, contra reports by scholars such as Mansfield.
The 2002 Commission Report reflects many of these goals. First, the Report adopts
an industry specific, or perhaps a technology specific, approach to IPRs with separate
emphasis on health care, educational, and agricultural needs in developing countries.
Second, the Report adopts a more analytical approach to intellectual property, question-
ing the linear model connecting IPRs, growth, and economic development. Third, this
analytical approach takes into consideration both the economics and politics of intel-
lectual property institutions and the needs of particular developing countries and regions
in designing IPRs. Finally, the Report frames IPRs in terms of the path of development
with the strength of intellectual property evolving with advances in income and economic
well-being. With the 2002 Report, the global public goods model has evolved into a richer
analytical framework for understanding the heterogeneous benefits and technologies.
Putting the various strands of scholarship together, we start with the highly analytically
abstract model of neoclassical growth theory which reduces economic growth to changes
in population and technological change. Starting from this useful, but decontextualized,
model, we move towards approaches that add more degrees of context and analytical rigor.
The New Growth Theory attempts to unpack the black box of technological change, and
the New Law and Development literature focuses on contextualized case studies. The
strands of the theoretical literature map onto the continuing debate over intellectual
property institutions and economic development. Where we are now appears to be a

need to recognize technological change and global public goods in heterogeneous terms
taking account of a range of interests and institutional contexts in developing countries.
Against this background, I turn in the second half of this chapter to empirical studies
of IPRs and economic development, looking first at macroeconomic studies and then
at microeconomic studies. The chapter closes with questions to guide future researchers.
IV. MAPPING THE CONTOURS OF WHAT WE KNOW ABOUT

LAW AND ECONOMIC DEVELOPMENT
This section presents an analytical overview of important empirical studies of the relation-
ship between law and economic development. While the second volume of this handbook
turns our attention to empirical methods in scholarship, the purpose of this section is to
assess results, rather than methods. On occasion, I will point to methodological issues to
highlight the substantive theoretical concerns about the various pathways and variables
that connect law and economic development.
The scholarly empirical literature divides into macroeconomic and microeconomic
studies. These broad categories capture respectively the high level economic variables that
aggregate market and social behavior, such as trade patterns and national income, and the
low-level variables on the ground, that describe behavior within a firm or a household.
Empirical scholarship is limited often by available data, which is often gathered through
aggregate government statistics or lower level surveys of individuals, whether managers in
a company, farmers, or consumers. However, theoretical considerations guide the statisti-
cal analyses of available data sources. In surveying the literature, I glean broader lessons
for the theoretical discussion presented in Section III. These lessons in turn inform policy
options for shaping development policy nationally and globally.
What the literature suggests is that the case for the macroeconomic effects of intellec-
tual property law on promoting trade and foreign direct investment is at best a mixed one.
Stronger IPRs do not necessarily invite more foreign investment. Creating an environment
hospitable to foreign investment is not solely a matter of intellectual property law, but
also one of institutional arrangements that create an environment for markets. At the
same time, the creation of market- and investment-friendly environments comes at a cost
for issues of distribution, cultural change, and social stratification. Micro-level studies,
particularly the work of Banerjee and Duflo (2011), call our attention to these broader
institutional concerns by examining incentives and the structure of social organization.
Consequently, I conclude, future research should explore more completely how law and
institutions shape each other and their influence on the course of economic development.
The remainder of this chapter adds substance and details to these arguments.
A. Macroeconomic Studies
Maskus and Fink (2005) summarize the empirical literature on the relationship between
trade and IPRs, between foreign direct investment and IPRs, and between licensing tech-
nology and IPRs. The scholarship they cite and include in their edited volume reminds
us of the ambiguous predictions of the theoretical analyses of economic growth and
intellectual property. Predictions depend on specific analytical models and the a ssumptions

built into these models about technologies, preferences, and population growth rates. In
addition, the empirical research highlights the importance of contextual variables such as
the presence of social institutions and the stage of development on the effect of stronger
intellectual property protections. Countries can vary on the development scale with
mid-level developing countries (such as India or China) offering a different environment
for the causal pathways from intellectual property than low-level developing countries,
such as several countries in sub-Saharan Africa or South America. Further research can
potentially identify these contextual and institutional variables.
Researchers have identified a strongly positive statistical relationship between trade
flows to developing countries and the strength of the country’s IPRs, as measured by
standard scales of legal protection. This positive relationship holds true for traded prod-
ucts that are relatively easy to copy, such as information or entertainment based works
or goods whose consumption status value derives from trademarks. However, studies
find that stronger patent protection does not necessarily lead to greater trade flows. The
explanation is that patented goods are inherently difficult to copy and so added patent
protection has a minor, and perhaps insignificant effect, on trade flows to developing
countries. Empirical studies also demonstrate differential effects between middle income
and low income developing countries. Heightened intellectual property protection affects
trade flows to middle income countries more than flows to low income developing
countries. This difference reflects the role of institutional factors in driving trade. Weaker
infrastructure and less advanced consumer markets in low income developing countries
make them less desirable as a recipient of trade flows.
Empirical support for the effect of strong IPRs on foreign direct investment is much
weaker than the effect on trade flows. Foreign direct investment includes the creation of
new business enterprises, the purchase of existing enterprises, and investment in physical
capital, such as infrastructure, machines, and buildings within developing countries. A
complicating factor is measuring these various dimensions of foreign direct investment
as business enterprises, nation-states, and governmental bodies do not collect data on
these diverse range of investments. Focusing on foreign direct investment by United States
companies, for which there is relatively stronger data, scholars have not found a strong
relationship between IPRs and investment. The weak statistical relationship reflects
in part the importance of institutional variables such as infrastructure, education and
training of the workforce, and cultural and political environment, all of which shape the
perceived risk of investing in developing countries. Stronger IPRs do not alleviate the risks
associated with foreign direct investment. Furthermore, stronger IPRs are not necessarily
associated with other dimensions of legal reform, such as more effective enforcement of
contract, or stable and predictable political systems.
One would predict a positive correlation between stronger IPRs and transborder
licensing to developing countries. Although companies in developed countries do not
tend to expand foreign direct investment to a developing country that expands its intel-
lectual property protections, there is evidence of increased licensing activity as measured
by greater flows of licensing revenues to developed countries from developing countries.
Empirical studies show some increase in licenses to companies in developing countries
that strengthen their intellectual property laws. But it is not clear from measures of
licensing revenue whether resulting increases in revenue are from higher royalty rates
and other payments or a higher volume of licensing transactions. Stronger intellectual

property protections may allow intellectual property owners to obtain more rents from
their licensing activities rather than stimulate more licenses to companies in developing
countries. How these changes in licensing activities translate into technology transfer is
not well understood.
Synthesizing the diverse studies summarized by Maskus and Fink (2005), stronger
IPRs lead to more imports of readily copied information and trademarked products into
a developing country but may not lead to increased technology transfer whether in the
form of foreign direct investment or technology licenses. Future research would need to
untangle these various effects on international transactions with an eye on the broader
question of whether stronger worldwide IPR regimes do support technology transfer
from developed to developing countries. It may be the case that intellectual property law
promotes new markets in developing countries for certain products from the developed
world.
Complexities of the relationship between intellectual property law and trade arise
with respect to parallel imports, or the reimportation of intellectual property protected
products from developing countries to developed countries. Whether electronics, books,
music CDs, movie DVDs, pharmaceuticals, or cosmetics, parallel importation rests on
a country’s choice of exhaustion rules. If the initial sale of a product exhausts global
or regional rights in patents, copyrights, or trademarks embodied in the product, then
parallel trade between countries is likely to result. Whether exhaustion and the resulting
parallel importation are socially desirable is not clear as a matter of theory. The effects
on social welfare depend on the benefits of global price discrimination and the value of
free trade.
Maskus and Fink (2005) summarize empirical studies that show differential effects
of parallel trade by industry and product type. This result is not surprising in light of
the ambiguities in the theoretical predictions. Whether price discrimination is beneficial
depends on relative sensitivities of demand to changes in price and income and on
the market structure of industries that meet the demand. Consequently, depending
on the industry, price discrimination might serve either to allow companies to extract
excessive profits from consumers or to provide a product to a market that otherwise
would not be served because of high prices. Empirical studies show that the benefits of
exhaustion can vary depending on whether rights are exhausted in all countries or only
among a few countries that are connected regionally (such as the European Union).
There seems to be some support for regional based exhaustion rules. However, there
is no empirical work on the benefits of worldwide exhaustion. The lack of such work
stems from the failure of countries to adopt an effective rule of global exhaustion, until
quite recently.
Studies of parallel importation point to the need for better understanding of the role of
market structure on development. Although the economics literature on intellectual prop-
erty rests on the debate between Schumpeterian theories of monopolies and Arrovian
theories of competition in promoting innovation, details of specific industries matter
for shaping policies to guide development. Large scale technological change may require
some degree of market concentration. But innovation can also occur through start-ups
and small companies in certain industries. Design of intellectual property laws needs to
take account of these differences across industries and markets. In addition, intellectual
property law design often must occur in tandem with the implementation of competition

laws. Both sets of law establish an interrelated environment for the promotion of innova-
tion, economic growth, and development.
Policy advocates and analysts draw on the empirical work on the relationship among
IPRs, development, and trade in many ways. Stiglitz and Charlton (2005) advise that a
strong patent system can limit economic development by impeding technology transfer.
They recommend the implementation of compulsory licensing schemes as a limitation
on patent rights in situations of global health emergencies as well as in situations where
patent owners refuse to deal. Compulsory licensing, they argue, can effectively move
innovative technologies into the hands of companies and thereby promote competition.
Yamane (2015) draws on Japanese success in promoting economic development to
examine that country’s regulation of intellectual property licensing schemes. This study
identifies what the author calls a ‘fairness’ approach to competition law based limitations
on intellectual property licensing. This approach emphasized access to technologies,
economic development, and competitive dynamics. Yamane recommends the Japanese
post-war model as one for currently developing countries to emulate.
Trade policy oriented towards development rests not only on national lawmaking and
implementation but also on cross border trade and multilateral agreements. Farley (2014)
and Yu (2015) document the proliferation of bilateral trade and investment agreements
and their role in promoting, even coercing, strong intellectual property laws in developing
countries. The authors separately caution against the conditioning of trade agreements
on the implementation of intellectual property laws that go beyond the requirements of
the TRIPS Agreement. Instead, their articles remind us of the need for contextualized
lawmaking in order to meet the development needs of individual nation-states. Appeal
to context echoes the recommendations of the 2002 Report of the Commission on
Intellectual Property Rights which concluded that the effects of trade and investment
on development depended on the particular economic progress and history of a country.
Greater benefits from trade and strong IPRs may accrue to high-level developing coun-
tries with advanced technological capacities. Mid-level or low-level developing countries
may experience little or no benefit from trade and IPRs especially if stronger rights are
used for creating new consumer markets for companies in developed countries rather than
transferring technology and know-how to developing countries.
How can policy makers in developing and developed countries shape intellectual
property and trade policy in order to promote development? The answer rests in part
on identifying the relevant contextual variables that shape the effectiveness of policy as
implemented through intellectual property and trade laws. Helpman and Krugman (1985)
provide a helpful conceptual toolkit for formulating strategic trade policy. Their analytical
structure rests on appreciating the role of market structure in capturing the benefits from
cross border trade. Some benefits accrue in greater consumption opportunities in devel-
oping countries. Some also accrue to the development of new companies and industries
that can compete in newly created markets. The authors show how targeted policy can
help promote new industries as they enter these new markets opened by trade.
Country specific studies provide helpful guidance in identifying potentially relevant
context for the success or failure of development policy. The World Bank’s collected
volume on poor people’s knowledge (2004) brings together case studies on efforts to
promote local industry and trade through intellectual property in South Asia, Latin
America, and the Caribbean. The various papers focus on efforts to create markets for

traditional knowledge and for cultural products designed to encourage tourism and busi-
nesses among indigenous communities.
Millaleo and Cadenas (2015) offer a model of country specific studies in their account
of developments in intellectual property law in Chile. Their analysis not only reveals the
political economy of legal reform but also connects changes in intellectual property to
the needs and interests of constituencies in Chile. As many intellectual property reforms
in Chile were responses to demands from free trade agreements, the case of Chile
demonstrates how efficacious trade led reforms can be. Specifically, the authors conclude
that countering the move for strengthening IPRs has been the worsening of income
inequality arising from poorer access to medicines and from limits on sharing cultures on
the Internet. These tensions poise the country for continuing debates on legal reform and
development goals.
Perullo and Eisenberg (2015) add to the richness of country specific studies in their
examination of copyright and the music industry in Tanzania and Kenya. Their compara-
tive article shows how the implementation of copyright in the two east African countries
facilitated the development of the music industry at the national and local levels with vary-
ing results. In Kenya, the music industry developed along lines that paralleled the industry
in developed countries with extensive licensing between creators and music publishers that
have tended to be concentrated and corporatized. By contrast, the national government
has taken more of a role in promoting artists and music in Tanzania, following what the
authors call a socialist model. What the study of music and copyright in Tanzania and
Kenya teaches is the range of context and forms in which cultural industries can flourish
within intellectual property systems.
Considerations of context not only should shape development policy and intellectual
property reform but also can influence conceptions of economic development. The 2002
Commission Report discussed at many points in this chapter concludes that the connec-
tions between foreign direct investment and intellectual property law may depend upon
the stage of development for a particular country. The authors of the Report also point
to how intellectual property law can affect education, food and agriculture, and public
health. Each of these variables influences growth and economic development. Schultz
(1981) provides a valuable theoretical and empirical framework supporting the connection
between growth and measures of population quality, such as healthiness and education.
Aptly titled ‘Investing in People,’ Schultz’s lectures on human health and education build
on traditional economic growth theory which looks solely at rates of population growth.
But Professor Schultz looks beyond the quantity of people in a given economy to include
other contextual factors. What the 2002 Commission Report does is show how intellectual
property law can affect measures of population quality through transfers of new technol-
ogy to benefit society through improvements in health and education.
This section has offered an analytical summary of scholarship on the path from intellec-
tual property reform to development at the macro level. Starting with research on foreign
direct investment and trade, the analysis moved to a recognition of contextual variables
that can affect the macroeconomic links between legal reform and development. Ending
with a discussion of studies that provide substance on the relevant contextual variables,
this section ends with a question. What institutional and behavioral mechanisms affect
innovation and technology transfer? This question is the focus of the research discussed
in the next section.

B. Microeconomic Studies
What studies of macroeconomic relationships between intellectual property law and

economic variables, such as trade and investment, teach is the importance of context
and institutions. Discovering the relevant context and institutions requires an analytical
approach that starts from examining behavior from the bottom up, starting from the
behavior of individuals within firms and households in developing countries. A bottom
up approach would provide insights into how invention and creativity occur, how tech-
nology is transferred and implemented across and within firms, and how returns from
innovation, growth, and development are distributed.
Banerjee and Duflo (2011) is a provocative example of a bottom up approach to law
and development. Economists at MIT, Professors Banerjee and Duflo are prominent for
their work on randomized control trials to examine through experimental techniques
the effect of various development policies on households and businesses in developing
countries. Their work is grounded in microeconomic theory enriched by sociological and
cultural studies as a response to contentious debates at the macroeconomic level on the
efficacy of United States foreign aid and development assistance policies. On the one
hand, the authors are critical of scholars like Jeffrey Sachs who point to the inexorability
of poverty traps and the need for government assistance to bring individuals out from
poverty. Professors Banerjee and Duflo express skepticism that poverty traps are inevi-
table although they acknowledge the existence of situations where extremely low income
individuals are unable to move up the economic level. Similarly, the authors are skeptical
of scholars who claim that foreign aid and development assistance are inexorably harmful
in dampening private incentives.
As a counter to these two extreme positions on foreign aid, Professors Banerjee and
Duflo analyze the behavior of poor people, identifying the psychological, social, and
economic constraints that shape their choices about financial investment, adoption of
technology, and consumption. Their methodology is rooted in part in behavioral econom-
ics, which examines how human reasoning and choice operates in practice, revealing
psychological processes ignored by pure rational choice theory with its emphasis on
comparing benefits and costs. In addition, institutional context and the lack of institu-
tions, such as banks and courts, also shape actual decision making in developing countries.
Behavioral limitations and institutional gaps help to explain why poverty traps may occur
and technology may fail to transfer for the stimulus of growth and development. Through
their empirical studies, the two development economists identify a role for conditioned aid
and grants to better pursue innovation and development goals to address psychological
and institutional constraints on investment behavior.
Although Professors Banerjee and Duflo do not directly address legal reform, their
identified impediments to investment and entrepreneurial decision making point to
institutional reform guiding law and policy. For example, their research uncovers the
problem of financing loans to poor individuals who want to form their own small business
that arises from difficulties in monitoring high risk loans. The authors make the case for
microcredit to fund small scale projects that can generate sufficient returns to help attain
profitability of business enterprises. In addition, institutional constraints limit the ability
to create start-ups and new technologies in developing economies. What they discover is
that the very poor are often faced with investment opportunities that have high marginal

returns but relatively low total returns. This disparity stems from the lack of markets
and institutions that make initial investments incrementally profitable but not sizable in
aggregate terms. Consequently, investment in infrastructure and large scale enterprises,
like factories, may have important effects on stimulating private investment in small,
individually owned and managed businesses. These investments would include a shift
from traditional to more modern technologies.
The case of a Chinese businesswoman discussed by Banerjee and Duflo illustrates
the role of institutional and behavioral constraints on technology transfer and business
development. The businesswoman opened a garment shop in her home with a single
sewing machine. Her initial investment in the sewing machine had very large marginal
returns, but the business soon reached capacity in meeting the local demand. This example
illustrates investments that have huge marginal return but low total returns. As Banerjee
and Duflo point out, this disparity between marginal and total returns describes many
business investments in developing countries. A consequence is while there are incentives
to invest in new enterprises and technologies, the investment often does not aid in suf-
ficient generation of income to allow expansion of a business.
The entrepreneur in the case study, however, enjoyed a windfall in receiving an export
order that allowed her to expand into a factory. This opportunity allowed the entrepreneur
to achieve scale and make high marginal return investments that yield high total returns
as well. This example illustrates how development policy should focus on overcoming
institutional and behavioral impediments. Legal reform, such as adoption of intellectual
property laws, is not sufficient in promoting growth and development. A further implica-
tion is that expansion in trade and foreign direct investment may not be fruitful without
addressing institutional and behavioral limitations.
Banerjee and Duflo (2011) offer an innovative methodological approach to the
microeconomics of development economics. Their attention to psychology and institu-
tions potentially supports a range of legal reforms which future research should explore.
Equally importantly, the theoretical and empirical scholarship they present complements
contemporaneous scholarly work demonstrating the role of institutions in the relationship
between law and economic development.
One illustration of institutional and behavioral understandings of development policy
and intellectual property is Nussbaum and Sen’s work on capabilities and economics
(1993). Capabilities expand the relevant measures of economic growth beyond income,
prices, and other traditional economic metrics. While these traditional metrics capture
aggregate well-being as quantified in market measures, capabilities capture improve-
ments in the standard of living as critical, but ignored development goals. Measures of
capabilities would include improvements in literacy, access to education, professional
training, health and welfare, and psychological well-being. Material metrics, such as
income or wealth, fail to acknowledge ‘the quality of life,’ the title of Nussbaum and
Sen’s volume.
There is a superficial similarity between capabilities and Schultz’s notion of population
quality, discussed in the previous section. The key difference is that Schultz views measures
of population quality, such as health and education, as inputs to obtain improvements in
income and wealth rather than as independent indicators of economic development. The
studies in Nussbaum and Sen’s 1993 volume make the case for how to gauge the quality of
life and identify aspects of the standard of living that can serve as the basis for assessing

development. Health and education are desirable independent of their contributions to

material measures of economic progress. Conceptually, income and wealth perhaps serve
as inputs for the realization of improved health and education, the outputs from economic
growth and development.
Chon (2006, 2007) provides the valuable connections among development, capabili-
ties, and intellectual property. Her works lay the foundation for a bottom up approach
to intellectual property that has as its starting point the activities and needs of people.
Development, according to her conception, arises from education (which is the focus of
her research) rather than from macro variables of trade and investment. However, access
to education depends on intellectual property laws adopted by a particular nation. As
Chon shows, copyright laws on access to education materials vary widely across countries
demonstrating different commitments to using intellectual property laws to improve
educational opportunities. Although she does not directly establish whether variations
in intellectual property laws promote stronger educational outcomes, she does show that
countries have managed to use copyright laws as strategic tools for promoting develop-
ment through access to education.
Detailed studies of institutions and behavior involving intellectual property can better
inform how to structure legal reform and policy. Lerner and Schankerman (2010) examine
the propagation of open source and proprietary software in developing countries. Scaria
(2014) provides an excellent example of this line of research. Using survey techniques,
Scaria identifies the extent of and reasons for copyright piracy of film in India. His work
examines both the demand side of piracy among Indian households and the supply side
responses by film copyright owners. The empirical findings are framed within a theoreti-
cal model of crime that identifies both the negative and positive incentives for engaging
in illegal activity like piracy. With this methodology, Scaria is able to identify how piracy
correlates with income, education status, and the price for various works. In addition, he
measures the use of peer to peer and other file sharing networks for purposes of piracy
as well as legally recognized uses of copying. While there are limits to survey methods,
even ones as broad and comprehensive as Scaria’s, this study is unique in providing rich
institutional details about copyright and usage in a developing country.
Cimoli et al. (2014) collect a diverse set of case studies on the role of intellectual
property in development. The studies in this collection complement the theoretical
approach of Banerjee and Duflo to illuminating the institutional and behavioral dimen-
sions of economic development. For example, the case study on the Green Revolution
demonstrates how technology transfer can occur parallel to intellectual property regimes
through government-industry collaboration. The case studies focus on the role of technol-
ogy diffusion in development and the limits placed on diffusion by intellectual property
laws at critical linkages from law to institutions to behavior by firms and consumers in
the adoption of new technologies.
Micro-level studies of economic development can potentially suffer from being too dif-
fuse and ungeneralizable. The methodology of Banerjee and Duflo addresses this concern
with an emphasis on experimental design combined with attention to institutional and
behavioral details. Future research can complement their approach with more careful and
varied case studies about specific industries in specific countries at specific points in time.
The resulting findings can be valuable in identifying contextual variables for the success
and failure of economic development within differing regimes of intellectual property.

V. DIRECTIONS FOR FUTURE RESEARCH
Discussions of intellectual property are often fueled by rhetoric about promoting pro-
gress, incentivizing creativity, and sparking innovation. There should be no surprise that
similar rhetoric informs the relationship between intellectual property law and economic
development. This chapter has tried to temper the rhetoric with a synthesis of scholarly
work that examines the determinants of economic growth and development, the relation-
ship between law and the economy, and the role of intellectual property law in fostering
development. Synthesis is necessarily a broadly interdisciplinary one, reaching across law,
economics, social theory, cultural theory, and political theory. While such a multifaceted
discussion may seem diffuse, the synthesis attempts to identify common themes that are
relevant to policy and scholarly research. Synthesis can promote greater analysis as a cure
to the rhetoric.
The volume divides the topic of development into economic development and cultural
development, with economic development centering on state efforts to promote improve-
ments in material well-being and cultural development, on state efforts to promote aspects
of national culture. As one recognizes that context matters for the relationship between
law and economic development, culture—whether language, history, religion—is an
important context for economic development. Therefore, the division may be artificial
and is largely adopted for purposes of exposition. Future scholars can fruitfully draw the
connections between these two interrelated strands.
Future scholars can also provide more specific context for understanding how intel-
lectual property and development interact in their complex, nonlinear relationship.
The topics identified in this chapter are rich with ideas for engagement. Whether better
understanding the intricacies of trade and agreements, or identifying experiments for
understanding technology adoption, finance, and business entrepreneurship at the micro
level, scholarship has much to build on here. In this way, rhetoric can give way to light
and to more thoughtful and effective policy debate at the national, regional, and global
levels.
REFERENCES
Aoki, Masahiko, Timur Kuran, and Gérard Roland, eds. 2012. Institutions and Comparative Economic
Development. New York, New York: Palgrave Macmillan.
Banerjee, Abhijit V., and Esther Duflo. 2011. Poor Economics: Rethinking Poverty and Ways to End It. New
York, NY: Random House.
Barro, R.J., and Xavier Sala-i-Martin. 2003. Economic Growth. Cambridge, MA: The MIT Press.
Basberg, Bjørn L. 1987. ‘Patents and the Measurement of Technological Change: A Survey of the Literature,’
16 Research Policy 131–41.
Casper, Steven. 2007. Creating Silicon Valley in Europe: Public Policy Towards New Technology Industries.
Chon, Margaret. 2006. ‘Intellectual Property and the Development Divide,’ 27 Cardozo Law Review 2821–911.
Chon, Margaret. 2007. ‘Intellectual Property “From Below”: Copyright and Capability for Education,’ 40
University of California Davis Law Review 803–47.
Cimoli, Mario, Giovanni Dosi, Keith E. Maskus, Ruth L. Okediji, Jerome H. Recihman, and Joseph E.
Stiglitz. 2014. Intellectual Property Rights: Legal and Economic Challenges for Development. Oxford: Oxford
University Press.
Coase, Ronald H. 1937. ‘The Nature of the Firm,’ 4 Economica 386–405.
Coase, Ronald H. 1960. ‘The Problem of Social Cost,’ 3 Journal of Law and Economics 1–44.

Commission on Intellectual Property Rights. 2002. Integrating Intellectual Property Rights and Development
Policy. London: Commission on Intellectual Property Rights.
Cooter, Robert D., and Hans-Bernd Schäfer. 2012. Solomon’s Knot: How Law Can End the Poverty of Nations.
Princeton, NJ: Princeton University Press.
Deaton, Angus. 2013. The Great Escape: Health, Wealth, and the Origins of Inequality. Princeton, NJ: Princeton
University Press.
Farley, Christine Haight. 2014. ‘TRIPS-Plus Trade and Investment Agreements: Why More May Be Less for
Economic Development,’ 35 University of Pennsylvania Journal of International Law 1061–72.
Fields, Gary. 2001. Distribution and Development: A New Look at the Developing World. Cambridge, MA.: The
MIT Press.
Ghosh, Shubha. 2012. ‘Propriedade intelectual no novo projeto de direito e desenvolviemento comentarios e
exemplos da India,’ in Mario G. Schapiro and David M. Trubek, eds., Direito e Desenvolvimento: Um Diálogo
Entre Os Brics. Sao Paulo, Brazil: Editora Saraiva.
Gilson, Ronald J., Charles F. Sabel, and Robert E. Scott. 2013. ‘Contract and Innovation: The Limited Role of
Generalist Courts in the Evolution of Novel Contractual Forms,’ 88 New York University Law Review 170–215.
Griliches, Zvi. 1990. ‘Patent Statistics as Economic Indicators: A Survey,’ 28 Journal of Economic Literature
1661–707.
Heald, Paul M. 2003. ‘Misreading a Canonical Work: An Analysis of Mansfield’s 1994 Study,’ 10 Journal of
Helpman, Elhanan, and Paul Krugman. 1985. Market Structure and Foreign Trade: Increasing Returns,
Imperfect Competition, and the International Economy. Cambridge: The MIT Press.
Hovenkamp, Herbert. 2016. ‘The Emergence of Classical American Patent Law,’ 58 Arizona Law Review 263–306.
Hurst, Willard. 1956. Law and Conditions of Freedom in the Nineteenth Century United States. Madison, WI:
University of Wisconsin Press.
Lerner, Josh, and Mark Schankerman. 2010. The Commingled Code: Open Source and Economic Development.
McCauley, Stewart. 1963. ‘Non-Contractual Relations in Business: A Preliminary Study,’ 28 American
Sociological Review 55–67.
MacNeil, Ian. 1978. ‘Contracts: Adjustments of Long-Term Economic Relations under Classical, Neoclassical,
and Relational Contract Law,’ 72 Northwestern University Law Review 854–902.
MacNeil, Ian. 1980. The New Social Contract. New Haven, Conn.: Yale University Press.
Mandel, Gregory N. 2014. ‘Leveraging the International Economy of Intellectual Property,’ 75 Ohio State Law
Journal 733–78.
Mansfield, Edwin. 1994. ‘Intellectual Property Protection, Foreign Direct Investment, and Technology
Transfer.’ International Finance Corporation Discussion Paper No. 19.
Maskus, Keith E., and Carsten Fink, eds. 2005. Intellectual Property and Development: Lessons from Recent
Research. Washington, DC: The World Bank.
Maskus, Keith E., and Jerome H. Reichman, eds. 2005. International Public Goods and Transfer of Technology
Under a Globalized Intellectual Property Regime. Cambridge: Cambridge University Press.
Millaleo, Salvador, and Hugo Cadenas. 2015. ‘Intellectual Property in Chile: Problems and Conflicts in a
Developing Society,’ in Matthew David and Debora Halbert, eds., The Sage Handbook of Intellectual
Property. Los Angeles, CA: Sage.
Mokyr, Joel. 2017. A Culture of Growth: The Origins of the Modern Economy. Princeton, NJ: Princeton
University Press.
Neal, Larry, and Jeffrey G. Williamson. 2014. The Cambridge History of Capitalism Volume I: The Rise of
Capitalism—From Ancient Origins to 1848. Cambridge: Cambridge University Press.
Neal, Larry, and Jeffrey G. Williamson. 2014. The Cambridge History of Capitalism Volume II: The Spread of
Capitalism—From 1848 to the Present. Cambridge: Cambridge University Press.
Perullo, Alex, and Andrew J. Eisenberg. 2015. ‘Musical Property Rights Regimes in Tanzania and Kenya after
TRIPS,’ in Matthew David and Debora Halbert, eds., The Sage Handbook of Intellectual Property. Los
Angeles, CA: Sage.
Robinson, Joan. 1979. Aspects of Development and Underdevelopment. Cambridge: Cambridge University Press.
Romer, P.M. 1994. ‘The Origins of Endogenous Growth,’ 8 The Journal of Economic Perspectives 3–22.
Sabel, Charles F. 2005. ‘Bootstrapping Development: Rethinking the Role of Public Intervention in Promoting
Growth,’ Paper presented at the Protestant Ethic and Spirit of Capitalism Conference, Cornell University,
Ithaca, New York, October 8–10, 2004, accessed August 2, 2015, at http://www2.law.columbia.edu/sabel/papers/
bootstrapping%20deve%20send5.pdf.
Sachs, Jeffrey D., and Andrew M. Warner. 1997. ‘Fundamental Sources of Long-Term Growth,’ 87 The
Scaria, Arul George. 2014. Piracy in the Indian Film Industry: Copyright and Cultural Consonance. Cambridge:

Schmookler, Jacob. 1972. Patents, Invention, and Economic Change: Data and Selected Essays. Cambridge, MA:
Schultz, Theodore W. 1981. Investing in People: The Economics of Population Quality. Berkeley, CA: University
of California Press.
Solow, Robert M. 1956. ‘A Contribution to the Theory of Economic Growth,’ 70 Quarterly Journal of Economics
65–94.
Solow, Robert M. 1988. Growth Theory: An Exposition. Oxford: Oxford University Press.
Spiegler, Peter. 2015. Behind the Model: A Constructive Critique of Economic Modeling. Cambridge: Cambridge
University Press.
Stiglitz, Joseph E., and Andrew Charlton. 2005. Fair Trade for All: How Trade Can Promote Development.
Trebilcock, Michael J., and Mariana Moto Prado. 2014. Advanced Introduction to Law and Development.
Trubek, David M., and Alvaro Santos. 2006. The New Law and Economic Development: A Critical Appraisal.
Williamson, Oliver. 2013. The Transaction Cost Economics Project: The Theory and Practice of the Governance
of Contractual Relations. Cheltenham, UK and Northampton, MA, USA: Edward Elgar Publishing.
World Bank. 2004. Poor People’s Knowledge: Promoting Intellectual Property in Developing Countries.
Yamane, Hiroko. 2015. ‘Competition Analyses of Licensing Agreements: Considerations for Developing
Countries under TRIPS.’ Geneva: ICTSD Discussion Paper.
Yu, Peter. 2015. ‘Déjà Vu in the International Intellectual Property Regime,’ in Matthew David and Debora
Halbert, eds., The Sage Handbook of Intellectual Property. Los Angeles, CA: Sage.

24. Economic development and intellectual property
rights: key analytical results from economics
Keith E. Maskus*
Contents
I. Introduction
II. Endogenous Intellectual Property Regimes
A. Basic Tradeoffs
B. Global Tradeoffs
III. Intellectual Property, Innovation, and Technology Diffusion
A. IPRs and Innovation
B. IPRs and Technology Diffusion
IV. Patents and Access to Medicines
V. Concluding Remarks
References
I. INTRODUCTION
Economic development is the result of numerous interrelated processes, including the

adoption and development of new technologies, the organization of markets to facilitate
production and consumption, and the establishment of facilitating institutions (Cimoli
et al., 2014). For countries behind the technological frontier, development is a process of
‘catching up,’ featuring a combination of imitation, learning, knowledge absorption, and
ultimately innovation. Many and varied elements affect the pace and direction of this
evolution, including the governance of intellectual property rights (IPRs).
Intellectual property laws, and their implementation and enforcement, aim to strike
a balance among numerous objectives and constraints that affect the complex processes
undergirding economic and social development. These laws determine the scope and
strength of IPRs, which in turn may influence the pace and direction of economic activity.
In their most direct conception, patents are thought to encourage innovation and deepen
technological markets, while raising the costs of potential rivals in imitating new tech-
nologies. Trademarks are supposed to sort out information problems in markets where
consumers might be confused about the provenance of goods and services, even as they
potentially diminish employment in counterfeiting firms. Copyrights are presumed to
secure market returns to successful content creators, but may diminish access to cultural
and scientific knowledge.
* Arts and Sciences Professor of Distinction in the Economics Department at the University
of Colorado, Boulder. Email: keith.maskus@colorado.edu.
656

Key analytical results from economics 657
These basic views dominate received economic analysis of IPRs to a considerable

degree, as will be evident from the review in this chapter. By focusing on specific incentive
impacts of patents or copyrights, economists, at least in theoretical models, are able to
isolate their potential effects on development and growth and to identify how those effects
may vary with economic and technological conditions.1 This approach has unearthed a
range of important insights that should be accounted for in any assessment of intellectual
property law and regulation.
At the same time, however, IPRs reside in a far broader and more complex economic
and social ecosystem than economists can hope to capture with tractable theoretical
models or econometric analysis. Thus, for example, development economists point to the
roles patents may play in the full ‘national innovation system,’ involving infrastructure,
investment taxes, R&D subsidies, factor markets, competition rules, educational attain-
ment, and even trade policy, not to mention accidents of history and geography (Odagiri
et al., 2010). How such factors interact over time and at different levels of economic
development is a frightfully complex issue, making it difficult to say much with confidence
about the true significance of IPRs in the development context.
Beyond this general complexity lie the many details of policies and effects that, while
extremely important, render straightforward statements about IPRs and economic
development all but meaningless.2 To begin, even the best economists make the mistake
of failing to distinguish among the types of IPRs, which may have quite different impacts
on development prospects. Patents and plant variety rights operate differently from
copyrights, which in turn are distinct from trademarks and geographical indications.
Moreover, these devices feature complex regulatory components that influence their true
protective scope, such as compulsory licenses of patents, copyright limitations and excep-
tions, and parallel trade. Next, economic sectors vary widely in their interrelationships
with IPRs of various types, with pharmaceuticals, chemicals, and biotechnological inven-
tions most dependent on patents and literature, music, software, and digital entertainment
goods closely associated with copyrights. Finally, IPRs are national policy constructs and
as such their structure and scope are, to some extent, dependent on economic and social
conditions in each country. In turn, there is two-way causation between economic activity
and the IPRs regime. Such complications should be kept in mind when contemplating the
development aspects of intellectual property protection.
Perhaps the greatest challenge in organizing a review chapter is simply deciding which
elements to cover among the vast array of development problems that may be affected
by IPRs. There are many economic processes that have been related to patents, including
innovation, technology diffusion, trade, competition, monopoly power, and price-setting.
Analysts may be concerned about sectoral issues involving medicines, green technologies,
agricultural inputs, software, and e-commerce. They have broader concerns, such as how
1
As will become evident later, satisfactory empirical identification of such effects is far more
elusive.
2
Regrettably, such statements are common where observers have a strong economic or
political interest, whether for or against IPRs. Thus, for example, some see IPRs as an unalloyed
‘power tool for development’ (Idris, 2002). Others paint IPRs as largely a mechanism for blocking
development of poor countries through sustaining monopolistic rents of existing firms (Stiglitz
and Charlton, 2005).

the productivity of IPRs depends on other policies, their implications for access to scien-
tific and technological information, and barriers they may raise to cultural development.
A judicious treatment is therefore called for to avoid being too shallow in the treatment
of too many subjects. Here I focus on three key questions that have been studied closely
by economists and about which most may be said with some confidence. First, I overview
the political economy of decisions made by countries at different development levels to
adopt stronger IPRs and characterize international trade rules in that context. Second,
I critically discuss findings on the roles of IPRs in innovation, technical change, and
technology transfer across borders. Third, I describe major results about how patents
may be affecting pricing decisions and product availability of medicines in developing
economies.
II. ENDOGENOUS INTELLECTUAL PROPERTY REGIMES
The vast majority of economic theorizing in this area treats IPRs as exogenous, or
determined independently outside the economic and social environment. This approach
is useful for analysing how policy changes might influence the behavior of specific firms
and industries or alter the well-being of households. It is reasonable to suppose that any
individual firm or consumer simply takes such policy changes as given by legislators or
trade negotiators seeking to achieve a broad set of outcomes.
A. Basic Tradeoffs
The difficulty, of course, is that IPRs, like taxes and tariffs, are the endogenous outcomes
of a policymaking process that depends on a large variety of competing interests, both
domestic and foreign. Moreover, these interests change over time because the dynamics
of economic growth fundamentally alter underlying circumstances. Focusing for the
moment solely on domestic factors, a government devoted to maximizing national welfare
would take account of at least the following considerations.3 First is consumer welfare,
which consists of static consumer surplus from having access to existing goods and
dynamic benefits from having access to more and newer goods from future innovation.
Second is the profits (more accurately, producer rents) of domestic firms, made up of
both imitators and innovators or creative content providers. Third would be any spillover
benefits or costs of IPRs, such as reduced infection rates from faster access to newer
medicines or diminished learning of new technologies through higher-cost imitation or
reverse engineering.
From this description two basic and interrelated tradeoffs emerge in policymaking
(Maskus, 2000). One is the purely static distribution of welfare between consumers,
3
The government might also care about tax revenues, employment, and other objectives.
Alternatively, a government may be self-interested, with legislators seeking to maximize the chances
of staying in office or garnering lobbying contributions, calling for a model of political economy
along the lines of a tariff-setting model. See, for example, Grossman and Helpman (1994). The
self-interested approach has not yet been studied in a rigorous empirical model of patent policy
formation.

including input users, and producers of existing goods and technologies. Another is the
dynamic tension between current access gains and the benefits from future innovation,
incentivized by IPRs. This simple logic underlies the primary political-economy view
about economic development and intellectual property. Countries with limited innova-
tion capacity are likely to favor short-term access benefits through weak or absent IPRs.
As firms and industries acquire greater capacities to invent new goods, which could
be the result of numerous economic and policy factors, interests arise endogenously
in strengthening patents and other rights (Chen and Puttitanun, 2005; Ginarte and
Park, 1997; Park, 2008a). A variant of this story is that endogenous protection may be
U-shaped in economic development, which was demonstrated in an early study (Maskus
and Penubarti, 1995).4 Countries at the lowest income levels may favor the ability of
moderately strong IPRs to bring more products to their markets from abroad, while
having little domestic production capacity that would oppose such a regime. As nations
gain more imitative capacity at middle-income levels, however, local firms gain the ability
to imitate products, which is facilitated by weaker patents. Beyond some level of real
per-capita incomes the emerging innovation interests dominate and the strength of IPRs
rises along with development.
It is consistent with economic history that IPRs are expanded in scope as economies
grow richer and more technologically capable (Odagiri et al., 2010). For example, develop-
ment of the US pharmaceutical industry was boosted by the vacating of German-owned
patents in World War I, with the rights to produce such goods given to domestic firms
as compulsory licenses (Moser and Voena, 2012). Switzerland’s chemical industry grew
from imitating foreign formulations in the absence of domestic patents (Maskus, 2012).
These countries now have highly protective patent regimes. Through the late 1980s Japan’s
patent system favored widespread technology diffusion through the use of utility models
and narrow claims, but that country also now strongly protects novel inventions (Ordover,
1991; Maskus and McDaniel, 1999). South Korea has experienced a similar transition
of technology development and IPRs, while the rapid solidification of patent protection
in China since 2000 surely reflects the emergence of such high-technology industries as
electronics, solar power, and biotechnology.
Two important qualifications to this dynamic must be mentioned. First, because
countries vary in their industrial composition even relatively poor nations may have
endogenous preferences for strong components of IPRs, while richer countries may
limit them. India, for example, has long protected copyrights due to the importance
of its domestic film and publishing sectors. More recently, many developing countries
have entertained legislation to protect geographical indications as a potential boost to
their agricultural sectors (Maskus, 2012). In contrast, Canada has deployed compulsory
licenses in pharmaceuticals and limited the scope of digital copyrights in order to
favor consumer access. These stories suggest that attempts to identify the evolution of
‘appropriate’ intellectual property regimes as countries develop inevitably are subject to
numerous exceptions (World Bank, 2002; Kim et al., 2012).
4
Caution should be exercised in interpreting this U-shaped outcome, however, for to some
degree, it may reflect simply the legacy of imported colonial laws.

B. Global Tradeoffs
A second and more fundamental qualification is that countries may not be fully free to
select their desired IPR regimes, even where governments are welfare maximizers, in a
world of open trade and investment. The primary reason is that a developing country’s
domestic protection may be inadequate for the interests of international corporations
seeking to export or invest there.5 These companies can press directly for upgraded
standards or encourage indirect strengthening via international trade agreements. Thus,
for example, the unprecedented expansion and partial convergence of patent rights since
1995 is largely the result of the Agreement on Trade-Related Aspects of Intellectual
Property Rights (TRIPS) at the World Trade Organization and even stronger demands
in subsequent preferential trade agreements (Maskus, 2012). It follows that negotiated
intellectual property rules, in themselves, likely exceed the individually welfare-optimizing
levels within poorer countries.
These considerations were spelled out carefully in landmark theoretical contributions
(Grossman and Lai, 2004; Scotchmer, 2004). In determining their own policies (so-called
‘Nash non-cooperative patent regimes’), governments will take into account the welfare
gains from access to goods and the profits of domestic firms, including imitators. They
will not consider the profits earned in their markets by foreign firms. There are two
critical implications. First, countries with large markets and strong capacities to invent
goods with commercial potential will choose considerably stronger protection than will
those with small markets and limited inventiveness. Second, these individual policies,
by failing to offer incentives to foreign innovators, suffer from a global coordination
problem: patent and copyright systems are inadequate to produce the globally optimal
level of innovation and growth is diminished as a result. It follows that international
agreements to internalize this spillover through more integrated standards can expand
global welfare.
This logic offers the key principled justification for a tendency to harmonize inter-
national IPRs within TRIPS or increase them further via preferential trade agreements
(PTAs). In the latter approach, agreement partners are pressured by industrialized coun-
tries, primarily the United States and the European Union, to adopt certain ‘TRIPS-Plus’
standards, which go beyond the requirements of TRIPS. Typically included among these
demands are stronger patent rights and data-secrecy protection in pharmaceuticals and
chemicals, enhanced protection from online digital piracy, elevated rights for geographical
indications, and strong enforcement mechanisms (Maskus, 2012). Indeed, policymaking
in TRIPS and its aftermath have been affected, for measured patent rights in the era
after the agreement became higher than their prior Nash levels.6 Note, however, that the
TRIPS and TRIPS-Plus patent standards themselves may not be globally optimal, despite
their partial convergence, if they settle at weighted-average levels higher than needed to
5
Again, a politically driven system could generate excessive protection even on a national
basis if industries seeking strong exclusive rights have more lobbying influence, a situation that
arguably has characterized the United States in recent years and underlies domestic debates about
copyright limitations and patent scope.
6
The issue of measurement will be discussed below. On the departure of TRIPS rules from
Nash standards, see the estimates in Lai et al. (2008).

correct the international externalities. Moreover, because these agreements generated

rules that may exceed currently useful levels for developing countries they may diminish
welfare there unless their costs are offset by other benefits. These questions remain largely
unanswered in the economics literature.7
III. INTELLECTUAL PROPERTY, INNOVATION, AND

TECHNOLOGY DIFFUSION
In granting and protecting the exclusive rights embodied in various forms of IPRs,
governments attempt to manage these static, dynamic, and international tradeoffs. The
fundamental social bargain in patents, for example, is to offer temporary rights to exclude
others from using or copying protected technologies and goods in return for disclosure
of the patented information. The ability to exclude addresses the dynamic innovation
problem by creating a limited monopoly in ideas, permitting originator firms to earn
enough profits to pay for R&D costs or to finance the costs of marketing their technolo-
gies. Copyrights address the similar problem of promoting creative expression. Whether
IPRs actually achieve these goals, and under what circumstances, is a hugely complex
question, even among developed economies. In this section, I review primary evidence
and draw some important lessons.8
A. IPRs and Innovation
As suggested above, a primary justification for IPRs is their presumed ability to incentiv-
ize invention, innovation, creativity, and knowledge diffusion. These concepts are not
easily measured or even explained, especially in poor countries. For example, economists
know very little about the drivers of artistic creativity and inventive activity in informal
sectors in developing countries. Even within developed economies, the primary measures
of innovation used by economists are indirect and focused on invention and diffusion,
whether input-based (R&D expenditures) or output-based (patents and technology
spillovers). These variables, while imperfect, permit a historical and statistical record
linking patents or patent laws to innovation.9
7
No particular answers could be definitive given the complexity of the subject and attendant
empirical uncertainties. For example, many have argued that the World Trade Organization (WTO)
agreements failed to generate sufficient market access and technology transfer to developing
countries to offset the costs of stronger IPRs (Maskus, 2012). However, if one takes account of
the likely positive impacts of globalized patent rights on foreign direct investment, even high-level
harmonization generates global increases in technology diffusion (Lai and Yan, 2013).
8
A full review in this chapter is impossible, given space constraints. See Maskus (2012), Park
(2008b), and Cimoli et al. (2014). Readers may wish to begin with a review of the broader roles of
innovation in economic development, such as Fagerberg et al. (2010).
9
Copyrights are difficult to assess in this context because they need not be registered to have
legal weight. Analysts sometimes measure their importance by the shares of employment or output
in such ‘copyright industries’ as music, publishing, software, and digital entertainment (Pouris and
Inglesi-Lotz, 2011; European Commission, 2013). This approach says relatively little about the
causal effects of copyrights, however.

As an initial matter, note that patents are, in principle, neither necessary nor sufficient
to induce the optimal degree of innovation. As long argued by innovation scholars, the
primary motivation for investments in new goods is the anticipation of making future
profits. This observation has been the foundation of innovation theory by economists
from Schumpeter (1942) to Romer (1990) and beyond. The ability to appropriate profits
from these inventions may come from many factors other than IPRs, including secrecy,
market lead times, the high costs of imitation, and barriers to competition, though the
relative importance of such factors, including patents, varies sharply across industries
(Cohen et al., 2000). This explains why the historical anecdotal record may be used to
support any view of the necessity of patents. For example, James Watt’s fundamental
improvement of the steam engine was patented and defended rigorously. The role of
patents in this case is alternately described in the literature as critical for the invention
itself and unnecessary, in that the patent was procured after the invention was stabilized,
and a socially wasteful means of blocking entry ex post by others.10
Anecdotes prove little so we must consider evidence from econometric studies. In this
context, innovation historians have fruitfully analysed detailed data from the past. For
example, Chen (2008) studied how 614 major inventions and innovations from 1750–1950
were related to the existence of patent laws in 14 Western European countries and the
United States. These countries introduced their initial patent laws at different times in
this period. Chen found a positive and highly significant impact of the existence of a
patent law on the number of domestic inventions over this long period, often with a
long lag, suggesting that a legal patent regime ultimately supports domestic inventive
activity. However, this evidence must be treated cautiously for a number of reasons. Most
importantly, he did not include in the analysis many confounding variables that surely
influence national inventiveness, nor did he satisfactorily account for reverse causality
between inventions and adoption of patent laws.
In a landmark study, Lerner (2002) compiled information on 177 legal patent reforms
in 51 countries over the period 1852–1998. He focused on substantive reforms, such as
implementation of patent laws and extensions of patent duration. He estimated how
these policy reforms led to changes in home patent applications by domestic residents and
foreign applications in the reforming nations, using data normalized by contemporaneous
trends in the propensity to patent. The study window ranged from five years before to five
years after each reform and the regressions included a number of controls. Lerner’s results
were striking. The volume of both domestic and foreign applications rose after patents
were strengthened. However, after normalization only foreign applications increased
significantly, while domestic applications actually fell. Thus, in most cases national patent
reforms induced far more inward applications than domestic inventions, at least in the
short run. Applications by domestic firms were, in fact, crowded out, perhaps by increased
competition from abroad. Additional work found evidence of diminishing returns to
increasing patent protection over time, meaning that countries with weaker initial regimes
saw more patenting post-reforms, and that domestic innovation gains were concentrated
in larger and higher-income economies.
10
The pro-patent history is exemplified by Rosen (2010), while the anti-patent argument is
made by Boldrin and Levine (2008).

Lerner’s findings are sobering for policymakers expecting that new and stronger patent
laws in the developing world will induce considerably more domestic innovation. This
outcome may pertain in large countries with initially weaker patents and rapidly rising
incomes, such as China and Brazil. In much of the developing world, however, the impact
over several years is likely to be greater growth in incoming patent registrations by foreign
firms seeking to protect the new products and technologies they export or transfer. Put
into context, this situation is not surprising. Many of the reforms Lerner analysed were
adopted in response to foreign pressure, rather than domestic commercial interests.
In another important study, Moser (2005) looked at whether patent laws alter the
sectoral distribution of inventive activity, rather than overall innovation.11 She compiled
data on nearly 15,000 inventions exhibited at either the 1851 world’s fair in London or
the 1876 world’s fair in Philadelphia. Her hypothesis was that inventions in countries
without patent laws should be concentrated in industries with other means of appropria-
tion, such as secrecy and lead time, while those from nations with patent laws should be
more broadly distributed. This proposition was confirmed: inventions from countries
without patents were more likely to be in textiles, food processing, and instruments (with
low patent intensities), while those from countries with patent laws were more evenly
distributed, though significantly higher in machinery. Notably, after the Netherlands
abolished its patent law in 1869, the share of Dutch innovations in food processing rose
sharply. Moreover, the cross-industry distribution of inventions was significantly different
between countries with short patent durations versus long patent durations. Thus, it seems
that both the existence and strength of patents can profoundly affect the type of goods
invented. In turn, legal patent reforms could help determine the evolution of industrial
specialization over time.
This historical record is informative but we also wish to study how IPRs influence
innovation currently, particularly across countries at different levels of economic develop-
ment. Econometric analysis of this question is rather recent, mainly because there were
no consistent and international measures of patent protection over time until the famous
Ginarte-Park (GP) index appeared (Ginarte and Park, 1997). This index, now extended
to 2010 across a comprehensive set of countries, amounts to an adding up of the presence
or absence of particular legal provisions in five components of patent laws, generating
an index running from zero to five. Its use has been criticized for a number of reasons,
including its inability to measure the extent of legal enforcement. This index has increased
sharply among developing and emerging economies since 1995, with a considerable
degree of harmonization with the most protective countries (Maskus, 2012). The earliest
studies of innovation using such measures suffered considerably from an inability to
control for missing variables and endogeneity. However, enough well executed studies
have been published recently to support certain conclusions, which primarily suggest that
the evidence is far from clear.
An important question is whether the effects of patent reforms on innovation activities
are different between rich and poor nations. In this context, Schneider (2005) analysed a
sample of 19 developed and 28 developing countries, taking innovation as the number of
patent applications residents of each nation registered in the United States from 1970 to
11
For further information on the historical evidence, see Moser (2013).

1990. The explanatory variables included the GP index and several national variables that
should affect technological change. In the basic regressions, Schneider found a positive
and significant elasticity of 0.6 between patent applications and patent rights. However,
when the sample was divided this positive impact remained only for developed countries,
with a highly elastic coefficient of around 2.0. In developing countries, the effect was nega-
tive, though significant in only some specifications. We should interpret these findings
cautiously, given some econometric problems with the approach, including the failure to
deal with causality. However, they suggest that patent laws have limited or even negative
impacts on contemporaneous patentable innovation in developing countries. It would be
useful to revisit this question to analyse whether there is a longer-lagged effect that may
be positive in emerging economies.
A more sophisticated approach was used by Chen and Puttitanun (2005). They used
data for 61 developing countries over 1975–2000 and accounted for the simultaneity
between IPRs and innovation. The authors found that the GP index had no effect on US
patent applications by residents of lower-income countries but the impact was positive
and significant for middle-income and emerging economies. Thus, there is an important
threshold effect, in that increases in the scope of patent rights seem to induce more
innovation only above relatively high levels of gross domestic product (GDP) per capita.
Similar conclusions were drawn by Allred and Park (2007a).
While suggestive, the results of aggregate cross-country regressions are of questionable
reliability for several reasons. More recently scholars have incorporated firm-level data
sets, which increase sample sizes and permit greater focus on strategic aspects of the
IPR-innovation relationship. For example, Allred and Park (2007b) related firm-level
real R&D expenditures for 2,446 multinational enterprises in ten industries to their
headquarter-nation’s GP index in 1990, 1995, and 2000, controlling for firm size, GDP,
and time and industry fixed effects. They found a strongly positive impact of patent rights
on R&D in the developed countries but no effect in developing countries. This evidence
indicates that elevated patent laws stimulate R&D in nations where there are both high
incomes and significant technological capabilities. However, there is little evidence of such
impacts in low-income countries.
A highly notable study is by Branstetter et al. (2006), which analysed the responses of
affiliates of US multinational enterprises to major reforms in patent laws in 16 countries,
14 of which were developing or emerging, between 1982 and 1999. The authors performed
an event analysis, considering changes in aggregate resident and non-resident patent fil-
ings in a six-year window surrounding the dates of reforms. In their econometric model
the patent reforms had no impact on domestic applications. However, the results indicated
that reforms had a positive impact on foreign patent applications, both in the short and
long run, raising non-resident filings in the average nation by at least 52 percent. Thus,
these authors reinforced the basic wisdom that international firms are more responsive to
increases in patent rights in developing countries than are domestic firms.
A last important work is by Qian (2007), who analysed 26 countries that, between 1978
and 2002, implemented laws establishing patent protection for pharmaceutical products
and how those policies affected innovation in that industry. Her primary innovation meas-
ure was the log of citation-weighted drug-patent applications registered in the United
States after legal changes, comparing matched country pairs that differed in whether they
adopted reforms. Various national and industry control variables were included in the

regressions. Qian found that there were no significant direct impacts of legal changes on
US drug-patent applications, even up to ten years later. However, there were important
interactions, in that patent reforms in countries with higher educational attainment,
per-capita income, and greater measured market freedom significantly increased such
applications. Thus, Qian’s results offer more evidence that the innovation-inducing impact
of IPRs depends on other factors. Low-income economies with limited educational
attainment and technical skills are unlikely to see much impact.
This review points out that there are no clear and unidirectional relationships between
patent rights and subsequent innovation, however measured.12 Much depends on other
factors that vary across countries and industries and over time. A few general conclusions
are worth drawing, however. First, the initial effect of legal revisions in developing coun-
tries is to attract more applications from abroad as multinational firms seek to exploit and
protect their technologies. Second, even in middle-income countries it takes time for any
domestic responsiveness to emerge.13 IPRs reforms have little, if any, impacts on innova-
tion in poor countries, perhaps because of weak business and investment environments
and poor governance institutions, including an inability to enforce such laws.
An important qualification is that virtually all of the available evidence refers to
strengthening patent laws, which may simply be irrelevant for innovation in the poorest
countries. But innovation and creativity are hardly absent in those countries, even if they
reside largely in informal sectors. It would be of great interest to study closely whether, in
the post-TRIPS era, new copyright systems have encouraged creative activity in concert
with greater access to the internet, or whether small producers are registering more
domestic and international trademarks as they expand their marketing reach.
B. IPRs and Technology Diffusion
For most developing economies, incoming international technologies are the primary
source of new information, productivity gains, and economic growth (Keller, 2004).
International technology diffusion is therefore a major determinant of global technical
change and increasing such flows is a critical part of economic development policy.
It is equally vital to adapt technologies to local conditions and learn how to use and
improve them. Countries seeking access to foreign technologies therefore build into their
innovation systems the entire policy complex involving skill accumulation, investment,
competition, R&D support, and IPRs.
In order to summarize a complex set of relationships, it is useful to distinguish between
market-mediated technology transfer and informal means of diffusion into the broader
12
Available measures are themselves deeply flawed. For example, definitions and coverage
of R&D can vary considerably across countries. Patent applications may not reflect underlying
innovation so much as a need for firms to engage in defensive patenting in industries with overlap-
ping claims and cumulative innovation. Moreover, some countries, notably China, see patent
applications in themselves as socially desirable and governments may absorb the application costs,
resulting in excessively high filings.
13
There are exceptions, as technology-oriented firms in South Korea, China, and India
quickly expanded their R&D spending and turned to patenting in the wake of domestic reforms
(Maskus, 2012).

economy. One major market-based channel is trade in high-technology goods and ser-
vices. Imported capital goods and technological inputs can directly improve productivity
by being placed into production processes. A second is foreign direct investment (FDI)
through multinational enterprises (MNEs), which tend to transfer to their subsidiaries
newer and more productive technological information (Markusen, 2002). Yet a third is
technology licensing, which typically involves the transfer of production or distribution
rights, protected by some combination of IPRs, and the associated technical information
and know-how. In this context patents, trade secrets, copyrights, and trademarks serve as
direct means of information transfer. Licensing of IPRs is overwhelmingly performed via
voluntary contracts. However, governments may on occasion issue a compulsory license.
Also important is the cross-border movement of engineers and technicians who transfer
knowledge. Evidence from patent citations suggests that there is substantial diffusion
through this channel (Hovhannisyan and Keller, 2015).
There are also important non-market means of technology diffusion. The first is imita-
tion, involving efforts to learn the technological or design secrets of an incoming technol-
ogy, whether by product inspection, reverse engineering, or another task. Imitators pay
no compensation to the technology owner, making this an attractive form of learning.
However, imitation can be costly and divert investment from local innovation, so its full
impacts on development are not straightforward. A related form of non-market diffusion
is for technical and managerial personnel to take technical information to a rival firm,
which can be particularly significant in industries and locations where cross-fertilization
of knowledge is important. Firms also access technology through reading patent applica-
tions, which, in principle, offer enough information that a skilled person should be able
to use them to invent competing products that do not infringe the original claims. Patents
therefore provide both a direct vehicle of technology transfer, through FDI and licensing,
and an indirect form through inspection and experimentation.
It is evident that one important factor determining how readily technologies may be
diffused through these various channels is the scope of multiple IPRs. On the one hand,
patents, trademarks, and enforceable contracts for licensed trade secrets can do much to
reduce the information costs and uncertainty of market-based technology transfer (Yang
and Maskus, 2001; Hoekman et al., 2005). On the other hand, if patents have extensive
scope, say through broad claims and a ban on experimental use, they can greatly increase
the costs of imitation. Similarly, rigorous trade-secrets protection against labor mobility
and patent applications that fail to disclose useful technical information do not support
much local diffusion. And, as always, the full effects depend on numerous conditioning
factors.
In this context, it is important again to consider the recent and most credible empirical
evidence. Regarding patents as a source of technical change, Eaton and Kortum (1996)
discovered in an elegant structural approach that the bulk of productivity growth in
smaller and less technologically advanced Organisation for Economic Co-operation
and Development (OECD) countries came from having foreign inventors patent in their
economies, resulting in related technology spillovers. This result likely would hold in small
developing nations, which overwhelmingly remain net importers of technology, so long as
they build the needed technical capacity to adopt and improve such technologies. Indeed,
there is direct evidence of this possibility from East Asian developing economies (Hu
and Jaffe, 2003). Citations listed in patents awarded to Korean and Taiwanese inventors

by the United States Patent and Trademark Office suggested that innovators in both
countries discerned and mastered considerable information from recent Japanese and US
inventions. Further, there are increasing citations across patents in East Asia, indicating
an expanding regionalization of knowledge flows (Hu, 2009).
The major channels of technology transfer include trade, investment, and licensing
contracts, raising the question of how these flows to developing economies are affected
by IPRs. This question supports an extensive literature that, while pointing to some
ambiguities, generally finds a positive relationship among emerging and middle-income
countries. For example, in the first study of the trade impacts of TRIPS reforms, Ivus
(2010) analysed the growth of high-technology exports from 24 OECD countries to
55 developing countries. Taking the 18 countries with relatively larger policy reforms
post-TRIPS as the treatment group, she found that high-technology exports to those
nations grew significantly faster than low-technology exports after 1994. Her estimates
suggested that the rise in the GP index in this period increased the value of OECD exports
of patent-sensitive goods to those countries by 8.6 percent. A more recent study finds
strong evidence that such reforms also raise the exports of high-technology goods from
middle-income economies (Maskus and Yang, 2018).
Regarding FDI, available evidence also points to positive impacts of patent rights in
developing countries. For example, increases in the GP index were a significantly positive
determinant of the FDI location decisions of US multinational firms between 1995 and
2000 (Nunnenkamp and Spatz, 2004). Similarly, the extent and enforcement of patents
in Eastern European and former Soviet Union economies positively affected the deci-
sions of European multinational firms to locate production facilities in those countries
(Smarzynska Javorcik, 2004). Similar evidence was found in Chinese provinces (Du et al.,
2008).
Intellectual property protection could affect multiple activities of multinational firms.
One prominent study analysed the impacts on licensing of US parents with affiliates after
patent-law changes in 16 developing economies (Branstetter et al., 2006). The authors
found that royalty payments to parents rose by 34 percent on average, mostly reflecting an
increased volume of technology sold rather than higher royalty charges. There was also a
significant increase in R&D investments at local subsidiaries. Both of these effects, which
were much stronger for companies in high-technology industries, implied a substantial
shift to reforming economies in the use and development of new technologies.
In a later study the authors related various measures of affiliate activity in high-
technology US multinational companies to patent reforms (Branstetter et al., 2011).
There were significantly positive increases after patent reforms in affiliate sales, net plant
and equipment, and employee compensation. Further, they found that value added in
local competing firms rose significantly, especially in technology-intensive sectors. The
preferred estimates indicated that there was an average increase of 20 percent in the local
industries of reforming economies. There was also strong evidence that firms in these
countries expanded the range (‘extensive margin’) of their exports to the United States
after patent rights were broadened. These results run counter to concerns that stronger
IPRs would shut down domestic enterprises. Rather, these policies seem to encourage
growth in the most competitive local firms. Again, these findings related to reforms in
larger and middle-income economies and there remains no evidence about whether they
would apply in smaller and poorer developing countries.

To summarize, the best available evidence supports the claim that patent reforms have
positive effects on inward technology transfer through market-based channels. They
attract foreign patents, though there is little evidence of a domestic innovation gain for
some years. They raise imports of high-technology goods and may also stimulate export
growth. Stronger IPRs expand the local activities of multinational firms, while increasing
licensing to both affiliated and unaffiliated parties. They particularly stimulate these
responses among high-technology firms.
While these are important benefits, the conclusions come with major qualifications.
First, to date, these impacts were found only in larger and middle-income countries. There
is little evidence of such effects in the poorest and smallest developing economies, where
patents are not of much relevance for technology transfer or industrial development.
Moreover, these positive impacts are subject to important threshold effects in the levels
of income and education. Second, the fact that international activities expand does not
necessarily imply a stimulus to domestic production. Local firms may have to change
product lines or close down if they cannot adapt to the new competitive environment
post-reforms, a possibility about which we have little systematic evidence.
A final major qualification is that the evidence reviewed above offers insights only on
issues where extensive data exist, which overwhelmingly means such market transactions
as exports, investment, and patenting. Adopting stronger IPRs to support such direct and
indirect technology markets is surely going to raise such activity, at least in countries that
can absorb new information. The subsequent spillovers into local productivity growth can
be substantial (Keller, 2010).
These potential gains reflect just one side of a complex process, however. Stronger
IPRs also may diminish prospects for imitation and learning from non-market channels,
removing a central channel for poorer countries to move up the critical lower rungs of the
technology ladder. Indeed, it is difficult to find historical evidence of a now-developed
economy that did not take considerable advantage of weak technology protection in the
early and middle stages of its development.14 Unfortunately, systematic data do not
exist for studying this fundamental claim because the counterfactual scenarios cannot
readily be measured without extensive industrial surveys applied consistently over time
in a selection of poor countries, combined with analysis of exogenous events affecting
imitation prospects. We are left far short of a balanced depiction of the full roles of IPRs
and economic development. This is the primary shortcoming in economic analysis of
intellectual property reforms, international technology flows, and innovation, one that
needs to be addressed.
In place of that lacuna, however, the literature does offer some important indirect
observations about IPRs and technology diffusion. Specifically, since the seminal work of
Zvi Griliches (1957), agricultural economists have studied how rapidly new crop varieties
are diffused and adopted across countries, based on numerous economic and technical
factors. Regarding intellectual property, two thought-provoking studies considered
the global diffusion of different major crops from 1960–2000 (Goeschl and Swanson,
14
Examples of those which did are many, including the United States, Switzerland, Japan,
South Korea, and China, albeit with different characteristics in each case. For a good historical
overview of such experiences, see the chapters in Odagiri et al. (2010).

2000; Swanson and Goeschl, 2014). An essential difference between crops is that corn
and maize hybrids had automatic ‘use restriction technologies’ because they produced
sterile seeds and could not be replanted, while others did not have this feature. Here, the
‘natural experiment’ was that such hybrids were fully protected by a technological form of
restriction on diffusion, while others may have been protected by weaker legal regimes of
plant breeders’ rights or patents, which varied across countries. They found that the strong
protection form produced higher levels of technological growth in those industries among
primarily developed economies, but materially impeded the diffusion of innovations to
developing countries. Thus, to the extent that corn hybridization can proxy for strongly
exclusive rights, this result suggests that enhanced IPRs may indeed slow the progress of
lower-income economies in approaching the technological frontier.
IV. PATENTS AND ACCESS TO MEDICINES
While studying the interplay between patents and innovation is important, it is hardly the
only relevant development issue regarding IPRs. More specific issues arise in considering
specific economic and cultural sectors. Full chapters could be devoted to the development
aspects of IPRs in agriculture, biogenetic resources, environmental technologies, health,
education, information technology, and software and digital goods. Indeed, all are the
subject of extensive qualitative analyses, often by interested observers.15 Again, however,
systematic evidence from which to draw analytical lessons is largely missing.
One key exception is pharmaceuticals, where issues of patents, market power, pricing,
and access to medicines loom large. The primary concern in developing economies as
they have implemented stronger patent rules is the potential for sharply increased prices,
diminished generic competition, and reduced availability of new drugs. Experience in the
United States and other developed economies shows that generic products entering at the
end of a patent take major shares of the market and drive prices down toward marginal
costs (Frank and Salkever, 1997; Reiffen and Ward, 2005). There is also the possibility that
originator firms suffer large market-share losses upon entry but the prices of their drugs
actually rise due to brand loyalty built under patent protection. As developing countries
register and enforce new drug patents the time of such entry likely will be delayed, perhaps
considerably. Generic companies may close down, consolidate or be taken over, generat-
ing even less competition and potentially longer waits before new medicines are imitated.
Thus, new patent regimes seem likely to raise significant challenges for both health and
competition authorities in developing economies.
Recent economic analysis has shed light on a few fundamental issues.16 First, consider
three detailed studies of potential impacts of new patents on drug prices in India, a
15
A partial list of comprehensive books would include Maskus and Reichman (2005),
Melendez-Ortiz and Roffe (2009), and Gervais (2007).
16
There are other critical issues, such as the need for additional public funding to meet global
needs for R&D in neglected diseases, the scope for advanced market commitments in new drugs,
and the effectiveness of exhaustion-based policy regimes to encourage price differentiation across
markets at different income levels. There is little in the way of serious empirical analysis of these
matters and I leave them aside to conserve space in this chapter. See Maskus (2012) for a discussion.

country with very low prices prior to its new patent law in 2005, extensive data, and
a deep generic industry. A study of data before 2005 suggested there was potential for
considerable price increases (Chaudhuri et al., 2006). The authors developed a structural
econometric model of the Indian market for quinolones, a family of broad-spectrum
antibiotics, such as ciprofloxacin. Using monthly data on prices and sales by firm from
January 1999 through December 2000, they estimated a demand system permitting drug
substitution across competing products. Using the estimated elasticities, they simulated
the impacts of patent protection by removing domestic competition in some or all of
these drugs. Thus, eliminating just domestic ciprofloxacin would increase prices of
three foreign competing drugs by up to 315 percent and also increase prices of related
domestic molecules by more than 100 percent. Removing domestic competition in all four
quinolones would raise foreign prices by a factor of between four and six. The associated
Indian welfare losses were predicted to be $156 million to $400 million per year. However,
the rise in profits to foreign pharmaceutical companies was estimated at just $53 million,
suggesting that the large static welfare loss would not be offset by comparable dynamic
incentives for innovation.
A more comprehensive analysis used data in India for 155 drugs in five therapeutic
groups, including many products with a foreign presence (Dutta, 2011). Using data
from 2001–03, the author estimated a structural model of the market, accounting for a
number of important demand and cost features. The model was simulated to compute
the potential price effects of providing patents in the 40 goods that had a foreign pres-
ence in India and did not face price controls. On average those drug prices would go up
by 18 percent, though the effects ranged from 3.5 to 80 percent. However, in another
simulation where some patents were accompanied by the elimination of price controls,
the price increases were considerably larger. Overall, the author computed a consumer
welfare loss of around $380 million per year, with perhaps 8.5 million patients choosing
not to buy the drugs.
Thus, simulation analyses based on pre-patent prices predicted notable price hikes in
India. However, a more recent study considered impacts on actual prices after the 2005
patent law and found far smaller impacts (Duggan et al., 2016). The authors developed
a database of 6,000 products in approximately 1,000 molecules, around 1/3 of which
were afforded patents by late 2011. They found a modest impact of patents, with prices
going up an average of 3 percent. Much of this increase came in newer molecules, which
received stronger protection in the law. This small price effect, however, likely was related
to competition: Indian law permitted existing firms that competed in newly patented
drugs to continue to produce under license. For those molecules with just one producer,
prices rose an average of 20 percent. The results point strongly toward the importance of
other policies, specifically price regulation and compulsory licensing where possible, to
limit the price impacts of patents.
This logic is reinforced in the only study to date of the drug-price effects of TRIPS-
related pharmaceutical patent laws across many countries (Kyle and Qian, 2014). The
sample covered 60 nations, about 2/3 of which were developing or transition economies,
permitting the authors to exploit changes in the implementation timing of such laws
among the latter group. They found that patented drugs have higher average prices than
those not patented, as expected. However, the price premiums associated with drugs pat-
ented after TRIPS compliance in the middle-income countries were modest and perhaps

negative in the poor nations. The authors attributed this outcome to the possible impacts
of price controls and other regulations, though they did not test this claim.
Another issue that has attracted attention is the impact of patent availability on the
willingness of pharmaceutical companies to launch their new products in different
markets. Two recent studies are particularly noteworthy and both point to the same basic
conclusion. The first studied the timing of launches of 642 new drugs in 76 countries over
1983–2002, thus covering a period before most TRIPS changes were made (Cockburn
et al., 2016). Controlling for a variety of endogeneity concerns, the authors found that
launches were accelerated in countries with longer and broader patents and in countries
with health policy institutions and demographic factors that favored profitability.
Launches were delayed by price regulations. The second study, undertaken in the TRIPS
era, found that the absence of patents significantly reduced the likelihood of a new drug
entering a market, while patent availability encouraged faster launches (Kyle and Qian,
2014).
A third critical issue for development purposes is whether the adoption in developing
countries of pharmaceutical patents and minimum protection standards, as set out in
TRIPS, is likely to incentivize more R&D into the particular medical needs of poor
nations. The new regime’s potential for such impact was a key promise of TRIPS advo-
cates and deserves serious scrutiny. It may be that implementation is too new, and the
potential impacts on R&D too delayed by time lags, to reach any conclusions at this point.
However, two observations may be made. First, there is one preliminary study of how
patent-law changes affected R&D investments from 1990 to 2003 (Kyle and McGahan,
2012). The fact that TRIPS compliance occurred at different times and across countries
with different relative disease burdens allowed the authors to study how global disease-
specific R&D investments (measured as clinical trials) were affected, distinguishing global
diseases from ‘neglected diseases’ of greatest interest in poor regions. The authors found
no indications of an increase in clinical trials in neglected diseases after TRIPS, although
there were significant increases in investments in global maladies with a large presence in
high-income countries. Second, early analysis suggests that major Indian pharmaceutical
companies sharply increased R&D and product development in the period surrounding
the 2005 patent law (Arora et al., 2008). That country is now among the largest global
suppliers of lower-cost drugs and a number of global pharmaceutical companies have
established R&D facilities in India. Thus, the industry is growing and consolidating,
perhaps as a result of patenting opportunities. However, the investments to date do not
seem to have focused on developing new drugs for neglected diseases.
To summarize, the evidence on patents and pricing power in developing nations is
scarce but the emerging evidence points to mixed messages. On the one hand, newly
protected patents in countries with limited competition may support markedly higher
prices, though this impact can be effectively countered with well-designed price regula-
tions and licensing regimes. On the other, countries with weak patent scope and extensive
price controls suffer lengthy delays before new products arrive in their markets. Thus,
policy institutions matter a great deal for access to medicines and health authorities have
deep tradeoffs to consider in the wake of TRIPS. Finally, there is little evidence to date
that the globalized patent regime is raising incentives for private R&D into the diseases of
poor countries. A solution for this last issue, therefore, remains in the purview of public
authorities and foundations.

V. CONCLUDING REMARKS
It is difficult to characterize the roles IPRs may play in the economic development
process, given the great complexity of the issue and the variability of potential impacts
across sectors and time. Broadly speaking, the selection of intellectual property regimes
is endogenous. We would expect lower-income economies with limited technological
capabilities to adopt weaker systems with broad limitations and exceptions. As countries
get richer and move into more advanced manufacturing and service sectors interests
emerge in deeper protection. Two immediate implications are that international attempts
to harmonize IPRs at TRIPS levels or even higher TRIPS-Plus standards may be sub-
optimal for many participants. A first-order issue for economists going forward is to
investigate whether emerging impacts are harming or helping development prospects, and
under what circumstances.
Despite this limited knowledge, economic analysis has made progress in understanding
some important development issues and their relationship to IPRs, especially patent laws.
In brief, and again noting that circumstances are highly variable across countries, the
following conclusions may be drawn. First, stronger global patents do seem to stimulate
marginally more R&D investments, but such effects are concentrated in the developed and
higher-income emerging economies. There is no evidence that measurable innovation is
growing in lower-income countries, nor is there any suggestion that the new regime has
increased private R&D incentives in important products for those markets.
Second, there are strong indications that enhanced IPRs encourage more and higher-
quality technology diffusion through market-based channels, including trade, FDI, and
licensing. The associated spillover gains in domestic productivity should offer a welcome
long-term boost to recipient economies. Again, however, this impact is prevalent only in
larger and middle-income countries that have a sound basis of intermediate technological
skills and education, which are important for absorbing and improving these technologies.
Technology transfer to lower-income economies in the TRIPS era has not expanded
significantly, raising numerous questions about the reasons for this lack of responsiveness.
Moreover, there remains no systematic evidence about how IPRs may be limiting the
scope for learning and diffusion through non-market means, including reverse engineer-
ing and imitation. This is another first-order area for additional research.
Finally, there are many important questions that could not be covered here and about
which we have inadequate information. For example, how are creativity and innovation
sustained in poor economies with large informal sectors, and is there any real role for IPRs
in that context? If product counterfeiting and unauthorized copying of digital products
limit development of new products and services in developing economies, how effective
are trademarks and copyrights in addressing such problems, and at what social cost? Does
the need to invest public resources in administering and enforcing a TRIPS-compliant
IPRs regime divert enough scarce talent to retard growth prospects? Most importantly,
what are the key thresholds, in terms of education, science, infrastructure, and factor
markets, that developing countries need to achieve before patents and other IPRs help
improve the dynamic efficiency of developing countries? There is a large research agenda
remaining.

REFERENCES
Allred, Brent B., and Walter G. Park. 2007a. ‘The Influence of Patent Protection on Firm Innovation Investment
in Manufacturing Industries,’ 13 Journal of International Management 91–109.
Allred, Brent B., and Walter G. Park. 2007b. ‘Patent Rights and Innovative Activity: Evidence from National
and Firm-Level Data,’ 38 Journal of International Business Studies 878–900.
Arora, Ashish, Lee Branstetter, and Chirantan Chatterjee. 2008. ‘Strong Medicine: The Impact of Patent
Reform on the Indian Pharmaceutical Industry,’ Working Paper, Carnegie-Mellon University, presented at
National Bureau of Economic Research Workshop on Location Decisions of Multinational Pharmaceutical
Enterprises, Savannah, Georgia.
Boldrin, Michaele, and David K. Levine. 2008. Against Intellectual Monopoly. Cambridge: Cambridge University
Press.
Branstetter, Lee, Ray Fisman, and C. Fritz Foley. 2006. ‘Do Stronger Intellectual Property Rights Increase
International Technology Transfer? Empirical Evidence from US Firm-Level Patent Data,’ 121 Quarterly
Branstetter, Lee, Ray Fisman, C. Fritz Foley, and Kamal Saggi. 2011. ‘Does Intellectual Property Rights Reform
Spur Industrial Development?,’ 83 Journal of International Economics 27–36.
Chaudhuri, Shubham, Pinelopi Goldberg, and Panle Jia. 2006. ‘Estimating the Effects of Global Protection of
Pharmaceuticals: A Case Study of Quinolones in India,’ 96 American Economic Review 1477–513.
Chen, Qiang. 2008. ‘The Effect of Patent Laws on Invention Rates: Evidence from Cross-Country Panels,’ 36
Journal of Comparative Economics 694–704.
Chen, Yongmin, and Thitima Puttitanun. 2005. ‘Intellectual Property Rights and Innovation in Developing
Countries,’ 78 Journal of Development Economics 474–93.
Cimoli, Mario, Giovanni Dosi, Keith E. Maskus, Jerome H. Reichman, and Joseph E. Stiglitz. 2014.
‘Innovation, Technical Change, and Patents in the Development Process,’ in Mario Cimoli, Giovanni Dosi,
Keith E. Maskus, Ruth L. Okediji, Jerome H. Reichman, and Joseph E. Stiglitz, eds., Intellectual Property
Rights: Legal and Economic Challenges for Development. Oxford: Oxford University Press.
Cockburn, Iain M., Jean O. Lanjouw, and Mark Schankerman. 2016. ‘Patents and the Global Diffusion of New
Drugs,’ 106 American Economic Review 136–64.
Cohen, Wesley M., Richard R. Nelson, and John P. Walsh. 2000. ‘Protecting their Intellectual Assets:
Appropriability Conditions and Why U.S. Manufacturing Firms Patent (or Not),’ Cambridge, MA: National
Bureau of Economic Research Working Paper 7552.
European Commission. 2016. Intellectual Property Rights Intensive Industries Contribution to Economic Performance
and Employment in the European Union. Brussels: Office for Harmonization in the Internal Market, accessed at
9 April 2019, http://www.europeantrademark.it/Files/intellectual_propery_rights_intensive_industries_study_
epo_ohim.pdf.
Du, Julian, Yi Lu, and Zhigong Tao. 2008. ‘Economic Institutions and FDI Location Choice: Evidence from
US Multinationals in China,’ 36 Journal of Comparative Economics 412–29.
Duggan, Mark, Craig Garthwaite, and Aparajita Goyal. 2016. ‘The Market Impacts of Pharmaceutical Product
Patents in Developing Countries: Evidence from India,’ 106 American Economic Review 99–135.
Dutta, Antara. 2011. ‘From Free Entry to Patent Protection: Welfare Implications for the Indian Pharmaceutical
Industry,’ 93 Review of Economics and Statistics 160–78.
Eaton, Jonathan, and Samuel Kortum. 1996. ‘Trade in Ideas: Patenting and Productivity in the OECD,’ 40
Journal of International Economics 251–78.
Fagerberg, Jan, Martin Srholec, and Bart Verspagen. 2010. ‘Innovation and Economic Development,’ in
Bronwyn H. Hall and Nathan Rosenberg, eds., Handbook of the Economics of Innovation: Volume 1.
Amsterdam, Netherlands: Elsevier-North Holland Publishing Co.
Frank, Richard G., and David S. Salkever. 1997. ‘Generic Entry and the Pricing of Pharmaceuticals,’ 6 Journal
of Economics and Management Strategy 75–90.
Gervais, Daniel J., ed. 2007. Intellectual Property, Trade and Development: Strategies to Optimize Economic
Development in a TRIPS-Plus Era. Oxford: Oxford University Press.
Ginarte, Juan Carlos, and Walter G. Park. 1997. ‘Determinants of Patent Rights: A Cross-National Study,’ 26
Research Policy 283–301.
Goeschl, Timo, and Timothy Swanson. 2000. ‘Genetic Use Restriction Technologies and the Diffusion of Yield
Gains to Developing Countries,’ 12 Journal of International Development 1159–78.
Griliches, Zvi. 1957. ‘Hybrid Corn: An Exploration in the Economics of Technical Change,’ 25 Econometrica
501–22.
Grossman, Gene M., and Elhanan Helpman. 1994. ‘Protection for Sale,’ 84 The American Economic Review 33–850.
Grossman, Gene M., and Edwin L.-C. Lai. 2004. ‘International Protection of Intellectual Property,’ 94 American

Hoekman, Bernard, Keith E. Maskus, and Kamal Saggi. 2005. ‘Transfer of Technology to Developing
Countries: Unilateral and Multilateral Policy Options,’ 33 World Development 1587–602.
Hovhannisyan, Nune, and Wolfgang Keller. 2015. ‘International Business Travel: An Engine of Economic
Growth?,’ 20 Journal of Economic Growth 75–104.
Hu, Albert G. 2009. ‘The Regionalization of Knowledge Flows in East Asia: Evidence from Patent Citations
Data,’ 37 World Development 1465–77.
Hu, Albert G., and Adam B. Jaffe. 2003. ‘Patent Citations and International Knowledge Flow: The Cases of
Korea and Taiwan,’ 21 International Journal of Industrial Organization 849–80.
Idris, Kamal. 2002. Intellectual Property: A Power Tool for Economic Growth. Geneva, Switzerland: World
Intellectual Property Organization.
Ivus, Olena. 2010. ‘Do Stronger Patent Rights Raise High-Tech Exports to the Developing World?,’ 81 Journal
of International Economics 38–47.
Keller, Wolfgang. 2004. ‘International Technology Diffusion,’ 42 Journal of Economic Literature 752–82.
Keller, Wolfgang. 2010. ‘International Trade, Foreign Direct Investment, and Technology Spillovers,’ in
Bronwyn H. Hall and Nathan Rosenberg, eds., Handbook of the Economics of Innovation: Volume 1.
Amsterdam, Holland: Elsevier-North Holland.
Kim, Yee Kyoung, Keun Lee, and Walter G. Park. 2012. ‘Appropriate Intellectual Property Protection and
Economic Growth in Countries at Different Levels of Development,’ 41 Research Policy 350–75.
Kyle, Margaret K., and Anita M. McGahan. 2012. ‘Investments in Pharmaceuticals Before and After TRIPS,’
94 Review of Economics and Statistics 1157–72.
Kyle, Margaret K., and Yi Qian. 2014. ‘Intellectual Property Rights and Access to Innovation: Evidence from
TRIPS,’ Cambridge, MA: National Bureau of Economic Research, Working Paper 20799.
Lai, Edwin L.-C., and Isabel K.M. Yan. 2013. ‘Would Global Patent Protection Be Too Weak Without
International Coordination?,’ 89 Journal of International Economics 42–54.
Lai, Edwin L.-C., Samuel Wong, and Isabel K. Yan. 2008. ‘International Protection of Intellectual Property: An
Empirical Investigation,’ Hong Kong University of Science and Technology, manuscript.
Lerner, Josh. 2002. ‘Patent Protection and Innovation Over 150 Years,’ Cambridge MA: National Bureau of
Economic Research, Working Paper 8977.
Markusen, James R. 2002. Multinational Firms and the Theory of International Trade. Cambridge, MA: MIT
Press.
Maskus, Keith E. 2000. Intellectual Property Rights in the Global Economy. Washington DC: Peterson Institute
for International Economics.
Maskus, Keith E. 2012. Private Rights and Public Problems: The Global Economics of Intellectual Property in
the 21st Century. Washington DC: Peterson Institute for International Economics.
Maskus, Keith E., and Christine McDaniel. 1999. ‘Impacts of the Japanese Patent System on Productivity
Growth,’ 11 Japan and the World Economy 557–74.
Maskus, Keith E., and Mohan Penubarti. 1995. ‘How Trade-Related Are Intellectual Property Rights?,’ 39
Journal of International Economics 227–48.
Maskus, Keith E., and Jerome H. Reichman, eds. 2005. International Public Goods and Transfer of Technology
Under a Globalized Intellectual Property Regime. Cambridge: Cambridge University Press.
Maskus, Keith E., and Lei Yang. 2018. ‘Patent Rights, Access to Technologies, and the Structure of Exports,’
51 Canadian Journal of Economics 483–509.
Melendez-Ortiz, Ricardo, and Pedro Roffe, eds. 2009. Intellectual Property and Sustainable Development
Agendas in a Changing World. Cheltenham, UK and Northampton, MA, USA: Edward Elgar Publishing.
Moser, Petra. 2005. ‘How Do Patent Laws Influence Innovation? Evidence from Nineteenth-Century World’s
Fairs,’ 94 American Economic Review 1214–36.
Moser, Petra. 2013. ‘Patents and Innovation: Evidence from Economic History,’ 27 Journal of Economic
Moser, Petra, and Alessandra Voena. 2012. ‘Compulsory Licensing: Evidence from the Trading with the Enemy
Act,’ 102 American Economic Review 396–427.
Nunnenkamp, P., and J. Spatz. 2004. ‘Intellectual Property Rights and Foreign Direct Investment: A
Disaggregated Analysis,’ 140 Review of the World Economy 393–414.
Odagiri, Hiroyuki, Akira Goto, Asushi Sunami, and Richard R. Nelson, eds. 2010. Intellectual Property Rights,
Development, and Catch-Up. Oxford: Oxford University Press.
Ordover, Janusz A. 1991. ‘A Patent System for Both Diffusion and Exclusion,’ 5 Journal of Economic Perspectives
43–60.
Park, Walter G. 2008a. ‘International Patent Protection: 1960–2005,’ 37 Research Policy 761–6.
Park, Walter G. 2008b. ‘Intellectual Property Rights and International Innovation,’ in Maskus, Keith E., ed.,
Intellectual Property, Growth and Trade: Frontiers of Economics and Globalization. Amsterdam, Netherlands:
Elsevier North-Holland.
Pouris, A., and R. Inglesi-Lotz. 2011. The Economic Contribution of Copyright-Based Industries in South Africa.

Geneva, Switzerland: World Intellectual Property Organization, accessed at 9 April 2019, http://www.thedti.
gov.za/industrial_development/docs/Economic_Contribution.pdf.
Qian, Yi. 2007. ‘Do National Patent Laws Stimulate Domestic Innovation in a Global Patenting Environment?
A Cross-Country Analysis of Pharmaceutical Protection, 1978–2002,’ 89 Review of Economics and Statistics
436–53.
Reiffen, David, and Michael R. Ward. 2005. ‘Generic Drug Industry Dynamics,’ 87 The Review of Economics
Romer, Paul M. 1990. ‘Endogenous Technological Change,’ 98 Journal of Political Economy S71–102.
Rosen, William A. 2010. The Most Powerful Idea in the World: A Story of Steam, Industry, and Invention. New
York, NY: Random House.
Schneider, Patricia Hingo. 2005. ‘International Trade, Economic Growth and Intellectual Property Rights: A
Panel-Data Study of Developed and Developing Countries,’ 78 Journal of Development Economics 529–47.
Schumpeter, Joseph A. 1942. Capitalism, Socialism and Democracy. New York, NY: Harper and Brothers.
Scotchmer, Suzanne. 2004. ‘The Political Economy of Intellectual Property Treaties,’ 20 Journal of Law,
Economics and Organization 415–37.
Smarzynska Javorcik, Beata. 2004. ‘The Composition of Foreign Direct Investment and Protection of
Intellectual Property Rights: Evidence from Transition Economies,’ 48 European Economic Review 39–62.
Stiglitz, Joseph E., and Andrew Charlton. 2005. Fair Trade for All: How Trade Can Promote Development.
Swanson, Timothy, and Timo Goeschl. 2014. ‘The Distributive Impact of Intellectual Property Regimes: A
Report from the “Natural Experiment” of the Green Revolution,’ in Mario Cimoli, Giovanni Dosi, Keith
E. Maskus, Ruth L. Okediji, Jerome H. Reichman, and Joseph E. Stiglitz, eds., Intellectual Property Rights:
Legal and Economic Challenges for Development. Oxford: Oxford University Press.
World Bank. 2002. Global Economic Prospects and the Developing Countries. Washington DC: World Bank.
Yang, Guifang, and Keith E. Maskus. 2001. ‘Intellectual Property Rights, Licensing, and Innovation in an
Endogenous Product Cycle Model,’ 53 Journal of International Economics 169–87.

Index
Abramowicz, M. 51, 121, 272, 273, 293, 332, patent proliferation issues 34
350–75, 379, 384, 620 patent trolls 36, 37
Abramson, B. 382, 393 patented genes in genetic tests 30–32, 38
abstraction levels 177, 178, 388, 558, 559, 560, and privatization 27
561, 573 survey results and form of questions 31
Abulude, S. 589, 597 tragedy of the anticommons 27, 28–9
Adams, W. 382 upstream research impact 30–33
Adler, A. 100, 102, 104, 106 see also commons
Adler, J. 355, 372 anticommons, transaction costs, and patent
advance purchase contracts 603 aggregators, risk mitigation changes 37–42
adverse possession doctrine 18, 22, 431, 446–7 aggressive patent assertions 38–9
aggressive patent assertions 38–9 ambiguous language use and future
Aghion, P. 232 innovation risk 41
Aitken, H. 195 claim interpretation issues 40–41
Akerlof, G. 126 doctrine of equivalents restrictions 41
Akpotaire, U. 608 Federal Circuit invalidation of broad
Alfano, M. 88 university patents 39
Allison, J. 41, 123, 216, 387, 390, 393, 417, 418 information costs 40–41, 42
allocative efficiency 288, 290, 385–6, 595, 614 injunctions, limits on 39–40
Allred, B. 664 and patent aggregation 42
Alschuler, A. 86 patent challenges, lowering costs of 41–2
Altman, R. 569 publicly available information about patent
Amabile, T. 394, 563 ownership, legal changes suggestion 41
Ames, J. 447 upstream rights and assertions, limits on
Anand, B. 335 38–9
Anderson, C. 595, 603, 604, 622 utility requirement for patentability 38
Anderson, J. 233, 443, 475, 476, 481, 483 Anton, J. 135, 428
Anderson, T. 54, 413 Aoki, M. 642
anti-competitive effects 35, 37, 234, 236–41, application program interface (API) protection
245–53, 255, 622 160, 162–3, 167, 172, 175–6, 179–80,
see also competition 189–91, 218–19
anticommons, transaction costs, and patent appropriability issues
aggregators 27–46, 242 network effects 166, 167, 173, 188, 211, 218,
anticompetitive effects 35, 37 220
bundling of rights 28 open innovation and ex ante licensing 338–9,
clinical genetic tests 34 341
copyright ownership fragmentation 29 product differentiation 270, 271
double marginalization problem 29, 35 uniformity cost reduction 379, 381, 391, 392,
downstream product development, impact 394, 395–7
on 33–5, 38–9 Arai, K. 232
fragmented ownership of intellectual Archibald, G. 264, 266, 269
property (IP) rights 28 Areeda, P. 210, 237, 241, 243, 246, 251, 252,
future research 31 253
information asymmetries 28 Arewa, O. 587
litigation costs 36 Armond, M. 133
patent aggregators, rise of 35–7 Armstrong, M. 287
patent portfolios and enhanced bargaining Arora, A. 122, 333, 334, 338, 671
position 35–6 Arora, S. 52
677

Arrow, K. 6, 120, 232, 331, 332, 334–5, 378, Beschorner, P. 272

385, 427–8, 553 Besen, S. 134, 164, 272, 290, 296, 301, 498
Arsenault, A. 619 Bessen, J. 4, 10, 16, 35, 36, 38, 40, 49, 59, 137,
Article III trial courts 481, 483, 486 139–40, 212, 335, 384, 396, 416, 417, 439,
Aschauer, D. 552 446, 447, 448
Athique, A. 583, 590, 591, 605–6, 617 best practice statements, custom role 532–3,
Aufderheide, P. 29, 529–30, 533 543–4, 545
Ayres, I. 130, 134, 388, 446 Bhaskar, V. 269, 385
Biagioli, M. 314
Bagley, M. 395 Bibb, M. 19
Bain, J. 275 bilateral (ex ante) license 333–6
Baker, E. 282, 292, 585, 597, 613, 618, 620, 621 Bishop, R. 266
Baker, J. 210, 292, 395 Bitzer, J. 395
Bakker, G. 356 Bjornstad, D. 29
Bakos, Y. 134, 300, 301 Blackstone, W. 527, 528, 543
Baldwin, C. 52, 332, 336, 339, 340, 341, 342 Blair, R. 132, 135
Balganesh, S. 102, 291, 409, 412, 508–25 blocking rights, innovation, cumulative
Ball, G. 483 innovation 141–2, 148
Band, J. 186 Bogers, M. 337
Banerjee, Abhijit 650, 651, 652 Bohannan, C. 205, 206, 232, 234, 235, 237,
Banerjee, Arpan 589 249, 250, 252, 255, 256
Banks, J. 113 Boldrin, M. 234, 379, 417, 662
Bar-Gill, O. 125, 292, 334, 335 Bonanno, G. 269
Barak, B. 52 Bone, R. 411, 418
Barnes, D. 274, 275 Bontis, N. 570
Barnett, J. 28, 29, 35, 37, 56, 61, 66, 166, 212, Booth, G. 591, 607, 608, 613, 625
313, 334, 335, 385, 396, 411, 448, 563, 566, Borenstein, S. 269, 270
595, 611, 617 Bork, R. 299
Barro, R. 639 Bose, D. 600
Barron, A. 565 Bouchard, R. 395
Barros, B. 7 Boudreau, K. 337
Barrot, P. 606 Bourdieu, P. 552
Basberg, B. 637 Bowman, W. 237, 238, 292, 296
Baseman, K. 164 Boyle, J. 5, 19, 84, 102, 291, 292, 553, 560, 567,
Battacharyya, S. 324 569, 586
Baumol, W. 269 Bracha, O. 263, 273, 378, 382, 384–5, 411, 594,
Baxter, W. 293 620
Beath, J. 266, 270 Braithwaite, J. 384
Bebchuk, L. 411 Brander, J. 269
Beckles, T. 529–30 Branscomb, L. 552
Beebe, B. 103, 110, 418 Branstetter, L. 664, 667
Beggan, J. 89 Bratton, W. 523
Bell, A. 17, 391 Brauneis, R. 457
Bell, T. 287, 291, 414 Brennan, T. 356, 373
Belleflamme, P. 232 Bresnahan, T. 199
Benesch, C. 110 Breyer, S. 5, 83, 398–9, 603, 609
Benjamin, S. 480, 483 Bridy, A. 409, 415
Benkler, Y. 66, 185, 292, 313, 330, 333, 338, Briggs, J. 213
340, 341, 342, 551–2, 553, 554, 557, 560, Broache, A. 439
561, 562, 563, 567, 618, 620 Bronsteen, J. 103, 105, 106, 107, 108, 109, 114
Berkowitz, M. 390 Brown, D. 88
Berners-Lee, T. 185 Brown, E. 89
Bernheim, B. 412 Brown, R. 274
Bernstein, L. 313, 528–9, 542 Browne, J. 543
Berry, S. 273 Brudner, A. 510

Index 679
Bruni, L. 110 Chen, Yeh-ning 134

Brunt, L. 353–4 Chen, Yongmin 659, 664
Brynjolfsson, E. 300 Chesbrough, H. 330, 336, 337, 338
Buccafusco, C. 81, 98–118, 203, 313, 316 Chiang, T.-J. 16, 19, 61, 439
Buchanan, J. 28, 273 Chien, C. 35, 83, 124, 339, 345–6, 416, 417,
Bud, A. 587, 589, 626 436, 481
Budish, E. 81 Chile, intellectual property law 649
bundling, and price discrimination 285–6, China, creative development see under creative
297–300 development and copyright, emerging
Burk, D. 57, 108, 124, 141, 235, 334, 335, 381, industries in Global South
383, 384, 385, 386, 387, 391, 393, 394, Chisum, D. 173, 177, 183
426–7, 474, 478–9, 481, 483–4 Cho, M. 30, 34
Burrell, R. 353, 371 Choi, J. 300
Burri, M. 618, 619, 622, 623 Chomsky, N. 88
Burstein, M. 330–49, 355, 378, 385, 428, 463, Chon, M. 594, 618, 652
464, 484, 566 Chu, C. 477
business method patents 143, 197–8, 201, 384, Cimoli, M. 652, 661
392, 394–5 Clancy, M. 357, 362
Butler, H. 298 Clapes, A. 217
Byrne, D. 325 Clark, K. 52, 332
clearance culture, custom role 529–30, 532,
Cadenas, H. 649 539, 540–41, 542
Cai, M. 146 Coase, R. 14, 49, 250, 334, 361, 370, 385
Calabresi, G. 6, 134 Cockburn, I. 671
Calandrillo, S. 359, 366, 367–8, 369, 370–71 Cohen, J. 4, 57, 113, 291, 385, 387, 394, 414,
Campbell-Kelly, M. 188 553, 557, 560, 563, 567, 595
Canada, collective society definition 490 Cohen, W. 31, 36, 332, 382, 389, 662
Carbone, J. 30 collaborative innovation, open innovation and
Cardozo, B. 510 ex ante licensing 333–45
Carlton, D. 299 collective licensing 458–9, 600–601
Carniaux, M. 42 collective management 489–507
Carrier, M. 4, 5, 35, 51, 216, 384, 414 complementary works and substitutive
Carroll, J. 415 works, difference between 500–501
Carroll, M. 372–3, 377–406, 453 consumers and product variety versus
Carter, S. 527 monetary incentive 499
Carver, B. 162, 184 and copyright 495, 499, 501, 503–5
Casey, A. 595 delegation of authority to license 491–2
Cass, R. 100 distribution to represented parties 493–4
Castells, M. 619 economic analysis 494–503
Caulfield, T. 32 efficiency measurement 495, 502–3
Ceruzzi, P. 183 game theory model 496–7
Chafee, Z. 61 individual works’ valuation 498–501
Chamberlin, E. 263–4, 265, 266, 267, 274–5, membership numbers and administrative
277 costs 498–9
Chander, A. 4, 6 online notice-and-take-down system 504
Chandler, A. 336 and pricing 495, 501
Chang, H. 137, 141 proportional distribution issues 499–500
Chang, X. 232 repertory price setting 492, 495–8
Chang, Y. 54 technology and online uses 503–5
channeling doctrines 122, 136, 150 transaction costs reduction 495, 501
Chari, V. 361 user identification and licensing 492–3
Charlton, A. 648, 657 willingness-to-pay model 497–8
Chaudhuri, S. 670 collective-rights organizations 57, 58, 66
Che, Y.-K. 121 Collins, K. 50
Chen, Q. 662 Comanor, W. 274–5

commercialization effects 51–2, 58, 59, 60, 61, nonprice restrictions 245
125, 128, 331–2, 336, 337–8, 368–9 patent aggregation and market power 242
common law 508–25 patent misuse doctrine 238, 247, 255–6
and copyright 513–14, 516, 520–21 patent pools 248–51
and custom role 527–8, 536, 543 pay-for-delay settlements 239
evolutionary approach 512–14 political economy of intellectual property
fair use doctrine 513–14, 516, 520 law 231–5
and federalism 521–3 price-fixing 236, 238–9, 241, 243–4
freestanding 517, 518–20 producer involvement 235
future research 523, 524 product differentiation 242, 263–9, 272–4
interstitial 517–18, 519, 520–21 standards-essential patents and FRAND
as judge-made law 514–16 licensing 242
as method of lawmaking 511–14, 518 tying arrangement 246–8, 250–51
and misappropriation doctrine 519 vertical restraints 244–8
patent law’s doctrine of equivalents 521 complementary products 159–60, 162, 164,
precedent and stare decisis use 511–12 168, 183, 194, 269, 285–6, 297–300, 368,
as state law 521–3 500–501
as substantive law 517–21 compulsory licensing 57, 59, 60, 135, 148, 170,
and technological development 513 253, 455
trade secret law 519, 520 Conner, K. 134, 301
utility-maximization 515–16 Connor, J. 239
commons 556–7, 561–2, 601–4 consequentialism see under philosophical
Creative Commons 149, 343, 460, 533 foundations of IP law
open source 148, 149–50, 162–3, 171, 185, consumers, effects on 107–12, 159–60, 161,
216, 460 167, 235, 236, 266–8, 269, 271, 369–70,
semi-commons regime 56, 66, 559–60, 571–3 499
see also anticommons, transaction costs, and context-specific copyright 615–16
patent aggregators Conti, A. 137
competition 231–61 contract law 295–6, 519, 528, 542
anti-competitive effects 35, 37, 234, 236–41, Contreras, J. 38, 66, 148, 162, 165, 166, 195,
245–53, 255, 622 199, 242, 339, 428, 597
antitrust policy and IP policy, relationship Cook, P. 396
between 235–40 Cook, V. 88
and consumer welfare 235, 236 Coombe, R. 586
copying 233, 234, 243 Cooter, R. 15, 135, 511
copyright misuse 255 copying
development cost factors 242–3 competition 233, 234, 243
digital content duplication 243 and custom role 534, 536
doctrine of equivalents and patent law 238 innovation, cumulative innovation 138, 144,
economic growth 232, 242–3 146
and exclusionary practices 237, 239, 241–2, innovation, stand-alone innovation 129, 133,
245, 250, 251–3, 256 134
exclusivity rights 244, 245–6, 252, 254 negative space of IP 310–11, 314–15,
exhaustion (first sale) doctrine 238, 246, 254 317–18, 320, 321–2, 324–5
functionality protection limitations 256–7 network effects 176, 178, 179, 183, 189–90,
industry structure and market position 233 193–4
and innovation 129, 233–4, 237 and notice and disclosure 428–9, 437, 447,
interbrand and intrabrand restraints 245–6 461
and market power 233, 235–6, 241–3, 244 and price discrimination 291, 298–9
network effects see network effects, and uniformity cost reduction 388, 391, 393,
competition policy 396
non-patent IPRs 243 copyleft licensing model 184–5, 343
non-practicing patent holders 234 copyright
nonobvious subject matter and patent law anticommons, transaction costs, and patent
234, 236–7 aggregators 29

Index 681
and collective management 495, 499, 501, future research 629

503–5 lead-time advantage 603, 609
and common law 513–14, 516, 520–21 creative development and copyright, cultural
and competition 255 diversity 618–29
context-specific 615–16 and alternative revenue models 621–2
and creative development see creative anticompetitive abuse 622
development and copyright commodity culture 620–21
creativity not protected by see negative space developed markets, effects on 619–23
of IP developing world 623–6
and cumulative innovation 144, 145, 148 digitization effects 619, 622–3
and custom role 531–2, 534–5, 536–7, 545 global diversity 626–9
enforcement costs see enforcement costs, market structure 621, 627
copyright law enforcement winner-take-all biases of media markets 621
and fair use doctrine 60, 144, 146–7, 187, creative development and copyright, emerging
456, 458, 459, 531–2, 534–5, 536, industries in Global South 584–93
545 China and film industry 591–3, 598–600,
first-sale (exhaustion) doctrine 63–4, 207, 606, 609, 611, 616, 617, 627
238, 246, 254, 294 China and film industry, and piracy 592,
governance see governance of intellectual 605–6, 607, 608, 615, 618
property, patent versus copyright China and music industry 593, 610, 611, 613,
and innovation, stand-alone innovation 122, 617, 624
123, 133–4 digital technology effects 585–6, 588, 589,
intellectual property as property 10–11, 13, 590, 592, 605–10, 611, 613, 615, 617
18–19, 21–2 home-grown media and cultural progress
notice and disclosure see notice and 584–5
disclosure, copyright tracing and India and film industry 589–91, 599, 600,
ownership 606, 609, 624, 625–6, 627, 628, 652, 659
price discrimination see price discrimination, India and film industry, and piracy 590, 604,
copyright 607–8, 615
product differentiation 272–4 India and music industry 591, 607–8, 613,
and uniformity cost reduction 387–8, 617, 626
390–91, 394, 397 Nigeria and film industry 586–90, 596, 597,
welfare promotion 109–12, 115 598, 599, 600, 609, 611, 619, 624, 625–6,
work-for-hire doctrine 536–7 627, 628
Cornelli, F. 382 Nigeria and film industry, and piracy 605–6,
Cornes, R. 554 608, 615
Correa, C. 384 Nigeria and music industry 589, 608
Corts, K. 287 and TRIPS Agreement 586
costs see enforcement costs; information creative development and copyright, state
costs; transactions costs; uniformity cost support 593–604
reduction advance purchase contracts 603
Cotropia, C. 40, 381, 382, 387, 417, 418, 447 allocative efficiency 595, 614
Cotter, T. 132, 135, 252, 448 alternative business models 602–3
cover licensing 455 collective licensing 600–601
Cowen, T. 598, 600 commons-based creativity models 601–4
Creative Commons 149, 343, 460, 533 direct patronage 597–600
see also commons exclusivity issues 603–4
creative development and copyright 582–635 non-commercial creativity 601–2
context-specific copyright 615–16 prior commitments and path dependencies
copyright alternatives, hidden drawbacks 596
596, 610–14 private ordering and intermediaries 595–6
distribution of benefits 614–15 production incentives 594
exclusivity rights 593, 594, 597, 603–4, risk-reduction functions 595–6
608–10 creative expression and functionality,
formalization imperative 616–17 distinction between 217–18

creativity as benefit, welfare promotion 112–14 Dasgupta, P. 390

creativity, cumulative 429, 440, 457–60 D’Aspremont, C. 268
creativity not protected by copyright see Daughety, A. 410
negative space of IP (creativity not David, P. 29, 137, 552
protected by copyright) Davy, B. 444
Crouch, D. 233 De Beer, J. 336, 339
Csikszentmihalyi, M. 112 De Castor, I. 408
Cugno, F. 132 De Jong, J. 314
cultural diversity, and creative development De Laat, E. 121
see creative development and copyright, De Larena, L. 395
cultural diversity De Mot, J. 409
cultural environment as infrastructure, deadweight loss
infrastructure theory 560–62, 567, 570 prize and reward alternatives 355, 357, 358,
cumulative innovation see innovation, 364, 367, 369–70, 372
cumulative stand-alone innovation 101, 120, 123, 126,
custom role 526–50 130–31, 142
best practice statements 532–3, 543–4, 545 Deaton, A. 108
certainty of custom 543–4 DeBrock, L. 383, 390
classroom (educational) guidelines 531–2, decentralization effects 338, 341, 380, 381
535–6, 540, 545, 547 DeMartino, G. 560
clearance culture 529–30, 532, 539, 540–41, Demsetz, H. 4, 54, 57, 102, 287, 288, 385, 413,
542 448, 565
and common law 527–8, 536, 543 Deneckere, R. 289, 292
contract law 528, 542 Denicola, R. 453
and copying 534, 536 Denicolò, V. 65, 121, 140, 141, 271, 390
and copyright 531–2, 533, 534–5, 536–7, 545 Denton, F. 389
customary pricing 534–5 deontology see under philosophical
formal law role 533–7 foundations of IP law
future research 547 Depoorter, B. 4, 29, 281–308, 388, 407–23, 430,
implications of adoption of a role 545–6 456, 458, 500–501, 585, 596
in-house guidelines 532 depropertization 448, 460–62
incentive-rationale theory 538 Derclaye, E. 104, 107
industry practices 529–33, 538–41 derivative works 59–60, 113–14, 115, 141, 144,
IP-adjacent norms 530–31 162, 163, 184, 343, 563, 564, 566, 571–2
labor-reward analysis 538 Desai, D. 418
lock-in effect 539, 542–3 Desai, R. 628
market effects 534–5 design patents 123, 203–5, 218–19, 462–5
and parties’ expectations 541–3 Determann, L. 185
and private sector 540–41, 546 deterrence principle, network effects 171,
reliance issues 537–43 220–21
representativeness 545 Deuze, M. 113
special interest groups involvement 539–40 DeVany, A. 382
and tort law 528, 541 developing countries 360, 494, 623–6, 646,
trademark protection 538–9, 547 665–9
see also individual countries
Dagan, H. 4 Devlin, A. 112, 427
Dale, M. 595 DiCola, P. 29, 34, 569
Dam, K. 295 Ding, W. 357–8
damages 134, 135, 149, 448, 457 Dinwoodie, G. 40, 381, 418
see also injunctions direct patronage, creative development and
Dana, J. 299 state support 597–600
Danaher, B. 604, 623 disclosure requirements 144–5, 331, 334, 335–6
Darling, K. 313, 320, 411, 546, 610 and notice see notice and disclosure
Darwall, S. 73 district courts 40, 434, 444, 456, 474, 475, 476,
Das, P. 312 481

Index 683
Dixit, A. 266, 292 price discrimination 647

doctrine of equivalents 41, 129, 142–3, 238, technological advancement 659
447, 521 technological change 639–40, 641, 642, 643,
Doll, S. 569 644, 665–9
Donaldson, M. 457 technology transfers 642, 643, 647, 648,
Dore, M. 390 650–51
Drahos, P. 384 threshold effect 664
Drassinower, A. 7, 8 trade flows to developing countries and
Dreyfuss, R. 40, 123, 146, 216, 236, 324, 345, strength of IPR 646
418, 476, 482, 484, 569 and TRIPS 643, 644, 660–61, 667, 670–71
Du, J. 667 welfare distribution between consumers
Du Mont, J. 203, 204 658–9, 660
Duffy, J. 51, 135, 236, 254, 359, 361, 362, Edlin, A. 239, 240, 385
363–4, 367, 369, 370, 371, 381, 389, 390, education 494, 531–2, 535–6, 540, 545, 547, 652
479, 480, 482, 484, 565, 566 Ehrlich, I. 409
Duflo, E. 650, 651, 652 Eicher, T. 76
Duggan, M. 670 Eisenberg, A. 649
Dukeminier, J. 2 Eisenberg, M. 510
Dutta, A. 670 Eisenberg, R. 27–46, 66, 126, 143, 145–6, 341,
Dwyer, J. 431 372, 387, 389, 395, 408, 417, 438, 485
Dybvig, P. 164 Elazari Bar-On, A. 207
Dyson, J. 122 Elberse, A. 621, 622
Elhauge, E. 166, 248, 286, 299
Easterbrook, F. 3 Elkin-Koren, N. 63, 410
Eaton, B. 265, 268, 269, 270 Ellickson, R. 14, 50, 55, 312–13, 314, 413, 528,
Eaton, J. 666 540, 610
economic analysis, collective management Ellis, R. 50
494–503 emerging industries in Global South see
economic characteristics, infrastructure theory creative development and copyright,
see infrastructure theory, economic emerging industries in Global South
characteristics enforcement costs 407–23
economic development 636–55 adjustment of intellectual property rights
and competition 232, 242–3 413–14
cross border trade and multilateral copyright and small creators, effects on 412,
agreements 648 413
education access 652 distributive effects 412–13
endogenous regimes 658–61 doctrinal developments, effects of 409
foreign direct investment 646–7, 666, 667 future research 419
future research 647, 652, 653, 668, 672 and legal protection costs 408–9
global diffusion and convergence issues negative effects 410–11
638–9 opportunistic enforcement practices 411
global public goods 643–4 patent law enforcement 408, 416–18
global tradeoffs 660–61 patent law, repeat players and large firms,
innovation and patents 661–5 effects on 412–13
institutional and behavioral understandings patent trolls 416–18
650–51, 652 private enforcement as public good 411
legal policy 641–5, 652 technological advances, effects of 409–10,
mapping 645–52 414, 415–16
nation-state as unit of analysis 643 trademark law enforcement 418
neoclassical growth model 638–9, 641, 644 enforcement costs, copyright law enforcement
network effects 159, 161–2, 163 414–16
parallel imports 647–8 innovation spillovers 414
and patents 666–7, 669–71 small claims courts 416
population quality measures 649, 651–2 social norm complications 414–15
preferential trade agreements 660–61 statutory damages 415–16

Englmaier, F. 286 Feldman, A. 73, 81

Epstein, L. 485 Feldman, R. 35, 36, 61, 206, 417, 418
Epstein, R. 3, 7, 8, 11, 28, 29, 34, 35, 65, 146, Feldman, Y. 415
234, 528, 540–41 Fellmeth, A. 207
essential facilities doctrine 210 Fennell, L. 6, 14, 15, 28, 29, 388
Ethiraj, S. 54 Field, B. 413
EU, Directive on Collective Management Fields, G. 637
490–91, 495 Fierstein, R. 242
Evans, D. 188 film industry, emerging industries see under
Ewing, T. 35, 36, 417, 418 creative development and copyright,
exclusionary practices 5–6 emerging industries in Global South
and competition 237, 239, 241–2, 245, 250, Finger, J. 604
251–3, 256 Fink, C. 645–6, 647
governance of intellectual property 50–53, first mover advantage 234, 317–19, 338–9
57, 59–61, 63–4, 65–6 first-sale (exhaustion) doctrine 63–4, 133, 207,
infrastructure theory 562–6, 570 238, 246, 254, 294
network effects 199, 201, 202, 209, 211 Fischbaum, M. 361, 363
uniformity cost reduction 382, 383–4, 387, Fish, R. 438
397 Fisher, F. 247
exclusivity rights Fisher, W. 104, 290–91, 341, 356–7, 359, 360,
and competition 244, 245–6, 252, 254 365, 378, 381, 532, 553, 566, 600, 618
creative development and copyright 593, Fishman, J. 144, 623
594, 597, 603–4, 608–10 Fisk, C. 313
infrastructure theory 564, 567, 573 fixed ‘pot’ suggestion, prize and reward
innovation, cumulative innovation 140, 143, alternatives 358, 359, 367
144, 146, 148 fluid property 55–7
network effects 183, 195, 210, 211 follow-on innovation 102, 108, 114, 137,
open innovation and ex ante licensing 331, 140–41, 144–5, 147–8, 149, 398
334, 335–6, 338, 344–5 Foot, P. 103, 104
price discrimination 285, 295, 297, 299–300 Foray, D. 552
uniformity cost reduction 383, 391–4, 397 foreign direct investment 646–7, 666, 667
welfare promotion 100–101, 102, 103 Fraley, J. 7
exhaustion (first-sale) doctrine 63–4, 133, 207, franchisors, and competition 248
238, 246, 254, 294–5 FRAND licensing
expert valuation specialty 58–60 and competition 242, 253
and cumulative innovation 149–50
Fagerberg, J. 661 network effects 162, 166, 171, 199, 201–2,
Fagundes, D. 313, 316, 531, 546 203, 206, 208–10, 214, 215–16, 220
fair use doctrine notice and disclosure 428, 448
common law 513–14, 516, 520 see also licensing
and copyright 60, 144, 146–7, 187, 456, 458, Frank, J. 515
459, 531–2, 534–5, 536, 545 Frank, Richard 669
patents 573 Frank, Robert 396
Faller, G. 569 Frater, P. 624
Falvey, R. 233 free software movement (General Public
Farber, D. 385 License) 184–5, 188, 200–201, 343
Farley, C. 648 free-riding issues 169, 173, 295, 359, 563, 564,
Farrell, J. 163, 164, 168, 483, 530 565, 566
Fauchart, E. 82, 313, 316, 530, 546, 547 Freeman, A. 275
Faulhaber, G. 273 French, S. 12, 447
federalism and common law 521–3 Frey, B. 103, 110, 394
fee shifting proposal, and notice and disclosure Freyfogle, E. 56
458 Friedman, D. 57, 126, 291, 520
Feenstra, R. 269 Friedman, M. 266
Feit, I. 146 Friedman, O. 7, 89

Index 685
Frischmann, B. 56, 61, 66, 104, 340, 343, 378, Ghosh, R. 341
380, 381, 385, 397, 551–80 Ghosh, S. 4, 5, 6, 295, 389, 564, 636–55
Fromer, J. 10, 49, 74, 336, 427, 483 Giblin, R. 415
Fu, W. 586 Gibson, J. 10, 381, 382, 412, 456, 535, 539
functionality doctrine 193, 194–5, 204, 217–18 Gilbert, D. 109
future research Gilbert, R. 35, 130, 162, 212, 213, 214, 233,
anticommons, transaction costs, and patent 248, 271, 390
aggregators 31 Gilson, R. 333, 335, 641
common law 523, 524 Ginarte, C. 232, 659, 663
creative development and copyright 629 Ginsburg, J. 10, 414, 455, 504
custom role 547 Gladwell, M. 87
economic development 647, 652, 653, 668, Glovsky, S. 41
672 Godin, B. 568–9
enforcement costs 419 Goeschl, T. 668–9
negative space of IP (creativity not protected Goettler, R. 273
by copyright) 313, 326–7 Gogswell, L. 440
network effects 221 Gold, E. 30
patent institutions and legal actors 485–6 Goldberg, M. 294
philosophical foundations of IP law 94 Golden, J. 58, 59, 166, 395, 479, 484
price discrimination 302 Goldstein, A. 201
product differentiation 276–7 Goldstein, P. 122, 191–2, 287, 290, 595, 614
welfare promotion 111, 115 Gordon, W. 57, 147, 288, 291, 456, 562
Gould, D. 232
Gabriella, J. 295 Gould, J. 410
Gadde, S. 480 governance of intellectual property 47–70
Galanter, M. 412 collective-rights organizations 57, 58, 66
Galasso, A. 28, 137 commercialization effects 51–2, 58, 59, 60,
Gale, I. 121 61
Galetovic, A. 166 exclusion strategies 50–53, 57, 59–61, 63–4,
Gallagher, W. 418 65–6
Galle, B. 356 fluid property 55–7
Gallini, N. 122, 129, 131, 158, 271, 293, 295, governance through equity 64–5
334, 390 group institutions 65–6
Gallo, A. 379, 387 hybrid regimes 55–6
game theory model, collective management in rem and in personam rights, distinction
496–7 between 61–2
Gana, R. 586 information costs 57, 61, 62
Gandal, N. 159, 181, 195 infrastructure theory 559
Gans, J. 335, 338 and law of nuisance 53, 54, 56–7
Ganti, T. 590 licensing 61–4
Garcia, K. 622 licensing, compulsory 57, 59, 60
Gardiner, B. 613 modularity of property law 52–4
Gardner, E. 456, 457 opportunism 64, 65
Garon, J. 313 patent trolls 64–5
Gates, B. 173 property as platform 48–55
Gault, F. 314 semicommons 56, 66
Geistfeld, M. 411 thing-definition 48–55
gene sequence patents 143, 477–8, 479 trademark law definition of rights 50
General Public Licence (free software governance of intellectual property, patent
movement) 184–5, 188, 200–201, 343 versus copyright 58–61
genericide doctrine 147, 170, 194, 418, 555 activities as uses 58
Geradin, D. 166 derivative works 59–60
Gervais, D. 29, 66, 415, 489–507, 669 expert valuation specialty 58–60
Ghafele, R. 338 fair use doctrine 60
Ghosh, A. 552 form of property 58

individual uses 59 Heald, P. 51, 81, 334, 389, 644

licensing 60 Hegde, D. 82
governmental purchases and tax credits 372–3 Heins, M. 529–30
Grad, B. 212 Heller, M. 27–8, 29, 30, 38, 66, 143, 341, 371,
Grady, M. 409 417, 431, 438
Graham, S. 81, 135, 202 Helpman, E. 232, 552, 648, 658
Grasmick, H. 415 Hemel, D. 103, 295, 369–70, 371, 373, 378, 566,
Green, D. 415 601
Green, J. 32, 41, 124, 125, 137, 140, 141, 567 Hemphill, C. 82, 111, 313, 418
Greenberg, B. 394 Henkel, J. 339
Greene, K. 418 Henry, M. 477
Greene, P. 626 Herder, M. 31
Grennan, M. 293 Herrell, J. 207
Griffin, J. 105 Hess, C. 553, 557
Griliches, Z. 637, 668 Heverly, R. 56, 559, 571
Grill, B. 299 Hicks, D. 126
Grinvald, L. 408, 418 Hicks, J. 121
Grossman, G. 232, 658, 660 Higgins, R. 295
group institutions, and governance of Hill, P. 54, 413
intellectual property 65–6 Hilsenteger, J. 443
Gruben, W. 232 Hoekman, B. 666
Grushcow, J. 124 Hofmann, J. 29
Grynberg, M. 418 Hogendorn, C. 552, 565
Guell, R. 361, 363 Holbrook, T. 8, 9, 49, 483
Gupta, K. 166 Hollander, A. 499
Hollis, A. 293, 295, 358
Haber, E. 411, 412 Holmes, O. 85–6, 110, 510, 512
Haddock, D. 54 Holmes, T. 232, 287
Haemmerli, A. 78 Hoofnagle, C. 291
Hagelin, T. 397 Hopenhayn, H. 361
Hagiu, A. 35 Hotelling, H. 263, 267, 268, 271
Hall, B. 211, 382, 408, 413, 478 Hovenkamp, E. 242, 247, 252
Hammond, R. 623 Hovenkamp, H. 206, 211, 231–61, 287, 300, 447
Handke, C. 495, 498 Howe, J. 337
Hansen, D. 16, 19, 33 Howitt, P. 232
Hansen, R. 298 Howse, D. 353
Hansmann, H. 594 Hrdy, C. 369, 373
Hanson, R. 372 Hsee, C. 110
Hanssen, F. 300 Hu, A. 666, 667
Hardin, G. 4, 27 Huang, K. 31, 32
Harding, S. 19 Hubbard, T. 356, 358, 369, 381
Hardy, T. 290, 414 Hudson, T. 462
Hargreaves, I. 449 Hughes, J. 3, 18, 19, 22, 101, 147, 287, 410,
Harhoff, D. 340 462, 622
Harsanyi, J. 73 Hunt, R. 137, 142, 323
Hart, H. 474 Hunter, D. 29
Hashmi, A. 232 Hurst, W. 640
Haskell, J. 409, 411 Hurt, R. 411
Hastie, R. 110 hybrid regimes, and governance of intellectual
Hauser, M. 87 property 55–6
Hausman, J. 293 Hylton, K. 100, 299
Hay, D. 269 Hynes, R. 254
Haynes, J. 587
Hazard, G. 411 imitator’s entry to proprietary market, and
Hazucha, B. 615, 627 innovation 131, 132, 138

Index 687
independent invention, and innovation 131–3 supply-side problems 562–6

India user innovation 563
creative development see under creative Inglesi-Lotz, R. 661
development and copyright, emerging injunctions 39–40, 134, 135, 201–2, 448
industries in Global South see also damages
patents and drug prices 669–70, 671 innovation
individual preference satisfaction, welfare ambiguous language use and future
promotion 102–4, 108–9, 110, 111, 114 innovation risk 41
individual works’ valuation, collective and competition 129, 233–4, 237
management 498–501 and economic development 661–5
industry practices, and custom role 529–33, investment and network effects 160
538–41 and prior art 32, 36, 122, 124–5, 143, 149,
information 196–7, 198
ambiguous language use and future and prize and reward alternatives 355, 365–7
innovation risk 41 spillovers 414
asymmetries 28, 379–80, 392, 427–8 and welfare promotion 100–101, 102, 103,
costs 14–17, 40–41, 42, 57, 61, 62 104–5, 109, 114
sharing challenges, and open innovation innovation, cumulative 137–50, 361
331–3, 335 blocking rights 141–2, 148
see also knowledge; language ‘business method’ patent 143
infrastructure theory 551–80 channeling doctrines 150
abstraction levels 558, 559, 560, 561, 573 compulsory licensing 145, 146, 148
commons management 556–7, 561–2 copying 138, 144, 146
cultural environment as infrastructure copyright law 144, 145, 148
560–62, 567, 570 copyright law, fair use doctrine 144, 146–7
derivative works 563, 564, 566, 571–2 Creative Commons 149
exclusionary practices 562–6, 570 damages for past infringement 149
exclusivity rights 564, 567, 573 derivative works 141, 144
first- and second-generation inventors, disclosure requirements 144–5
relationship between 567 enhanced damages in patent law 149
genericness 555 evolving doctrines regarding subject matter
governance rules 559 143–4
in-system behavior 553 exclusive rights 140, 143, 144, 146, 148
input-output relationships 567–8 and follow-on innovation 137, 140–41,
intellectual-cultural resources 557–73 144–5, 147–8, 149
lead-time advantage 563, 566 FRAND licensing 149–50
linear model 568–9 ‘inventive application’ eligibility threshold
and market power 554–6, 565–6 143, 144
misappropriation risks 570 licensing failures 141
nonrival resources 554, 564–6, 567, 570 licensing and transaction costs 140, 143, 147
patent fair use doctrine 573 licensing virtues 138–9
private demand and social demand, lost royalty and damages for past
difference between 555–6 infringement 149
resources characteristics 553–4 network effects and technology sharing 148,
semi-commons regime 559–60, 571–3 150
sharable nature of resources 554 patent life and renewal rates 139–40
social returns versus private returns 553–4 prior art and threshold for protection 143,
spillover effects 561, 568, 569, 572 149
transaction costs 564, 570 profit shares 141
infrastructure theory, economic characteristics prospect patents 142
562–71 protection duration 139–40
demand-measurement problems 565 public interest exemptions 148
free riding issues 563, 564, 565, 566 remedies 149–50
patent law 564 reverse engineering 147
self-help mechanisms 563 and rights of others 145–50

and technical change 121 copyright and time limitations 18–19

threshold requirements and breadth 140–45 first possession theory 11
and trade secrets 147 information costs 14–17
and university research 145–6, 147 non-possessory property 12–14, 17, 19–21
innovation, stand-alone 122–36 orphan works and time limitations 18–19
breadth of intellectual property protection patent disclosure requirements 9–10
128–31 possession rationale 7–8, 9–10
channeling doctrines 122, 136 public domain works 11
commercialization effects 125, 128 recording and renewal requirements 20–21
compulsory licensing 135 standardization of property rights 15
copying 129, 133, 134 time dimension 17–22
copyright protection 122, 123, 129 trademarks as rivalrous resources 11–12
deadweight loss 101, 120, 123, 126, 130–31, intellectual property as property, tangible
142 objects distinction 4–7
design patent protection 123 distribution of physical resources 6–7
doctrine of equivalents 129 exclusionary practices 5–6
duration of protection 126–8 land titles and information mechanisms
exhaustion (first-sale) doctrine 133 16–17
expressive creativity protection 122 law of adverse possession 18
imitator’s entry to proprietary market 131, scarcity and rivalry 5, 11–12
132, 138 interbrand and intrabrand restraints, and
independent invention, rights arising from competition 245–6
131–3 interstitial common law 517–18, 519, 520–21
injunctions and damages 134, 135 ‘inventive application’ eligibility threshold 143,
licensing 122, 130, 131, 132, 135 144
patent law nonobviousness standard 124–6
patent law novelty requirement and first-to- Jackendoff, R. 88
invent system 124, 136 Jaffe, A. 219, 411, 484, 666
and price discrimination 134 James, S. 415
prior user right 131, 132, 133, 136, 447 Janis, M. 203, 204, 379, 476
and prior works 122 Japan, prize and reward alternatives to IP,
profitability of intellectual property right history of 354
127–8 Jasanoff, S. 287
remedies 134–6 Jaszi, P. 29, 529–30, 532, 533
rights of others 131–4 Jegan, R. 394
substantial similarity test 129 Jesien, K. 312
and technical change 121 Johnson, W. 134
threshold for protection 123–6 Jones, C. 552
trade secret law 123, 126, 132, 135–6 Joo, T. 623
unpatented inventions and independent Joyce, C. 388
inventors 132–3 Judd, K. 269
input-output relationships, infrastructure judge selection, randomization suggestion 483
theory 567–8 judicial tailoring, and uniformity cost
institutions reduction 391, 393
and economic development 641–5, 650–51, Justman, M. 552
652
patent see patent institutions and legal Kagan, S. 84
actors Kahn, A. 552
insurance pools and license on transfer (LOT) Kahneman, D. 103, 105, 108, 109
commitments 200 Kaldor, N. 264, 265, 267, 269
intellectual property as property 2–26 Kalil, T. 356, 367, 373
copyright and changed circumstances Kaminski, M. 594
concept 21–2 Kamm, F. 87
copyright as instrument of constraint 13 Kanngiesser, P. 90
copyright notice and registration 10–11 Kant, I. 78, 79, 86, 90

Index 689
Kapczynski, A. 6, 102, 105, 313, 364, 365, 372, Krackhardt, D. 560

594, 613, 614, 620 Krazit, T. 439
Kaplow, L. 73, 130, 134, 293, 296, 385, 386, Kremer, M. 359, 362–3, 364, 367, 371, 373,
409, 411 379
Karaganis, J. 604, 617 Kretschmer, M. 495
Karjala, D. 170, 381 Krings, M. 619, 628
Kastenmeier, R. 392 Krugman, P. 648
Katoh, M. 186 Ku, R. 411
Katsoulacos, Y. 266, 270 Kuhlik, B. 28, 29, 34, 35
Katz, A. 411 Kumar, S. 481
Katz, M. 161, 164, 168, 269, 288, 292 Kwall, R. 86, 101
Katz, R. 463 Kyle, M. 670–71
Keller, J. 552
Keller, W. 665, 668 La Manna, M. 132
Kelly, C. 353, 371 Laakmann, A. 393
Kelman, M. 386 Lai, E. 660, 661
Kennedy, D. 386 Lakhani, K. 337–8, 340
Kenney, R. 300 Lancaster, K. 267, 565, 570
Kenya, intellectual property law 649 Landes, W. 5, 10, 11, 19, 100, 126, 128, 132,
Kesan, J. 379, 387, 389, 483 138, 147, 173, 216, 275, 288, 295, 390, 395,
Khan, B. 363 409, 410, 411, 414, 496, 528, 564, 572
Khanna, D. 19 Lane, W. 269
Khanna, T. 335 Langlois, R. 52, 170
Khoury, A. 233 language, ambiguous language use and future
Kieff, F. 51, 125, 146, 332, 335, 368–9, 389, innovation risk 41
395, 397, 564 see also information
Kiesling, L. 54 Lanjouw, J. 139, 408, 413
Kim, Y. 76, 659 Lansing, P. 295
King, G. 310 LaPolt, D. 623
Kirby, S. 134, 272, 296, 301 Larkin, B. 619
Kitch, E. 58, 125, 126, 135, 142, 291, 334, 368, Lastowka, G. 313
380, 414, 564 Laursen, K. 339
Klein, B. 292, 298, 300 Lawson, J. 543
Klemperer, P. 129, 134, 164, 271 Laycock, D. 64
Klerman, D. 475, 481 Le, N. 566
Kline, D. 219 lead-time advantage 396, 563, 566, 603, 609
Kline, S. 469 Leahy, J. 395
Knight, F. 381 Lederman, L. 415
knowledge Lee, B. 49
commons, and open innovation 343–5 Lee, E. 113, 597
dissemination function, and notice and Lee, J.-A. 29
disclosure 426–7 Lee, P. 336, 480, 485, 560, 573
see also information Lee, T. 121, 128
Knudsen, C. 569 Lee, W. 65, 202, 440
Kobayashi, B. 300 Leeson, P. 286
Koch, J. 607 Lefstin, J. 144, 477
Koenigsberg, F. 504 Lei, Z. 31
Koenker, R. 266, 270 Leibovitz, J. 132
Komesar, N. 566 Lemley, M. 3–5, 9, 35–7, 39, 41, 53, 58, 61, 66,
Korman, B. 504 85, 88, 100, 108, 123–4, 135, 141, 143–4,
Korngold, G. 13 149, 162–3, 165–6, 199, 201, 203, 206,
Kornhauser, L. 313, 511 216, 235, 272–3, 275, 288, 311, 327, 331,
Kortum, S. 666 381, 383–97passim, 412, 414, 416–18, 436,
Kotowitz, Y. 390 439–40, 447, 474–84 passim, 530, 540, 552,
Kozinski, A. 455 553–7, 571, 573, 603–4, 622

Lemus, J. 240 Lobel, O. 570

Lenman, J. 81 lock-in effect, custom role 539, 542–3
Lerner, J. 185, 219, 249, 313, 342, 343, 395, Loeb, D. 88
411, 413, 484, 652, 662–3 Lohr, S. 397
Leslie, C. 369 Long, C. 10, 16, 59, 120, 335, 381, 383–4, 385,
Leslie, P. 290, 299, 300 389, 391, 412, 428, 478, 527
Lessig, L. 18, 378, 397, 553, 562, 567, 623 Loren, L. 16, 19, 29, 56, 559, 571
Leval, P. 57, 529 Loshin, J. 313
Levin, R. 382, 389, 395, 563 Loury, G. 121, 128
Levine, D. 234, 379, 417 Love, J. 356, 358, 369, 381, 382, 390
Levine, M. 287, 662 Lowery, D. 603
Levinsohn, J. 269 Lu, S. 628
Levinthal, D. 54, 332 Lueck, D. 7, 413, 431, 438
Levmore, S. 373 Lunney, G. 4, 272, 275, 290, 291, 292, 296–7,
Lewinsohn-Zamir, D. 104 300, 380, 381, 382, 385, 389, 392, 495, 499,
Liang, L. 589, 590, 591, 594, 605, 607, 608 564, 565
Libecap, G. 57, 413, 431, 438 Luo, H. 82
licensing Lyons, D. 77, 78
collective 458–9, 600–601
and collective management 491–3, 495 McAdams, R. 616
compulsory 57, 59, 60, 135, 148, 170, 253, McAfee, R. 289, 292
455 Macaulay, S. 313
cover licensing 455 Macaulay, T. 378, 594
and cumulative innovation 140–41 McCahery, J. 523
FRAND see FRAND licensing McCall, J. 587, 619
governance of intellectual property 57, 60, McCauley, S. 640
61–4 McClure, D. 275
innovation, cumulative 138–9 McDaniel, C. 659
innovation, stand-alone 122, 130, 131, 132, McDonnell, B. 334, 385
135 McDonough, J. 36
network effects 183–6, 199–201, 210–11, Macedo, C. 124
213 McFetridge, D. 390
and open innovation see open innovation Macfie, R. 354, 355
and ex ante licensing McGahan, A. 671
refusal 210–11, 252–3 McGeveran, W. 418, 532
and transaction costs, innovation, McGowan, D. 76, 83–4, 163, 165, 185, 396–7
cumulative innovation 140, 143, 147 Machlup, F. 83, 120, 379
voluntary 458–60 McJohn, S. 438
Lichtman, D. 29, 125, 132–3, 298, 357, 379, Mackaay, E. 504
409, 414, 418, 595–6 McKenna, M. 272, 273, 275, 453, 562, 566
Liebeler, L. 295 MacKie-Mason, J. 293
Lieberman, M. 396 McKusick, M. 184
Liebowitz, S. 133, 161, 300, 496, 604–5, 610 McLeod, K. 29, 34, 569
Liivak, O. 3, 4, 52, 566 McManis, C. 31, 32
Lim, D. 255 McManus, J. 54
Lipner, S. 295 Macmillan, F. 620
Lipsey, R. 265, 268, 270, 565, 570 MacNeil, I. 640
Litman, J. 10, 59, 311, 384, 411, 553, 562, 567, Madison, M. 48, 343, 344, 531, 546, 558, 560,
572, 623 561, 564, 568, 571
Litwak, M. 300 Magliocca, G 313
Liu, Jiarui 592, 599, 603, 607, 610, 611, 613, Mahoney, J. 20
614, 621, 624 Mahoney, P. 511
Liu, Joseph 10, 22, 147, 393 Mak, J. 137
Llewellyn, K. 510, 515 Malamud, C. 192
Lobato, R. 589, 609, 617, 619, 625 Malueg, D. 293

Index 691
mandatory registration, notice and disclosure, Merges, R. 4, 35, 52, 65, 66, 72–97, 101, 125,
copyright tracing and ownership 449, 452, 126, 135, 137, 141, 142–3, 146, 148, 185,
461 206, 212, 216, 233, 234, 291, 324, 331, 334,
Mandel, G. 643 335, 383, 384, 387, 398, 418, 448, 458, 460,
Maness, R. 211 483, 530, 553, 567, 596, 622
Mankiw, N. 270 Merrill, T. 7, 15, 17, 49, 51, 52, 57, 63, 431,
Manly, L. 542 446
Mansfield, E. 139, 382, 385, 396, 563, 644 Merz, J. 30, 34
Manta, I. 107, 287, 290 Meurer, M. 4, 10, 16, 36, 38, 40, 49, 59, 199,
mapping 281–308, 384, 385, 395, 396, 416, 417, 430,
economic development 645–52 438, 439, 442, 445, 446, 447, 448, 451
negative space of IP 313–24 Miceli, T. 411
notice and disclosure, patents 445–6 Mikhail, J. 86, 88–9, 90
Margolis, S. 133, 161, 300 Milgrom, P. 418
market power Mill, J. 364
and competition 233, 235–6, 241–3, 244 Millaleo, S. 649
and creative development, cultural diversity Miller, Jade 586, 587, 588, 596, 598, 600, 606,
621, 627 608, 609, 616, 619
custom role 534–5 Miller, Joseph 65, 66, 443
and infrastructure theory 554–6, 565–6 Miller, S. 418
negative space of IP 322–4 Miller, Z. 569
and network effects 161–5, 212, 214–16 Mireles, M. 395
prize and reward alternatives 362–3, 364, Mnookin, R. 313
366–7 Mock, D. 195
and product differentiation 269–70, Moglen, E. 601, 620
272–3 Mojibi, A. 233
Markoff, J. 212 Mokyr, J. 121, 637
Markusen, J. 666 monopoly power 171, 263–6, 291, 295,
Marshall, R. 496 354–5
Maskin, E. 137, 139–40, 289 Monroe, B. 464
Maskus, K. 76, 411, 643, 645–6, 647, Montgomery, D. 396
656–75 Montgomery, L. 592, 593, 603, 611
Massarsky, B. 493 Moore, K. 135, 389, 439, 446, 475, 477, 478,
Masur, J. 98–118, 477, 478, 484 481, 482
Matal, J. 481 Morris, C. 528
Mattioli, M. 195, 448 Mortimer, J. 287, 290
Matutes, C. 271 Moschini, G. 357, 362
Maurer, S. 32, 111, 121, 131, 132, 272 Moser, J. 456
Max, D. 530 Moser, P. 232, 354, 659, 663
Mazzeo, M. 417 Moskowitz, N. 219
Mazzone, J. 453 Mossinghoff, G. 124
Meagher, K. 552 Mossoff, A. 3, 53, 234
Medina, B. 84 Moullier, B. 590
Meeker, H. 162, 185, 188, 200 Mowery, D. 31, 395
Meese, A. 299 Mueller, C. 78
Melamed, D. 6, 35, 36, 37, 65, 134, 201, 202, Mueller, D. 78
417, 418 Mueller, J. 39, 146, 206
Melendez-Ortiz, R. 669 Murray, F. 29, 31, 32, 314, 355
Melichar, F. 499 Musgrave, R. and P. 554, 555
Mencken, J. 312 music industry
Menell, P. 4, 5, 6, 10, 16, 42, 49, 59, 75, 102, compulsory license for cover versions 148
119–230, 232, 233, 234, 236, 251, 330, emerging industries see under creative
331, 378, 395, 409, 411, 416, 424–71, 476, development and copyright, emerging
623 industries in Global South
Mensch, E. 275 negative space of IP 321, 325–6

Nachbar, T. 5 FRAND licensing 162, 166, 171, 199, 201–2,

Nadel, M. 619, 621, 622, 624 203, 206, 208–10, 214, 215–16, 220
Nadler, J. 415 free-rider problem 169, 173
Nalebuff, B. 248, 286, 299, 300 functionality and creative expression,
Nanda, V. 324 distinction between 217–18
Napoli, P. 619 future research 221
Nard, C. 474, 477, 479, 480, 482, 484 and market power 161–5, 212, 214–16
Neary, K. 7, 89 monopoly power issues 171
negative space of IP (creativity not protected network externality dilemma 168–9
by copyright) 309–29 open innovation and ex ante licensing 333,
comedians 314–15, 325 338, 339, 340
computer databases and first-mover open source software 162–3, 171, 216
advantage 317–18 parsimony principle 168–9, 217–19
copying 310–11, 314–15, 317–18, 320, 321–2, patent bargaining model 166
324–5 proportionality principle 169–70, 219–20
fashion industry 311–12, 321–2, 327 protectionist entrepreneurs 216
financial service industry 322–4 real and virtual networks 161–2
first-mover advantage 317–19 remedies against abusive and anti-
food industry 316–17, 320 competitive behavior 171
future research 313, 326–7 royalty stacking 166
industrial design 326–7 standard setting organizations (SSOs) 162,
legal consequences, unintended 324–6 166, 171
mapping 313–24 standard-essential patents (SEPs) 166, 171,
market power unrelated to IP 322–4 199, 201, 202–3
music business 321, 325–6 standardization 163, 164–5, 167–8
philosophical foundations of IP law 82 and technological advances 159–60, 169–70,
positive space comparison 311 171, 218
products versus performances 319–21 and technology sharing 148, 150
publishing trade 315–16 trade secret protection in software industry
social norms, role of 314–17 172–3
Nelson, R. 120, 121, 122, 137, 141, 146, 233, uniformity cost reduction 396–7
331, 383, 387, 398, 553, 567 network effects, competition policy 167–71,
neoclassical economic growth model 638–9, 205–16
641, 644 ‘essential facilities’ doctrine 210
Netanel, N. 134, 135, 287, 290, 585, 595, 596, exhaustion (first-sale) doctrine 207
600, 618, 621, 622 licensing guidelines 213
network effects 157–230 patent misuse doctrines 205–7
application program interface (API) patent thickets 211–12
protection 218–19 private antitrust liability 210–12
appropriability problem 166, 167, 173, 188, public enforcement 212–16
211, 218, 220 refusals to license 210–11
business method patents 197–8, 201 ‘rule of reason’ approach to patent licensing
complementary products 159–60, 162, 164, 213–14
168, 183, 194 standard setting processes, ambush of 208
compulsory licensing of patents 170 network effects, copyright protection in
and consumer demand 159–60, 161, 167 software industry 173–92, 218
copying 176, 178, 179, 183, 189–90, 193–4 abstraction-filtration-comparison test 177,
demand-side effects 167–8 178
derivative works 162, 163, 184 application program interface (API)
deterrence principle 171, 220–21 protection 160, 162–3, 167, 172, 175–6,
economic and social value 159, 161–2, 179–80, 189–91
163 command systems for computer software
exclusionary practices 199, 201, 202, 209, 180–82
211 copyleft licensing model 184–5, 343
exclusivity rights 183, 195, 210, 211 fair use doctrine 187

Index 693
free software movement (General Public New Zealand, Phonographic Performances

License) 184–5, 188, 343 (NZ) Ltd v. Radioworks Limited 507
idea/expression doctrine 173–4, 175, 177 Newell, R. 121
mobile platform development 189 Newiak, M. 76
open software movement (permissive Newman, C. 49, 61, 455
licenses) 185 Newman, J. 237
piracy protection 173, 175 Newson, M. 88
public domain dedication 185–6 Nicholas, A. 354
reverse engineering permissibility 182–3, Nicol, D. 30, 33, 34
186, 188 Nielsen, J. 30, 33, 34
software copyright jurisprudence 188–91 Nigeria, creative development see under
software licensing 183–6 creative development and copyright,
standards and codes 191–2 emerging industries in Global South
start-up phase, adoptions over revenues Nimmer, D. 453
strategy 188–9 ‘nine no nos’ of patenting 213, 237
trade secrets 182–3 Nock, J. 480
unprotectability of functional and network non-commercial creativity 601–2
features 176–82 non-possessory property 12–14, 17, 19–21
network effects, patent protection for software- non-practicing entities (NPEs) (patent trolls)
related technologies 195–205, 219–20 36, 37, 64–5, 83, 200, 416–18, 417–18
claim indefiniteness doctrine 199 nondisclosure agreements (NDAs) 428, 440,
design patents 203–5, 218–19 465
dot-com bubble burst effects 197, 201, 220 nonobviousness requirement 124–6, 198, 234,
FRAND licensing 199, 201–2, 203, 206 236–7, 366, 476, 478–9, 480
free software movement (General Public nonrivalness 57, 432, 554, 564–6, 567, 570
License) 200–201 Nordhaus, W. 126–7, 158, 271, 390
injunctive relief as remedy 201–2 North, D. 560
insurance pools and license on transfer notice and disclosure 424–71
(LOT) commitments 200 and copying 428–9, 437, 447, 461
licensing 199–201 design patents 462–5
machine-or-transformation test 197 FRAND licensing 428
non-practicing entities 200 freedom to operate function 428–9
nonobviousness requirement 198 informational asymmetries 427–8
patent holdup prevention and pre- intangible resources, costs of describing and
commitment strategy 200 disclosing 432, 436–8
patentability requirements 196–8 knowledge dissemination function 426–7
and prior art 196–7, 198 nondisclosure agreements (NDAs) 428, 440,
remedies 201–3, 204–5 465
royalty rates 202 nonrivalrous nature of intangibles 432
standard setting organizations (SSO) 199, notice costs 430–33
201–3, 206 notice externalities 438–9
standard-essential patents (SEPs) 166, 171, notice functions 426–9
199, 201, 202–3 orphan works and copyright 439
start-up businesses and patent portfolios 197 patent protection 426–7, 434, 436, 438–9
subject matter eligibility 196–8 patent thickets 438
network effects, trademark protection 192–5 remedies 440
and confusing product names 194 signaling function 427–8
functionality doctrine 193, 194–5, 204 standard setting organizations (SSOs) 428
genericide doctrine 194 trade secrets 427, 440, 465
platform sponsors and complementary trademarks 435, 465
product manufacturers 194 see also disclosure requirements
and standard setting organizations (SSOs) notice and disclosure, copyright tracing and
194 ownership 428–9, 434–8, 449–62
and unfair competition 194 collective licenses 458–9
Neven, D. 269 compulsory licensing 455

cover licensing 455 Oliar, D. 4, 8, 10, 82, 99, 313, 314, 315, 394,
Creative Commons 460 530–31, 547
cumulative creativity 429, 440, 457–60 Olson, D. 287, 290
depropertization 460–62 Olson, K. 90
digital registration proposal 451, 461 Olson, T. 384
fair use doctrine 456, 458, 459 O’Neil, O. 86
fee shifting proposal 458 Opara, J. 615
liability standards 456–7 Opderbeck, D. 560
mandatory registration 449, 452, 461 open innovation and ex ante licensing 330–49
open source projects 460 appropriability over free revealing 338–9,
orphan works 449 341
preclearance of works 453, 458 bilateral (ex ante) license 333–6
preregistration system 461 collaborative innovation 333–45
reforms 454–8 collective action problems 344
remedies 456–8 commercialization effects 331–2, 336, 337–8
standardized unique identification number ‘coupled’ open innovation 337–8
systems 451–2 Creative Commons 343
statutory damages 457 and decentralization 338, 341
technology advances 449–52, 455, 459 derivative works 343
trade secrets 461 exclusivity rights 331, 334, 335–6, 338,
unpublished works, protection of 461–2 344–5
voluntary licensing 458–60 first mover advantage 338–9
notice and disclosure, patents 441–8 General Public Licence (free software
adverse possession doctrine 18, 22, 431, movement) 184–5, 188, 200–201, 343
446–7 information sharing challenges 331–3, 335
application and maintenance fees 446 inside-out and outside-in implementation
claim clarity and standardized claiming 337, 338
system, call for 442–3 knowledge commons 343–5
claim construction and judicial evaluation network effects 333, 338, 339, 340
443–4 patent thickets, effects of 341
depropertization 448 peer production 341–3
disclosure timing 442 and reward theory of intellectual property
doctrine of equivalents 447 331–2
FRAND licensing 448 ‘sticky’ information 332–3, 335, 342
independent invention defense 447 supply chains 334–5
injunctive relief and damages awards technology transfer 331–3, 334–5, 338, 342
448 transaction costs 332–3, 334–5, 342
mapping and search tools 445–6 underproduction problem 330–31, 343–4
notice failure and harm reduction 447 user innovation 339–41, 344–5
ownership transparency 443 open source 148, 149–50, 162–3, 171, 185, 216,
remedies 447–8 460
novelty requirement 124, 136 see also commons
Novos, I. 134, 301 opportunism 64, 65, 411
nuisance law 53, 54, 56–7 Ordover, J. 134
Nunnenkamp, P. 667 O’Rourke, M. 146, 291, 393, 573
Nussbaum, M. 104, 637, 651–2 orphan works 18–19, 439, 449
Osenga, K. 395
Ochoa, O. 567 Ostrom, E. 47, 56, 57, 65–6, 250, 343–4, 528,
Ochoa, T. 235 531, 553, 557, 560
O’Connor, S. 99 Oswald, A. 108
Odagiri, H. 657, 659, 668 Ottoz, E. 132
Oddi, A. 125 Ouellette, L. 103, 295, 336, 369–70, 371, 373,
O’Donoghue, T. 121, 137, 138, 139, 141, 142, 378, 427, 566, 601
271, 389 overcompensation issues, prize and reward
Okome, O. 619, 628 alternatives 360–61

Index 695
Pager, S. 582–635 ‘business method’ patent, innovation,

Pakes, A. 139 cumulative innovation 143
Pallante, M. 21 design patents 123, 203–5, 218–19, 462–5
parallel imports 647–8 disclosure requirements, intellectual property
Parchomovsky, G. 5, 17, 36, 37, 125, 211, 334, as property 9–10
335, 389, 390, 391, 397, 412, 432, 446 doctrine of equivalents 41, 129, 238, 447,
Parente, M. 29 521
Parisi, F. 28, 29, 74, 291, 500–501 and economic development 661–5, 666–7,
Park, W. 232, 659, 663, 664 669–71
parsimony principle, network effects 168–9, and enforcement costs 408, 416–18
217–19 fair use doctrine 573
Partnoy, F. 389 and infrastructure theory 564
patent institutions and legal actors 473–88 innovation and incentive-to-invent theory
accuracy evaluation 483–4 81–2
administrative review 484 innovation, stand-alone innovation 124,
Article III trial courts 481, 483, 486 124–6, 132–3
claim construction 476–7, 482, 483 life and renewal rates, innovation, cumulative
Congressional involvement 474, 480–81, innovation 139–40
484 misuse doctrine 238, 247, 255–6
district courts 40, 434, 444, 456, 474, 475, network effects 200, 205–7
476, 481 ‘nine no nos’ 213, 237
en banc decision making 473, 479, 481, 484 nonobviousness requirement 124–6, 198,
Federal Circuit 474, 475–9, 480, 481, 482–4 234, 236–7, 366, 476, 478–9, 480
future research 485–6 notice and disclosure see notice and
gene sequence patents 143, 477–8, 479 disclosure, patents
ideology role 485 novelty requirement 124, 136
judge selection randomization suggestion and price discrimination 292–3
483 product differentiation 271–2
legal devolution 474–5, 483 prospect patents 142
non-obviousness determination 476, 478–9, ‘rule of reason’ approach to patent licensing
480 213–14
patent application increase 478 scope-of-the-patent test 238–40
patent clustering 481 standard-essential patents (SEPs) 166, 171,
patent denial reversals 478 199, 201, 202–3, 208–10, 214, 215–16,
prior case law effects 485 242
pro-patent bias 477, 484 uniformity cost reduction 381, 386–8, 389,
specialized trial court suggestion 482–3 390
Supreme Court 39, 465, 474–5, 476, 477, university patents 39, 145–6, 147
479–80, 483, 484–5 versus copyright governance see governance
teaching, suggesting, and motivation (TSM) of intellectual property, patent versus
approach 479–80 copyright
USPTO 474–5, 477–8, 480, 481, 483, 484 welfare promotion 107–9, 112
USPTO Patent Trial and Appeal Board Patry, W. 235, 384, 390, 391
(PTAB) 474, 481, 484, 485–6 Patterson, M. 208
validity presumption for issued patents 476 Paulson, C. 588
venue restriction suggestion 483 pay-for-delay settlements, competition and
patent pools 248–51 intellectual property 239
patent thickets 211–12, 341, 371, 438 peer production, and open innovation
patent trolls (non-practicing entities (NPEs)) 341–3
36, 37, 64–5, 83, 200, 416–18, 417–18 Peñalver, E. 3, 4
patents and patent law Penner, J. 49, 61
aggregators and anticommons see Penrose, E. 120
anticommons, transaction costs, and Penubarti, M. 659
patent aggregators Perel, M. 410
bargaining model, network effects 166 Perry, M. 266, 270

‘person having ordinary skill in the art’ population quality measures, economic
(PHOSITA) 387, 393 development 649, 651–2
Perullo, A. 649 Port, K. 408, 418
Perzanowski, A. 14, 17, 82, 207, 291, 313, 315, Posner, R. 5, 10, 11, 19, 74, 84–6, 88, 90, 100,
411, 546, 610 102, 104, 106, 109, 115, 126, 128, 132, 138,
Pessach, G. 621, 622 147, 173, 216, 275, 288, 293, 295, 299, 313,
Peterman, B. 295 386, 390, 395, 407, 409–11, 496, 510–12,
Petherbridge, L. 479, 482 515–16, 518, 528, 539, 541, 572
Phelps, M. 219 Pound, R. 510
Philips, L. 295 Pouris, A. 661
Phillips, D. 183, 185, 186 Powdthavee, N. 108
philosophical foundations of IP law 72–97 Powell, W. 333
consequentialism 73–84 Prado, M. 642
consequentialism, macro theory 75–6, 77–9, Pratt, L. 588
83–4 preferential trade agreements 660–61
consequentialism, morally reprehensible Prescott, E. 269
objection 79–80 Pressman, D. 438
consequentialism, and personal preferences Price, W. 121
78–9 price discrimination 281–308
consequentialism, productivity and allocative efficiency 288, 290
innovation investment relationship 75–6 complementary products 297–300
deontology 74, 79, 84–90 and copying 291, 298–9
deontology and consequentialism, digital technologies 286–7
reconciling 91–3 distributional effects 289
deontology, consistent moral judgments economic development 647
across cultures 87–8 exclusivity rights 285, 295, 297, 299–300
deontology, moral duties 84–6 exhaustion principle 294
deontology, moral judgment studies 86–90 future research 302
deontology, and universal intellectual geographic 293–5
property instinct 90 monopoly power 291, 295
efficiency principle 92, 93 patent exhaustion doctrine 294–5
future research 94 patents and social welfare 292–3
‘negative space’ industries (absence of IP) 82 product differentiation 284
non-practicing entities (patent trolls) and profitability 285, 286–7
small businesses 83 second and third degree, differences between
patents, innovation and incentive-to-invent 285, 288
theory 81–2 social welfare issues 287–93
patents and small businesses 83 stand-alone innovation 134
proportionality principle 92 tie-in metering benefits 299
utilitarianism 73–84 trademark free-riding 295
welfare economics 73–4, 99–101 trademarks, identical 295
Pigou, A. 283 tying, merchandising, and bundling 285–6,
Piller, F. 337 297–300
Pink, D. 563 uniformity cost reduction 385
Pinker, S. 88 and willingness to pay 282–3
piracy protection 173, 175, 291 price discrimination, copyright
Pitofsky, R. 206, 215 and cumulative creation 291
Plant, A. 398–9, 411 and delivery date differentiation 284
Png, I. 134 ‘first sale’ doctrine 294
Pogge, T. 358 and free access loss 292
Polanvyi, M. 354, 355, 379 and incentives for creative activity 290–91,
Polanyi, M. 333 292
Polinsky, A. 134, 135 and piracy 291
political economy 231–5, 384 restrictions on type of use 296
Pollack, M. 312 sharing 301–2

Index 697
and social welfare 290–92 valuation of contributions 360–64

and transaction costs reduction 291 valuation mechanism 351–2, 360–64
and unauthorized use 297 variable versus fixed payments 357–8
price-fixing 236, 238–9, 241, 243–4 pro-patent bias 477, 484
pricing producer-side effects, welfare promotion
and collective management 495, 501 112–15
real option pricing, tailoring through 388–91 product differentiation 262–80
Priest, E. 583, 592, 593, 599, 600, 603, 605, access and incentives, relationship between
606, 607, 608, 610, 611, 616, 617, 624, 626 269
Priest, G. 411, 511 appropriability problems 270, 271
prior art 32, 36, 122, 124–5, 143, 149, 196–7, and competition 242, 263–9, 272–4
198 complementary products 269, 285–6
prior user right, innovation, stand-alone 131, consumer preferences 266–8, 269, 271
132, 133, 136, 447 copyright 272–4
private antitrust liability, network effects, future research 276–7
competition policy 210–12 and market power assessment 269–70, 272–3
private demand and social demand, difference monopolistic competition 263–6
between 555–6 patents 271–2
private enforcement as public good 411 price discrimination 284
private interest, prize and reward alternatives spatial competition 266–9, 272–4
355, 367, 371 trademarks 274–5
private ordering 385–6, 595–6 trademarks, free imitation 274
private sector, and custom role 540–41, uniformity cost reduction 397
546 profitability 127–8, 141, 285, 286–7
prize and reward alternatives 350–75 Profitt, B. 218
administrative design issues 356–60, proportionality principle 92, 169–70, 219–20,
362 499–500
cash versus other incentives 356–7 prospect patents 142
commercialization, effects on 368–9 protectionist entrepreneurs, and network
complementary invention 368 effects 216
consumers, effects on 369–70 Pruitt, S. 212
cumulative invention 361 public enforcement, network effects,
deadweight loss 355, 357, 358, 364, 367, competition policy 212–16
369–70, 372 public fund investment in innovation, risks in
developing countries 360 379–80
fixed ‘pot’ suggestion 358, 359, 367 publicly available information about patent
free-riding incentives 359 ownership 41
governmental purchases and tax credits Puttitanun, T. 659, 664
372–3
innovation, effects on 355, 365–7 Qian, Y. 664–5, 670–71
international coordination 359–60
market mechanisms 362–3, 364, 366–7 Radin, M. 6, 386
monopolization concerns 354–5 Rafiquzzaman, M. 390
overcompensation issues 360–61 Rai, A. 38, 107, 293, 395, 473–88
patent nonobviousness 366 Rajadhyaksha, A. 589, 600, 626
and patent thickets 371 Rajec, S. 384
private interest 355, 367, 371 Ramsey, L. 418
redundant innovation, effects on 367–8 Rantanen, J. 41, 233, 480
social value 360–61, 365, 366, 367 Raskind, L. 290
timing of payment 356 Ratner, J. 210
traditional property systems comparison Rato, M. 166
364–73 Raustiala, K. 82, 111, 234, 309–29, 330, 394,
transaction costs 370–71 530, 546, 610
and TRIPS Agreement 360 Rawls, J. 84–5, 86, 90, 93
uniformity cost reduction 379, 380, 381 Raymond, E. 336

redundant innovation, and prize and reward Romer, P. 232, 552, 567, 639, 662
alternatives 367–8 Rooklidge, W. 476
Reese, A. 19, 416 Rose, C. 5, 6, 7, 10, 11, 12, 13, 14, 20, 47, 56,
refusal to license 210–11, 252–3 66, 386, 528, 542, 552–3
Reichman, J. 341, 411, 643, 669 Rose-Ackerman, S. 411
Reidel, I. 613 Rosen, S. 396
Reidenberg, J. 83 Rosen, W. 662
Reiffen, D. 669 Rosenberg, D. 410, 411
Reilly, G. 475, 481 Rosenberg, N. 137, 569
Reinganum, J. 121, 128, 410 Rosenblatt, E. 82, 310, 394
remedies Rosenbluth, G. 269
innovation, cumulative innovation 149–50 Rothman, J. 16, 412, 526–50
innovation, stand-alone innovation 134–6 Rowe, E. 395
network effects 171 Rowley, C. 73
notice and disclosure 440, 447–8, 456–8 Royall, S. 208
Remington, M. 392 royalties 149, 166, 202
repertory price setting, collective management Rubin, P. 295, 511, 512–13, 518
492, 495–8 Rubinfeld, D. 211
reverse engineering, cumulative innovation ‘rule of reason’ approach to patent licensing
147 213–14
reward alternatives see prize and reward Rumelt, R. 134, 301
alternatives Russell, A. 165
reward theory, and open innovation Rutschman, A. 589
331–2 Ruttan, V. 121
rewards distribution issues, uniformity cost
reduction 382–3 Sabel, C. 640–41
Rhodes, D. 608 Sachs, J. 638, 650
Rice, J. 148, 200, 448 Sacks, A. 474
Rich, G. 51, 443 Saeb, S. 109
Ridgway, W. 418 Sag, M. 409, 411, 416
Ridley, D. 356 Salinger, M. 299
Rierson, S. 418 Salkever, D. 669
Riley, J. 289 Saloner, G. 163, 164, 168
Risch, M. 36, 417, 418, 438 Salop, S. 268, 269, 270, 272–3, 418
risk mitigation changes, anticommons see Salter, A. 339
anticommons, transaction costs, and Sammi, P. 147, 178
patent aggregators, risk mitigation Sampat, B. 108, 389
changes Samuelson, P. 10, 29, 147, 170, 173, 183, 266,
risk-reduction functions, creative development 273, 274, 388, 395, 415, 416, 438, 553, 554,
and copyright 595–6 562, 564, 565
risk-spreading justification and costs of failure, Sanchez-Roig, R. 312
uniformity cost reduction 380 Sander, M. 42
Rivette, K. 219 Sandler, T. 554
Rizk, N. 603, 613 Santilli, M. 594
Roberts, J. 418 Santore, R. 248
Roberts, P. 212 Santos, A. 605, 608, 609, 627, 628, 640
Roberts, R. 298 Sarnoff, J. 238, 384, 387, 566
Robinson, G. 17, 61 Saulo, E. 381
Robinson, J. 263, 288, 638 Saunders, K. 211
Robinson, K. 448 Saunders, M. 40
Robinson, W. 396 Sawicki, A. 595
Rochelandet, F. 502–3 Sawyer, M. 459
Roffe, P. 669 Saxenian, A. 333
Roin, B. 300, 360, 362, 365, 368, 369, 372, Scanlon, T. 86
379 Scaria, A. 590, 591, 610, 652

Index 699
Schallnau, J. 408 Shapley, L. 496–7

Schankerman, M. 135, 137, 139, 149, 382, 408, Shavell, S. 73, 121, 134, 135, 358, 361,
413, 652 367, 379, 380, 394, 395, 410,
Schatz, T. 569 411
Schauer, F. 386, 515 Shaver, L. 380
Scheffler, S. 73, 78–9 Shaw, A. 89, 90
Scheffman, D. 418 Sheldon, J. 438
Scheibehenne, B. 110 Shrestha, S. 417
Scherer, F. 199, 271, 274, 366, 379, 382, 390 Shur-Ofry, M. 614, 618, 621
Schiller, D. 586 Shy, O. 134, 301
Schlag, P. 386 Sichelman, T. 52, 58, 60–61, 125, 238, 332,
Schlatter, S. 601 369, 411, 434
Schlosser, S. 418 Sidak, J. 58, 136, 166, 292
Schmalensee, R. 269, 275, 288 Siebrasse, N. 448
Schmidt, G. 552, 567 Siegel, J. 353
Schmitz, J. 232 Siegelman, P. 5, 397, 412
Schmookler, J. 638 Silberston, A. 137
Schneider, P. 663–4 Silbey, J. 113
Schuchman, R. 411 Simcoe, T. 164, 339
Schuler, P. 604 Simon, D. 453
Schultz, J. 14, 17, 148, 200, 207, 220, 291, 341, Simon, H. 52
343, 395 Singer, J. 4
Schultz, M. 313, 314, 316, 583, 586, 587, 601, Singh, J. 606
611, 614, 626 Siwek, S. 586, 626
Schultz, T. 649, 651 Skladany, M. 585, 620
Schulz, N. 28 small claims courts 416
Schumpeter, J. 232, 331–2, 662 Smart, J. 73
Schwartz, B. 110 Smarzynska Javorcik, B. 667
Schwartz, D. 35, 40, 418, 434, 481, Smith, H. 2, 5, 6, 7, 8, 15, 16, 17, 47–70,
482–3 383, 431, 436, 540, 559, 571
Schwartz, H. 65 Smith, J. 413
Schwartz, M. 289, 293 Smith, K. 415
Schwartz, W. 411 Smith, M. 622
scope-of-the-patent test 238–40 Snow, A. 495, 499
Scotchmer, S. 32, 41, 75, 102, 119–56, 158, 173, Sobel, G. 211
183, 192, 232, 234, 236, 251, 272, 288, 293, Sobel, R. 286
330, 331, 378, 384, 390, 395, 429, 553, 557, social costs of intellectual property rights
562, 567, 660 383, 384–5
Scott, J. and T. 232 social demand and private demand,
Seaman, C. 233 difference between 555–6
Seijo, B. 82 social norms
semi-commons regime 56, 66, 559–60, 571–3 enforcement costs, copyright law
see also commons enforcement 414–15
Sen, A. 103, 104, 561, 637, 651–2 negative space of IP 314–17
Serrano, R. 73 social returns versus private returns,
Sevcenko, C. 29 infrastructure theory 553–4
Seymore, S. 427 social value
Shachar, R. 273 network effects 159, 161–2, 163
Shankerman, M. 28 prize and reward alternatives 360–61,
Shankland, S. 212 365, 366, 367
Shannon, J. 191 social welfare see welfare promotion
Shapiro, C. 28, 35, 37, 39, 58, 130, 135, 159, software industry 143, 145, 147
161, 164, 166, 168, 172, 188, 201, 212, 213, network effects and copyright protection
233, 249, 271, 282, 341, 385, 390, 396, 408, see network effects, copyright
417, 438, 448, 530 protection in software industry

network effects and patent protection see Stiglitz, J. 266, 292, 360, 379, 390, 648, 657
network effects, patent protection for Stole, L. 269, 289
software-related technologies Story, A. 585, 620
trade secret protection 172–3 Strandburg, K. 146, 314, 333, 339, 340, 345,
Soini, S. 30, 34 393, 395, 573, 613
Sokoloff, K. 363 Strang, L. 62
Solow, R. 75, 232, 638, 639 Straus, J. 30
Solum, L. 99, 515 Stutzer, A. 103
Song, M. 592, 618 substantial similarity test, copyright 129
spatial competition, product differentiation substitutive works 500–501
266–9, 272–4 see also complementary products
Spatt, C. 164 Suk, J. 82, 111, 313
Spatz, J. 667 Sukhatme, N. 112
special interest groups involvement, custom Sullivan, L. 293
role 539–40 Sumner, L. 102, 104
specialized trial court suggestion 482–3 Sundaram, S. 589, 590, 591, 605, 607, 608
Spector, H. 85 Sunder, M. 585, 618, 623
Spence, M. 266 Sunstein, C. 93, 386
Spiegler, P. 638 Supreme Court 39, 465, 474–5, 476, 477,
Spier, K. 299 479–80, 483, 484–5
spillover effects, infrastructure theory 561, 568, see also individual cases
569, 572 Swanson, T. 668–9
Spoo, R. 313, 315–16 Sweeney, M. 20
Sprigman, C. 10, 82, 111, 234, 309–29, 330, Syed, T. 102, 104, 263, 273, 356–7, 359, 360,
394, 449, 452, 530–31, 546, 547, 610 364, 365, 372, 378, 381, 382, 384–5, 411,
Spulber, D. 269, 365 594, 620
Stake, J. 7 Sykes, A. 293
Stanca, L. 110
stand-alone innovation see innovation, Takeyama, L. 301, 396
stand-alone Tandon, P. 130, 271
standard-essential patents (SEPs) 166, 171, Tangri, R. 149, 440
199, 201, 202–3, 208–10, 214, 215–16, Tanzania, intellectual property law 649
242 Tassey, G. 552
standard-setting organizations (SSOs) 94, 148, Taylor, C. 137
149–50, 162, 166, 171, 199, 201–3, 206, Taylor, D. 58
208, 428 teaching, suggesting, and motivation (TSM)
standardization, network effects 163, 164–5, approach 479–80
167–8 technological advancement and change
standardized unique identification number and common law 513
systems 451–2 economic development 639–40, 641, 642,
standards and codes, network effects 164, 643, 644, 659, 665–9
167–8, 191–2 and enforcement costs 409–10, 414,
state support, and creative development see 415–16
creative development and copyright, state and innovation 121
support and network effects 159–60, 169–70, 171,
Stefanadis, C. 300 218
Stein, A. 412 notice and disclosure, copyright tracing and
Steinmueller, W. 553 ownership 449–52, 455, 459
Sterk, S. 5, 17, 300, 447, 456 technology
Stern, J. 4, 8, 10 and collective management 503–5
Stern, S. 29, 31, 335, 338 computer databases and first-mover
Stewart, S. 29 advantage 317–18
‘sticky’ information, and open innovation creative development and cultural diversity
332–3, 335, 342 619, 622–3
Stigler, G. 266, 282, 284 digital registration proposal 451, 461

Index 701
digital technology effects, emerging transaction costs

industries in Global South 585–6, 588, and anticommons see anticommons,
589, 590, 592, 605–10, 611, 613, 615, transaction costs, and patent
617 aggregators
and price discrimination 286–7 and collective management 495, 501
software industry see software industry infrastructure theory 564, 570
and welfare promotion 107–9 innovation, cumulative 140, 143, 147
technology transfer open innovation and ex ante licensing 332–3,
economic development 642, 643, 647, 648, 334–5, 342
650–51 and price discrimination 291
innovation, cumulative innovation 148, 150 prize and reward alternatives 370–71
open innovation and ex ante licensing 331–2, uniformity cost reduction 385–6
331–3, 334–5, 338, 342 Trebilcock, M. 642
Teece, D. 338 Treiger-Bar-Am, K. 78
Tehranian, J. 291, 301, 412 Trubek, D. 640
Telang, R. 590, 606, 607, 608, 622, 625 Tuckman, H. 395
Teubal, M. 552 Tufano, P. 323
thing-definition, governance of intellectual Tugend, A. 159
property 48–55 Tullock, G. 411
Thisse, J. 134, 287, 301 Tur-Sinai, O. 102, 104, 105
Thomas, J. 123, 478, 482 Turner, J. 477
Thompson, E. 528 Tushnet, R. 313, 531, 546
Thomson, J. 87 tying arrangements 246–8, 250–51, 285–6,
Thorpe, J. 495 297–300
threshold effect, economic development Tyler, T. 415
664
threshold requirements Uhlir, P. 341
innovation, cumulative 140–45 underproduction problem, and open
innovation, stand-alone 123–6 innovation 330–31, 343–4
Tiebout, C. 273, 521 uniformity cost reduction 377–406
Tiller, E. 387, 393 administrative benefits of uniform rights in
Tirole, J. 185, 249, 288, 289, 295, 313, 342, 343, patent system 383–4
366, 395, 412 allocative efficiency 385–6
To, T. 269, 385 alternative appropriability mechanisms
Tomlinson, B. 233 395–7
Torrance, A. 233, 343 alternative incentives to create or innovate
tort law, and custom role 528, 541 and distribute 394–5
Towse, R. 495, 498 business method patents 384, 392, 394–5
trade secrets and copying 388, 391, 393, 396
common law 519, 520 and copyright 387–8, 390–91, 394, 397
innovation, cumulative 147 decentralization argument 380, 381
innovation, stand-alone 123, 126, 132, demand side and positive spillovers 397
135–6 direct subsidies 395–6
network effects 172–3, 182–3 exclusionary practices 382, 383–4, 387, 397
notice and disclosure 427, 440, 461, 465 exclusive rights, tailoring 383, 391–4, 397
see also misappropriation doctrine follow-on innovation 398
trademarks information-asymmetry 379–80, 392
and custom role 538–9, 547 judicial tailoring 393
enforcement costs 418 lead-time advantage 396
and governance of intellectual property 50 market efficiency limits 380
network effects see network effects, network effects 396–7
trademark protection one-size-fits-all argument 383–5
notice and disclosure 435, 465 patent law and real option pricing 389,
and price discrimination 295 390
tragedy of the anticommons 27, 28–9 patentable subject matter definition 386–8

patents to facilitate financing of drug Sherman Act 210, 235, 245, 246, 251, 252
discovery 381 US, cases
‘person having ordinary skill in the art’ 1-800-CONTACTS v. WhenU.Com 256
(PHOSITA) 387, 393 321 Studios v. Metro-Goldwyn-Mayer
price discrimination 385 Studios 187
private ordering, tailoring through 385–6 Adams v. Burke 207, 238, 254
prize or reward strategy 379, 380, 381 Akzo v. Int’l Trade Comm. 296
product differentiation strategies 397 In re Alappat 219
public fund investment in innovation, risks Alice Corp. v. CLS Bank International 143,
in 379–80 198, 220
real option pricing, tailoring through 388–91 Allied Orthopedic Appliances v. Tyco Health
rewards distribution issues 382–3 Care Grp. 237
risk-spreading justification and costs of American Geophysical Union v. Texaco 492,
failure 380 535
social costs of intellectual property rights American Society for Testing and Materials
383, 384–5 v. Public.Resource.Org 192
standards instead of rules 386–8 Amini Innovation Corp. v. Anthony Cal. 204,
transaction costs and allocative efficiency 463
385–6 Apple Computer v. Formula Int’l 176, 194
uniformity presumption of IP 379–81 Apple Computer v. Franklin Computer Corp.
willingness-to-pay and ability-to-pay, gap 176, 178, 217
between 380–81 Apple Computer v. Microsoft Corp. 179–80,
university patents 39, 145–6, 147 217
unpatented inventions and independent In re Apple iPod iTunes Antitrust Litig. 237,
inventors 132–3 246
unpublished works, protection of 461–2 Apple v. Motorola 150, 242
unreasonable infringement claims 251–2 Apple v. Samsung Elecs. Co. 150, 202, 204,
Urban, J. 16, 19, 148, 200, 220, 341, 343 464
US Application of Bergel 198
America Invents Act (AIA) 41–2, 124, 125, Ariad Pharmaceuticals v. Eli Lilly 8, 39, 426
475, 481, 484 Aronson v. Quick Point Pencil Co. 100
Bayh-Dole Act 38, 336, 392, 395 Asahi Glass Co. v. Pentech Pharm. 243
Clayton Act 210, 234, 235, 236, 245, 246, Aspen Skiing Co. v. Aspen Highlands Skiing
252, 253, 255 Corp. 210
Copyright Act 145, 148–9, 174–5, 183, Assessment Tech. of Wisconsin. v. Wiredata
186–8, 192, 235, 245, 320, 390, 392, 394, 207, 255
415, 455–6, 459–60, 490, 520, 531, 590 Ass’n for Molecular Pathology v. Myriad
Digital Millennium Copyright Act (DMCA) Genetics 39, 143
160, 186–8 AT&T Corp. v. Excel Commc’ns 387
Hatch-Waxman Act 239, 392, 447 Athletic Alternatives v. Prince Mfg. 49–50
International Trade Commission 202, 209, 481 Avia Group International v. L.A. Gear
Lanham Act 192, 193, 195 California 462
Music Modernization Act (MMA) 435, 451, Baker v. Selden 173, 174, 191, 256, 462
455, 459, 460–61 Barclays Capital v. Theflyonthewall.com
National Commission on New Technological 519
Uses of Copyrighted Works (CONTU) Basic Books v. Kinko’s Graphics 536, 540
Final Report 174–5, 176, 190 Bateman v. Mnemonics 179
National Endowment for the Arts 599 Bement & Sons v. Nat’l Harrow Co. 236
Patent Act 125, 143–4, 196–8, 202, 204–6, Berkey Photo v. Eastman Kodak Co. 246
210, 236, 240–41, 244–7, 251–3, 255, Bernhardt v. Collezione Europa USA 387
294, 392–3, 395, 427, 520 Berry Sterling Corp. v. Pescor Plastics 204
Patent and Trademark Office (USPTO) 241, Best Lock Corp. v. Ilco Unican Corp. 203,
251, 386, 389, 393, 409, 426, 434–5, 439, 204, 257
443, 445, 474–5, 477–8, 480–81, 483–6, Bikram’s Yoga College of India v. Evolation
509 Yoga 256

Index 703
Bill Graham Archives v. Dorling Kindersley Creative Labs v. Cyrix Corp. 194
456 Crocs v. Int’l Trade Comm’n 204
In re Bilski 197 CSIRO v. Cisco Sys. 150, 203
Bilski v. Kappos 197, 220 Cuno Engineering Corp. v. Automatic Devices
Blaustein v. Burton 428 Corp. 124
Bleistein v. Donaldson Lithographic Co. 110, Cybor Corp. v. FAS Technologies 477
585, 599 Data General Corp. v. Grumman Systems
Bloomer v. McQuewan 235, 238, 294 Support Corp. 211
Bobbs-Merrill Co. v. Straus 254 Dawson Chem. Co. v. Rohm & Haas Co. 60
Bonito Boats v. Thunder Craft Boats 183, Dell Computer Corp. 215
462, 522 Dellar v. Samuel Goldwyn 529
Bowers v. Baystate Techs. 186 In re Dembiczak 198
Bowman v. Monsanto 14, 254 Dennison Manufacturing v. Panduit 476
Braun Med. v. Abbott Labs. 244 Desny v. Wilder 428
Brear v. Fagan 20 Diamond v. Chakrabarty 387
Bridgeport Music v. Dimension Films 530 Diamond v. Diehr 197, 219
Bright Tunes Music Corp. v. Harrisongs Dickinson v. Zurko 480
Music 429 Dippin’ Dots v. Mosey 237
Broadcast Music v. Columbia Broadcast 300 Dolbear v. Am. Bell Tel. Co. 144, 195
Brooke Group Ltd. v. Brown and Williamson DSC Communications Corp. v. DGI
Tobacco Corp. 235 Technologies 207
Brulotte v. Thys Co. 255, 297 Duke University v. Madey 30
Bryant v. Mattel 454 Dun & Bradstreet Software Servs. v. Grace
Building Officials & Code Admin. v. Code Consulting 537
Technology 192 Eastman Kodak Co. v. Image Technical
California Computer Prod. v. IBM Corp. 247 Services 210, 300
Cambridge Univ. Press v. Patton 535 eBay v. MercExchange 39–40, 60, 64–5, 150,
Campbell v. Acuff-Rose Music 456, 516 201–2, 220, 237–8, 448, 456, 480
Capitol Records v. Thomas-Rasset 415 Egyptian Goddess v. Swisa 203–4, 462
Carbice Corp. v. American Patents Eldred v. Ashcroft 235, 390
Development Corp. 205, 238, 247 Elliott v. Google 194
Cariou v. Prince 516 Eng’g Dynamics v. Structural Software 179
Chamberlain Group v. Skylink Techs. 188 Ericsson v. D-Link Sys. 150, 202, 203
Chevron v. Natural Resources Defense Erie R.R. Co. v. Tompkins 521
Council 484, 486 Exxon Research and Engineering Co. v.
Chicago Board of Trade v. United States 235 United States 40
Chrysler Motors Corp. v Auto Body Panels of Faus v. City of Los Angeles 446
Ohio 256–7 Feltner v. Columbia Pictures Television 457
Cincinnati Car Co. v. New York Rapid Festo Corp. v. Shoketsu Kinzoku Kabushiki
Transit Corp. 202 Co. 447
Clorox Co. v. Sterling Winthrop 237 Folsom v. Marsh 516
Code Revision Comm’n for General Assembly Forasté v. Brown Univ. 536
of Georgia v. Public.Resource.Org 192 F.T.C. v. Actavis 238–40, 485
In re Comiskey 197 F.T.C. v. Borden’s Real Lemon 275
Computer Assocs. Int’l v. Altai 177–8, F.T.C. v. Phoebe Putney Health Sys. 246
179–80, 188, 217, 388, 537 Gates Rubber Co. v. Bando Chem. Indus. 179
Continental Paper Bag Co. v. E. Paper Bag Gen. Talking Pictures Corp. v. W. Elec. Co.
Co. 236, 252 244
Continental TV v. GTE Sylvania 245 Gentry Gallery v. Berkline Corp. 439
Continental Wall Paper Co. v. Louis Voight Geophysical Union v. Texaco 534
and Sons Co. 255 Georgia-Pacific Corp. v. U.S. Plywood Corp.
Cornell Univ. v. Hewlett-Packard Co. 202 202
Cortese v. United States 21 Golden Bridge Tech. v. Motorola 242
Costco Wholesale Corp. v. Omega 14 Gorham Co. v. White 462
Coupe v. Royer 238 Gottschalk v. Benson 196, 197, 219

In re GPAC 387 L.A. Gear v. Thom McAn Shoe Co. 203, 204,
Graham v. John Deere Co. 361 462
Graver Tank & Mfg. Co. v. Linde Air Lasercomb America v. Reynolds 206, 255
Products Co. 129, 447 LaserDynamics v. Quanta Computer 202
Halo Electronics v. Pulse Electronics 202, 440 Lee v. Dayton-Hudson Corp. 203
Harbor Software v. Applied Sys. 453 Leegin Creative Leather Prod. v. PSKS 245
Harper & Row, Publishers v. Nation Enters Lenz v. Universal Music Corp. 459
388, 534–5, 536, 537 Lexmark Int’l v. Static Control Components
Harper House v. Thomas Nelson 180 188
Hays v. Sony Corp. of Am. 536 Lotus Dev. Corp. v. Borland 181–2, 188, 217
Henley v. DeVore 456 Lotus Dev. Corp. v. Paperback Software 180
Henry v. A.B. Dick Co. 236 Lotus v. Borland 176, 218
Herbert Rosenthal Jewelry Corp. v. Kalpakian L.P. v. Penguin Books USA 191
180 Lucent Techs v. Gateway 202
Hewlett-Packard Co. v. Nu-Kote Intern. 194 Madey v. Duke University 30, 145
Highmark v. Allcare Health Mgmt. Sys. 480 MAI Systems v. Peak Computer 298–9
IBM Corp. v. United States 298 Markman v. Westview Instruments 434,
Illinois Tool Works v. Indep. Ink 241, 247, 443–4, 476–7
286 Mattel v. MGA Entertainment 454
Image Tech. Servs. v. Eastman Kodak Co. Mattel v. Walking Mountain Prods. 191
211, 253 May v. Morganelli-Heumann & Assocs. 537
Impression Products v. Lexmark International Mayo Collaborative Services v. Prometheus
14, 63, 207, 248, 254, 294–5 Laboratories 39, 143, 198, 220
In re Independent Service Organizations Mazer v. Stein 462
Antitrust Litigation 211 MDY Indus. v. Blizzard Entm’t 188
In re Innovatio IP Ventures LLC Patent Litig. MedImmune v. Genentech 434
150, 203 Microsoft Corp. v. Lindows.com 194
In re Intel Corp. 215 Microsoft Corp. v. Motorola 150, 203,
Intel v. Advanced Micro Devices 194 209–10, 253
Intellectual Ventures I v. Capital One Microsoft v. i4i Ltd 42
Financial Corp. 252 Mitchell Bros Film Group v. Cinema Adult
Intergraph Corp. v. Intel Corp. 211 Theater 319–20
International News Service v. Associated MiTek Holdings v. ARCE Engineering Co.
Press 511 182
International Salt Co. v. United States 241, Mitel v. Iqtel 179
297 Morton Salt Co. v. G.S. Suppiger Co. 205,
Inwood Labs. v. Ives Labs. 193–4, 204 256, 298
ISO Antitrust Litigation 253 In re Morton-Norwich Prods. 204
Jacobsen v. Katzer 185, 533 Motion Picture Patents Co. v. Universal Film
Jefferson Parish Hosp. Dist. No. 2 v. Hyde Mfg. Co. 205, 206, 236, 255–6, 298
286, 300 In re Motorola Mobility 214
Jim Henson Prods. v. John T. Brady & Murphy Door Bed Co. v. Interior Sleep
Assocs. 537 Systems 170
Johnson & Johnston Assocs. v. R.E. Serv. Co. Nautilus v. Biosig Instruments 40, 199, 220
238 NBA v. Motorola 523
Jungersen v. Ostby & Barton Co. 236 In re Negotiated Data Solutions LLC
K-Mart Corp. v. Cartier 295 (N-Data) 214
Kellogg Co. v. National Biscuit Co. 193 Nero AG v. MPEG LA 251
Kewanee Oil Co. v. Bicron Corp. 9, 100 New Era Publ’ns Int’l v. Carol Publ’g Grp.
Kimble v. Marvel Entm’t Co. 255, 485 536
King-Seely Thermos Co. v. Aladdin Indus. New York v. Microsoft Corp. 215
170 Newton v. Diamond 388
Kirtsaeng v. John Wiley & Sons 14, 245, 254, Nichols v. Universal Pictures Corp. 177
294, 628–9 Nobelpharma v. Implant Innovations 251
KSR Int’l Co. v. Teleflex 198, 220, 479–80 In re NTP 439

Index 705
In re Nuijten 197 Sheldon v. Metro-Goldwyn Pictures Corp.

Octane Fitness v. ICON Health & Fitness 429
251, 440 Siegel v. Chicken Delight 241, 248
OddzOn Products v. Just Toys 203, 463 Smith v. Dravo 428
Oil States Energy Services v. Greene’s Energy Soc’y of Holy Transfiguration Monastery v.
Grp. 481 Gregory 535
Omega v. Costco Wholesale Corp. 14, 255 Sony BMG Music Entm’t v. Tenenbaum 416
Oracle v. Google 150, 174, 182, 189–91, Sony Computer Entertainment v. Connectix
218–19, 221 Corp. 179, 190, 191
O’Reilly v. Morse 143, 387 Sony Corp. v. Universal Studios 100, 301
Otter Tail Power Co. v. United States 210 Stac Elec. v. Microsoft Corp. 219
Pandora Media v. Am. Soc’y of Composers, Standard Oil Co. of Calif. v. United States
Authors, and Publishers 459 246
Parker v. Flook 197–8, 219 Standard Sanitary Mfg. Co. v. United States
PHG Techs. v. St. John Cos. 204 236
Pierson v. Post 7 State ex. rel. Thornton v. Hay 528
Pittsburgh State Univ. v. Kansas Bd. of State Street Bank v. Signature Financial
Regents 536 Group 197, 219, 323, 387, 478
Plains Cotton Coop. Ass’n v. Goodpasture Storage Technology Corp. v. Custom
Computer Serv. 177 Hardware Eng’g & Consulting 188
Practice Mgmt. Info. Corp. v. American Med. Sun Microsystems v. Microsoft Corp. 194,
Assoc. 192, 207 212, 220–21
Princeton Graphics Operating v. NEC Home Synercom Technology v. University
Electronics 194 Computing Co. 177
Princeton Univ. Press v. Mich. Document TC Heartland LLC v. Kraft Food Brands 483
Servs. 534, 535 Teleflex v. KSR Intern. Co. 198
Princo Corp. v. Int’l Trade Comm’n 249, 252 Teva Pharmaceuticals USA v. Sandoz 40,
ProCD v. Zeidenberg 61–2, 290, 295 249, 480, 481
Qualcomm v. Broadcom Corp. 208, 242 The T.J. Hooper 528
Quality King Distributors v. L’anza Research TrafFix Devices v. Marketing Displays 256,
Int’l 294 462
Quanta Computer v. LG Electronics 14, 63–4, Trebro Mfr. v. Firefly Equip. 237, 252
254 Tulk v. Moxhay 447
Queen City Pizza v. Domino’s Pizza 248 Twentieth Century Fox Film v. Marvel Enters
Raab v. Casper 446 537
Rambus v. Infineon Technologies 208 Tyco Healthcare Group v. Mutual Pharma.
Regents of the University of California v. Eli Co. 251
Lilly 39 Uniloc USA v. Microsoft Corp. 202
Richardson v. Stanley Works 203, 463 United States v. Aluminum Co. of Am. 235
Ringgold v. Black Entm’t Television 534 United States v. Am. Can Co. 246
Roche Prods. v. Bolar Pharm. Co. 145 United States v. AT&T 253
Rodi Yachts v. Nat’l Marine 541 United States v. Carroll Towing Co. 514
Rosco v. Mirror Lite Co. 204 United States v. Dentsply Int’l 245
Rosemont Enters. v. Random House 536, 542 United States v. E.I. du Pont de Nemours 235
Rosetta Stone v. Google 256 United States v. General Electric Co. 238
Roy Exp. Co. Establishment of Vaduz v. CBS United States v. Line Material Co. 238
535, 544 United States v. Loew’s 241
Samsung Elecs. v. Apple 205 United States v. Microsoft Corp. 215, 242, 246
Sanitary Refrigerator Co. v. Winters 129 United States v. Singer Mfg. Co. 240
SCM Corp. v. Xerox Corp. 211–12, 242 United States v. Univis Lens Co. 254
Sega Enterprises Ltd. v. Accolade 179, 183, Universal City Studios v. Corley 187
187, 188, 190, 191, 194–5, 217, 298 University of Rochester v. G.D. Searle 39
Seiko Epson Corp. v. Nu-Kote Int’l 204 U.S. Gypsum Co. v. National Gypsum Co. 256
Shapiro, Bernstein & Co. v. P.F. Collier & Son U.S. Steel Corp. v. Fortner Enterprises 286
Co. 536 USM Corp. v. SPS Tech. 255

Vault Corp. v. Quaid Software 186 Voon, T. 585

Veeck v. S. Bldg. Code Congress Int’l 192
Viacom Int’l v. YouTube 457 Wagner, R. 36, 37, 211, 387, 389, 390, 393, 397,
VirnetX v. Cisco Sys. 202 432, 446, 479, 482
Wal-Mart Stores v. Samara Bros. 326 Waldfogel, J. 273, 321, 411, 412, 590, 604–5,
Walker Process Equip. v. Food Mach. and 606, 607, 608, 621, 623, 625
Chem. Corp. 237, 251–2 Waldman, M. 134, 299, 301, 418
Walker v. Time Life Films 388 Walker, R. 388, 410, 585, 596
Warner-Jenkinson v. Hilton Davis Chemical Walls, W. 382
41, 129, 447 Walsh, J. 30, 31, 33–4, 143, 146
Weinstein v. Univ. of Illinois 536 Walton, G. 137
Whelan Associates v. Jaslow Dental Ward, M. 669
Laboratory 176–7, 178, 217 Ware, J. 444
White-Smith Music Publishing Company v. Wasserman, M. 477, 478, 484
Apollo Company 12–13, 19 Waterson, M. 271, 390
Whitmill v. Warner Bros. Entertainment 315 Watt, R. 495, 499
Whittemore v. Cutter 145 Weber, S. 184, 343, 460
Williamson v. Citrix Online 199 Wegner, H. 457
Witmark & Sons v. Berger Amusement Co. Weil, M. 476
206 Weinberg, H. 275
Witmark & Sons v. Jensen 206 Weinreb, L. 312, 529, 540
W.L. Gore & Assocs. v. Garlock 132 Weiser, P. 396
user identification, collective management welfare promotion 98–118
492–3 consequentialism and IP law 99–101
user innovation 339–41, 344–5, 563 consumer-side effects 107–12
utilitarianism 73–84 copyright law 109–12, 115
and derivative works 113–14, 115
Vacca, R. 475–6, 479, 484 and economic development 658–9, 660
Vaidhyanathan, S. 4, 567 exclusivity and innovation costs 100–101,
valuation mechanism, prize and reward 102, 103
alternatives 351–2, 360–64 and follow-on innovation 102, 108, 114
Van Gelder, A. 583, 601 future research 111, 115
Van Gompel, S. 16, 449 happiness (subjective well-being) 105–7,
Van Hiel, A. 415 108–9
Van Houweling, M. 2–26, 48, 49, 61, 62, 63–4, individual preference satisfaction 102–4,
207, 425, 447, 452, 459, 460 108–9, 110, 111, 114
Van Overwalle, G. 393 innovation promotion and funding 103,
Van Ypersele, T. 121, 358, 361, 367, 379, 380, 104–5, 109, 114
394, 395 patent law 107–9, 112
Vanneste, S. 29, 415 and price discrimination 287–93
Varian, H. 133, 159, 161, 172, 188, 282, 288, producer-side effects 112–15
301, 385, 396 science and useful arts promotion 101–2
Verrier, R. 82 technologies and well-being gains 107–9
Vertinsky, L. 484 virtue ethics 104–5
Vetter, G. 66, 571 welfare economics 73–4
Vickers, J. 287 Wenzel, M. 415
Viscusi, W. 288, 290 West, J. 162, 337, 338, 340
Vishnubhakat, S. 389 Wheatland, T. 415, 416
Visscher, M. 269 Wheeler, G. 438
Vives, X. 287 Whinston, M. 270, 299, 412
Voena, A. 659 Wilde, L. 121, 128
voluntary licensing 458–60 Wildman, S. 586, 626
Von Hippel, E. 82, 113, 122, 313, 314, 316, 330, Wiley, J. 292, 298
332, 333, 336, 339–40, 341, 342, 345, 530, Williams, B. 73
546, 547, 563 Williams, H. 31–2, 108, 359

Index 707
Williams, P. and R. 411 Wu, T. 291, 301, 381, 393, 398, 412
Williamson, O. 165–6, 334, 640 Wyatt, J. 300
Willig, R. 134 Wyman, K. 57
willingness-to-pay model 282–3, 380–81,
497–8 Yablon, D. 432, 462, 464, 465
Wilson, B. 213, 237 Yagi, B. 31, 32
Wilson, T. 275 Yamane, H. 656
Winn, A. 29 Yan, I. 661
winner-take-all biases of media markets 621 Yang, G. 666, 667
Winokur, J. 20 Yao, D. 135, 428
Winter, R. 334 Yelderman, S. 346
Winter, S. 121, 122, 333, 640 Yen, A. 8
Wolfstetter, E. 357–8 Yoffie, D. 35
Wooders, M. 268, 269 Yoo, C. 262–80, 384, 385, 620
work-for-hire doctrine 536–7 Yoon, Y. 28
Wright, B. 121, 354–5, 357, 358, 368, 369 Yu, P. 56, 411, 492, 571, 584, 593, 604, 648
Wright, D. 390
Wright, J. 293, 299 Zamir, E. 84
WTO TRIPS Agreement 360, 586, 643, 644, Ziedonis, A. 31, 34, 36, 166, 382, 408, 413
660–61, 667, 670–71 Zimmerman, D. 84, 603

RESEARCH HANDBOOK ON THE ECONOMICS OF

Series Editors: Richard A. Posner, Judge, United States Court of Appeals for the Seventh Circuit
and Senior Lecturer, University of Chicago Law School, USA and Francesco Parisi, Oppenheimer
Wolff and Donnelly Professor of Law, University of Minnesota, USA and Professor of Economics,
University of Bologna, Italy
Edited by highly distinguished scholars, the landmark reference works in this series offer advanced
treatments of specific topics that reflect the state-of-the-art of research in law and economics,
while also expanding the law and economics debate. Each volume’s accessible yet sophisticated
contributions from top international researchers make it an indispensable resource for students
and scholars alike.
Titles in this series include:
Research Handbook on the Economics of Property Law

Edited by Kenneth Ayotte and Henry E. Smith
Research Handbook on the Economics of Family Law

Edited by Lloyd R. Cohen and Joshua D. Wright
Research Handbook on the Economics of Antitrust Law

Edited by Einer R. Elhauge
Research Handbook on the Economics of Corporate Law

Edited by Brett McDonnell and Claire A. Hill
Research Handbook on the Economics of European Union Law

Edited by Thomas Eger and Hans-Bernd Schäfer
Research Handbook on the Economics of Criminal Law

Edited by Alon Harel and Keith N. Hylton
Research Handbook on the Economics of Labor and Employment Law

Edited by Michael L. Wachter and Cynthia L. Estlund
Research Handbook on Austrian Law and Economics

Edited by Todd J. Zywicki and Peter J. Boettke
Research Handbook on Behavioral Law and Economics

Edited by Joshua C. Teitelbaum and Kathryn Zeiler

Volume 1: Theory
Edited by Ben Depoorter and Peter S. Menell

Edited by Peter S. Menell and David L. Schwartz

Research Handbook on the
Economics of Intellectual
Property Law
Edited by
Peter S. Menell
University of California, Berkeley School of Law, USA
David L. Schwartz
Northwestern University, School of Law, USA
Cheltenham, UK • Northampton, MA, USA

© The Editors and Contributors Severally 2019
All rights reserved. No part of this publication may be reproduced, stored in a retrieval system or
transmitted in any form or by any means, electronic, mechanical or photocopying, recording, or
otherwise without the prior permission of the publisher.
Published by
Edward Elgar Publishing Limited
The Lypiatts
15 Lansdown Road
Cheltenham
Glos GL50 2JA
UK
Edward Elgar Publishing, Inc.

William Pratt House
9 Dewey Court
Northampton
Massachusetts 01060
USA
A catalogue record for this book

is available from the British Library
Library of Congress Control Number: 2019945636
This book is available electronically in the

Law subject collection
DOI 10.4337/9781789903997
ISBN 978 1 84844 536 9 (2 volume set)

ISBN 978 1 78990 399 7 (eBook)
Typeset by Servis Filmsetting Ltd, Stockport, Cheshire

Contents
List of contributorsviii
PART I EMPIRICAL METHODS
1 Data sources on patents, copyrights, trademarks, and other intellectual

property2
David L. Schwartz and Ted Sichelman
PART II EMPIRICAL STUDIES RELATING TO PATENTS
Section A Metrics
2 Patent citation data in social science research: overview and best practices 20
Adam B. Jaffe and Gaétan de Rassenfosse
3 Patent value 47
John R. Allison
Section B Patent Institutions and Litigation
4 Empirical scholarship on the prosecution process at the USPTO 77

Michael D. Frakes and Melissa F. Wasserman
5 The USPTO’s Patent Trial and Appeal Board 92
Arti K. Rai and Saurabh Vishnubhakat
6 The Federal Circuit as an institution 104
Ryan Vacca
7 Empirical studies of claim construction 159
J. Jonas Anderson and Peter S. Menell
8 Empirical studies of the International Trade Commission 175
Colleen V. Chien and David L. Schwartz
9 Technical standards, standards-setting organizations, and intellectual
property: a survey of the literature (with an emphasis on empirical
approaches)185
Jorge L. Contreras
10 Empirical studies of patent pools 236
Michael Mattioli

vi Research handbook on the economics of IP law volume 2
11 Empirical analyses related to university patenting 256

Arvids A. Ziedonis
PART III PATENT LAW DOCTRINES
12 Empirical studies in patentability 281

Ronald Mann and Christopher Cotropia
13 Patent duration 310
Brian J. Love
14 Infringement 326
Lee Petherbridge and Jason Rantanen
15 Presumption of validity 350
Christopher B. Seaman
16 Inequitable conduct and patent misuse 367
Lee Petherbridge and Jason Rantanen
17 Remedies 390
Thomas F. Cotter and John M. Golden
PART IV TECHNOLOGY-SPECIFIC STUDIES
18 Patent rights and innovation: evidence from the semiconductor industry 423
Rosemarie H. Ziedonis and Alberto Galasso
19 Patent trolls 445
Jay P. Kesan
20 Patents and innovation in economic history 462
Petra Moser
21 The political economy of intellectual property reforms 482
Jay P. Kesan and Andres A. Gallo
PART V EMPIRICAL STUDIES RELATING TO COPYRIGHT
22 Empirical studies of copyright litigation 511

Matthew Sag
23 Empirical studies of copyright registration 533
Dotan Oliar
24 Copyright and technological change in music, movies, and books 547
Joel Waldfogel
25 Music copyright 564
Peter DiCola

Contents vii
26 Experiments in intellectual property 579

Christopher Buccafusco and Christopher Jon Sprigman
27 The effect of copyright law on access to works 605
Paul J. Heald
PART VI EMPIRICAL STUDIES OF TRADEMARK LAW
28 Empirical studies of trademark law 617

Barton Beebe
PART VII EMPIRICAL METHODS IN TRADE SECRET RESEARCH
29 Empirical methods in trade secret research 638

Michael Risch
PART VIII KNOWLEDGE COMMONS
30 Knowledge commons 656

Michael J. Madison, Katherine J. Strandburg and Brett M. Frischmann
Index677

Contributors
John R. Allison, The Mary John and Ralph Spence Centennial Professor, McCombs
Graduate School of Business, University of Texas at Austin, USA
J. Jonas Anderson, Professor of Law, American University, Washington College of Law,
USA
Barton Beebe, John M. Desmarais Professor of Intellectual Property Law, New York
University School of Law, USA
Christopher Buccafusco, Professor of Law and Director of the Intellectual Property and
Information Law Program, Cardozo School of Law, Yeshiva University, New York, USA
Colleen V. Chien, Professor of Law, Santa Clara University Law School, California, USA
Jorge L. Contreras, Professor, University of Utah S.J. Quinney College of Law, USA
Christopher Cotropia, Professor of Law, University of Richmond School of Law, USA
Thomas F. Cotter, Briggs and Morgan Professor of Law, University of Minnesota Law
School, USA
Gaétan de Rassenfosse, Assistant Professor, Brandeis University (USA) and Queensland
University of Technology (Australia)
Peter DiCola, Professor of Law and Searle Research Fellow, Northwestern University
Pritzker School of Law, USA
Michael D. Frakes, Professor of Law and Economics, Duke University, USA
Brett M. Frischmann, Charles Widger Endowed University Professor in Law, Business
and Economics, Charles Widger School of Law, Villanova University, Philadelphia, USA
Alberto Galasso, Associate Professor, University of Toronto, Canada and Research
Associate, National Bureau for Economic Research, USA
Andres A. Gallo, Professor, Coggin College of Business, University of North Florida,
USA
John M. Golden, Loomer Family Professor in Law, University of Texas at Austin, USA
Paul J. Heald, Richard W. and Marie L. Corman Professor of Law, University of Illinois
College of Law, USA
Adam B. Jaffe, Research Professor, Brandeis University (USA) and Adjunct Professor,
Queensland University of Technology (Australia)
Jay P. Kesan, Professor of Law and Workman Research Scholar at the University of
Illinois at Urbana-Champaign, USA
viii

Contributors ix
Brian J. Love, Associate Professor and Co-Director of the High Tech Law Institute, Santa
Clara University School of Law, USA
Michael J. Madison, Professor of Law and Faculty Director, Innovation Practice Institute,
University of Pittsburgh School of Law, USA
Ronald Mann, Albert E. Cinelli Enterprise Professor of Law, Columbia Law School,
USA
Michael Mattioli, Professor of Law, Indiana University Maurer School of Law, Indiana
University, USA
Peter S. Menell, Koret Professor of Law and Director, Berkeley Center for Law &
Technology, Berkeley School of Law, University of California, USA
Petra Moser, Associate Professor of Economics, NYU Stern School of Business,
USA and Research Associate, National Bureau of Economic Research, Cambridge,
Massachusetts, USA
Dotan Oliar, Professor of Law, University of Virginia School of Law, USA
Lee Petherbridge, Professor of Law, Vachon Research Fellow, Loyola Law School, Loyola
Marymount University, USA
Arti K. Rai, Elvin R. Latty Professor of Law, Duke Law School; Faculty Director, Duke
Law Center for Innovation Policy; Duke Innovation and Entrepreneurship Initiative
Research Fellow, USA
Jason Rantanen, Professor, Ferguson-Carlson Fellow in Law, and Director of the
Innovation, Business, and Law Program, University of Iowa Law School, USA
Michael Risch, Professor of Law, Villanova University School of Law, USA
Matthew Sag, Professor, Loyola University Chicago School of Law and Associate
Director for Intellectual Property, Institute for Consumer Antitrust Studies, USA
David L. Schwartz, Stanford Clinton Sr. and Zylpha Kilbride Clinton Research Professor
of Law, Northwestern University School of Law, USA
Christopher B. Seaman, Associate Professor of Law, Washington and Lee University
School of Law, USA
Ted Sichelman, Professor of Law, University of San Diego Law School, USA
Christopher Jon Sprigman, Professor, New York University School of Law and
Co-Director, Engelberg Center on Innovation Law and Policy, USA
Katherine J. Strandburg, Alfred B. Engelberg Professor of Law, New York University
School of Law, USA
Ryan Vacca, Professor of Law, University of New Hampshire School of Law, USA
Saurabh Vishnubhakat, Associate Professor of Law, Texas A&M University School of
Law, USA

x Research handbook on the economics of IP law volume 2
Joel Waldfogel, Professor, Frederick R. Kappel Chair in Applied Economics, University

of Minnesota, USA
Melissa F. Wasserman, Charles Tilford McCormick Professor of Law, The University of
Texas at Austin, USA
Arvids A. Ziedonis, Professor of Strategy and Innovation, Faculty of Economics and
Business, KU Leuven, Belgium
Rosemarie H. Ziedonis, Associate Professor, Strategy and Innovation, Questrom School
of Business, Boston University, and Research Associate, National Bureau for Economic
Research, USA

PART I
EMPIRICAL METHODS

1. Data sources on patents, copyrights, trademarks
and other intellectual property
David L. Schwartz* and Ted Sichelman**
Contents
I. Patents
II. Trademarks
III. Copyrights
IV. Trade Secrets and Other Intellectual Property
V. Licensing Data
VI. IP Litigation
VII. Conclusion
References
In this chapter, we provide a roadmap of the sources of data on the various forms of
intellectual property (IP) protection. We first explain what data is available about patents,
copyrights, trademarks, and other types of IP, and where to find it. Then we identify and
analyze data sources specifically relating to IP licensing and litigation, growing areas of
research by scholars and lawyers.
I. PATENTS
There are numerous sources that permit advanced searching and free downloading of
individual patents. U.S. patents and published patent applications are freely searchable and
downloadable from the United States Patent and Trademark Office (USPTO) website.1
** Stanford Clinton Sr. and Zylpha Kilbride Clinton Research Professor of Law, Northwestern
University School of Law.
** Professor of Law, University of San Diego Law School. We thank Colleen Chien, Amit
Elazari, Paul Heald, Jason Rantanen, Michael Risch, Josh Sarnoff, Jeremy Sheff, Ryan Whalen,
and Heidi Williams for their helpful comments on previous drafts of this chapter.
1
U.S. patents can be searched at United States Patent and Trademark Office. 2017. ‘USPTO
Patent Full-Text and Image Database’, http://patft.uspto.gov/netahtml/PTO/search-adv.htm.
According to the website, the database includes all U.S. patents issued since 1976, whereas patents
from 1790 through 1975 are searchable by Issue Date, Patent Number, and Current Classification
(US, IPC, or CPC) categories. Full text images of U.S. patents back to 1790 are available at United
States Patent and Trademark Office. 2017. ‘US Patent Full-Page Images’, http://patft.uspto.gov/
netahtml/PTO/patimg.htm, although the results are not searchable. U.S. published patent applica-
tions can be searched at United States Patent and Trademark Office. 2017. ‘Patent Application Full
Text and Image Database’, http://appft.uspto.gov/netahtml/PTO/search-adv.html. The USPTO
began publishing certain patent applications in 2001.

Data sources on intellectual property 3
There are other free websites that permit searching of U.S. patents.2 These websites appear
to obtain their data from the USPTO.
Bibliographic information about individual U.S. patents is also downloadable for free
from the USPTO.3 Bibliographic information includes, for instance, inventor names,
assignee of patent at issuance, U.S. and International patent classification, filing date,
issue date, claims of priority, etc. The USPTO also provides basic aggregate statistics
about U.S. patent grants, filings, and other patent-related information.4 The availability
of bulk downloadable data will be discussed shortly.
International patent information is also available from a number of sources. European
Patent Office (EPO) patents are searchable at the EPO’s eSpaceNet website.5 The EPO
also operates PatStat,6 which permits users to generate bibliographic information on
all EPO patents, for a fee (de Rassenfosse et al., 2014). The World Intellectual Property
Organization (WIPO) hosts a free website, PatentScope, that permits searching of WIPO
patents.7 Google patents permits searching of some international patents, in addition to
U.S. patents.8 The Organization for Economic Co-operation and Development (OECD)
also provides basic statistical information on international patents.9
In addition to governmental sources, various commercial databases permit searching
of U.S. and international patents. For instance, LexisNexis, Westlaw, and BloombergLaw10
permit searching of U.S. patents. Thomson Innovation,11 which is owned by Westlaw,
permits more extensive searching and analysis of patents. For example, Thomson
Innovation maintains data on patent families and allows for examiner-provided citations
to be distinguished from applicant-provided citations (starting for patents and applica-
tions published in 2001). While these databases all contain the underlying patents, the
commercial databases permit easier downloading of demographic characteristics of
patents after a desired query.
2
Numerous websites permit free downloading and searching of patents, including, for
instance, 2017. Free Patents Online, freepatentesonline.com and Google Patents, https://patents.
google.com/.
3
The data, beginning in 2002 and through the present, is downloadable at ReedTech. 2017.
‘USPTO Data Sets’, http://patents.reedtech.com/pgog.php. Some scholars report finding it time-
consuming and generally difficult to parse the bibliographic information into a useful format.
4
United States Patent and Trademark Office. 2017. ‘Statistics’, www.uspto.gov/learning-and-re
sources/statistics.
5
EPO patents can be searched at European Patent Office. 2017. ‘Espacenet Patent search’,
http://worldwide.espacenet.com.
6
European Patent Office. 2017. ‘Patstat’, www.epo.org/searching/subscription/raw/product-1
4-24.html.
7
World Intellectual Property Organization. 2017. ‘Patentscope’, https://patentscope.wipo.int/
search/en/search.jsf.
8
The Help page for Google patents indicates that it obtains patent information from the
USPTO, the EPO, and WIPO. Google. 2017. ‘About Google Patents’, https://support.google.com/
faqs/answer/2539193.
9
OECD. 2017. ‘Intellectual property (IP) statistics and analysis’, www.oecd.org/science/inno/
oecdpatentdatabases.htm. The OECD website indicates that the data is mainly derived from the
EPO’s PatStat dataset.
10
Bloomberg Law. 2017. www.bloomberglaw.com/patent_search.
11
Clarivate Analytics. 2017. ‘Derwent Innovation’, http://info.thomsoninnovation.com/.

Patent citation information is available through the National Bureau of Economic

Research (NBER).12 The NBER website contains several free data files. These files
contain various demographic information about each U.S. patent issued from 1976 until
2006, including assignee, filing date, issue date, and International classification. They also
include the number of times that the patent was cited in patent prosecution by another
patent. Unfortunately, the NBER citation field does not distinguish between citations
provided by a patent applicant and those supplied by the patent examiner.13 An update
to the NBER patent citation data file to include more recent patents is expected to be
released shortly. The NBER patent citation data includes all citations to U.S. patents,
including citations by the same inventor. Such self-citations can be removed by using the
disambiguated inventor dataset. Lee Fleming and others have updated and refined the
NBER patent dataset, and made their data available to the public.14
Recently, the USPTO, through an agreement with ReedTech, has released bulk patent
data in a granular format, which is specifically geared for researchers.15 Using the NBER
classifications, the USPTO constructed the USPTO Historical Patent Data Files, four
research datasets containing time series and micro-level data by NBER sub-category on
applications, grants, and in-force patents spanning two centuries of innovation.16 The
USPTO also has datasets including the gender of inventors, the full text of claims, foreign
priority, the number of figure sheets, and other information.17 We understand that all of
the relevant data from grants and published applications—e.g., citations—can be parsed
directly from the full-text XML files available via bulk downloads from Google.
Information about patent prosecution is also available for download. Through the
Patent Application Information Retrieval (PAIR) system, one can view the prosecution
history of a U.S. issued patent or published patent application.18 For a majority of patents
issued and applications published in the last five to ten years, PDF images of the actual
12
National Bureau of Economic Research. ‘Patent Data Project’, https://sites.google.com/site/
patentdataproject/Home. The seminal article introducing and explaining the dataset is Hall et al.
(2001).
13
Beginning with patents issued around 2002, examiner citations were marked with an asterisk
on the face of the patent. Citations cited by the applicant in an information disclosure statement do
not have an asterisk. For patents issued before 2002, there is no way to identify examiner citations
from the face of the patent.
14
https://dataverse.harvard.edu/dataset.xhtml?persistentId=hdl:1902.1/15705.
15
The USPTO’s home page for its electronic bulk data (patent and trademark) is available at
United States Patent and Trademark Office. 2017. ‘Bulk Data Products’, www.uspto.gov/learning-
and-resources/electronic-bulk-data-products. ReedTech has released additional data to the public,
available at ReedTech. 2017. ‘Public PAIR Patent Access Download’, http://patents.reedtech.com/
Public-PAIR.php.
16
United States Patent and Trademark Office. 2017. ‘Historical Patent Data Files’, http://www.
uspto.gov/learning-and-resources/electronic-data-products/historical-patent-data-files. For a discus-
sion of the USPTO’s historical patent data files, see Marco et al. (2015).
17
http://www.patentsview.org/download/.
18
Members of the public can use Public PAIR, which requires entry of a Recaptcha code.
United States Patent and Trademark Office. 2017. ‘Patent Application Information Retrieval’,
http://portal.uspto.gov/external/portal/pair. Public PAIR has publicly available information on
patents and published applications. Private PAIR, which is only available to registered patent
attorneys or agents, does not require a Recaptcha and further provides access to non-public
information on applications that the attorney or agent is authorized to view. United States

prosecution documents (i.e., office actions, restriction requirements, originally filed

applications, amendments, etc.) can be accessed. For older patents and applications, only
the table of contents of the prosecution history is available. The full prosecution history
for these older patents is only available in hardcopy format from the USPTO. Through
cooperation with the USPTO, Google has made available certain PAIR data.19 We believe
that Google’s patent PAIR data is generally complete for applications published in the last
several years, although it may not be comprehensive historically. The USPTO has released
parsed datasets of patent examination data,20 including a separate research dataset on
office actions.21 Patent Advisor22 is a Reed Technology/Lexis product that has a database
of millions of prosecution histories and provides information about patents and patent
prosecution. Researchers should note that applications which have not been published or
issued as patents, whether because they are premature (less than 18 months from filing)
or because the applicant requested non-publication, are missing from the PAIR data.
PatentBox also permits searching of patent file wrappers.23 Patent Ninja provides detailed
statistical information relating to specific patent examiners, including information relat-
ing to examiner interviews at the USPTO.24
Google has also assembled additional patent data, which it makes available for free
download. The additional patent data includes text from patent assignments (1980–
current),25 patent maintenance fee filings and abandonments (1981–current),26 patent
classification,27 and Image File Wrapper petition decisions.28 The patent assignment data
only includes assignments and other documents recorded with the USPTO.29 Because
recording is voluntary (although encouraged), the assignment database represents an
incomplete subset of all assignments. The remaining dataset should be more complete,
although we have not independently verified as much.
Although the USPTO has released substantial amounts of data on patent prosecution,
data about the agency itself is less available. The Freedom of Information Act (FOIA)
Patent and Trademark Office. 2017. ‘Portal Applications’, https://ppair.uspto.gov/authenticate/

AuthenticateUserLocalEPF.html.
19
Google. 2012. ‘USPTO Bulk Downloads: PAIR Data, www.google.com/googlebooks/uspto-
patents-pair.html.
20
Graham, Marco and Miller (2017). The dataset is available for download at www.uspto.gov/
learning-and-resources/electronic-data-products/patent-examination-research-dataset-public-pair.
21
Lu, Myers and Beliveau (2017). The dataset is available for download at www.uspto.gov/
learning-and-resources/electronic-data-products/office-action-research-dataset-patents.
22
http://kr.lexisnexisip.com/products-services/intellectual-property-solutions/lexisnexis-patent
advisor.
23
The Patent Box. 2017, www.thepatentbox.com/.
24
https://examiner.ninja/.
25
Google. 2012. ‘USPTO Bulk Downloads: Patent Assignment Text’, www.google.com/google
books/uspto-patents-assignments.html.
26
Google. 2012. ‘USPTO Bulk Downloads: Patent Maintenance Fees’, www.google.com/google
books/uspto-patents-maintenance-fees.html.
27
Google. 2012. ‘USPTO Bulk Downloads: Patent Classification Information’, www.google.
com/googlebooks/uspto-patents-class.html.
28
Google. 2012. ‘USPTO Bulk Downloads: Patent Petition Decisions’, www.google.com/google
books/uspto-patents-petitions.html.
29
Various companies, such as Innography, offer seemingly similar information for a fee.

permits members of the public to request information from federal agencies. These
include the USPTO (via the Department of Commerce), the Copyright Office, and the
Food and Drug Administration. For example, Michael Frakes and Melissa Wasserman
obtained patent processing information (Frakes and Wasserman, 2013, p. 92), as well as a
full examiner roster from the USPTO (Frakes and Wasserman, 2017, p. 553).30
While there is no specific form to use in making a FOIA request, the government
provides specific contact information to use for each agency.31 The FOIA request itself
must reasonably describe the records sought. The agency is required to promptly respond
to the request and provide responsive documents or indicate a reason for denial of the
request. The government may charge educational requestors for the costs of duplication
of the records, and may charge commercial users search charges as well. Public Citizen
publishes an informative guide on the FOIA process.32 The courts, however, are not
subject to FOIA requests.
II. TRADEMARKS
Relative to patents, there are fewer sources of data on trademarks. Free online searches of
trademarks can be conducted at the USPTO website through the Trademark Electronic
Search System (TESS).33 TESS permits searching on text and images of registered
marks, and marks in pending and abandoned applications. Many other countries also
have government-run websites that permit trademark searching.34 Trademarks can also
be searched on Lexis, Westlaw, BloombergLaw and other fee-based places. One new fee-
based service, TrademarkNow, provides searching across many international trademark
offices and uses sophisticated artificial intelligence and natural language processing
techniques to find marks similar to those being searched.35
The USPTO, through the Office of the Chief Economist, has released bulk trademark
data.36 The USPTO trademark data set is updated annually, and contains detailed
information on 8.6 million trademark applications filed with or registrations issued by the
30
Other information, such as examiner wages, are also likely subject to a FOIA request.
31
United States Department of Justice. 2011. ‘FOIA.Gov: Make a Request’, www.foia.gov/
report-makerequest.html.
32
Public Citizen. 2017. ‘How to File a FOIA Request: A Guide’, http://citizen.org/Page.aspx?pid=
458.
33
TESS can be accessed at United States Patent and Trademark Office. ‘Trademark Electronic
Search System (TESS)’, http://tess2.uspto.gov/.
34
For instance, the U.K. operates Gov.UK. ‘Search for a trade mark’, www.gov.uk/search-for-
trademark, and the Canadian Intellectual Property Office operates Government of Canada. 2017.
‘Canadian Trademarks Database’, www.cipo.ic.gc.ca/app/opic-cipo/trdmrks/srch/bscSrch.do. See
also World Intellectual Property Organization. 2017. ‘Global Brand Database’, www.wipo.int/
branddb/en/. WIPO also provides statistical reports. See World Intellectual Property Organization.
‘Intellectual Property Statistics’, www.wipo.int/ipstats/en/. Japanese trademark records are avail-
able in digital format from the National Center for Industrial Property Information and Training,
www.inpit.go.jp/info/standard/download/standard_dl/sgml5.3.html.
35
TrademarkNow, www.trademarknow.com/.
36
United States Patent and Trademark Office. 2017. ‘Trademark Case Files Dataset’, www.uspto.
gov/learning-and-resources/electronic-data-products/trademark-case-files-dataset-0.

USPTO, beginning in 1870. It is derived from the USPTO main database for administer-
ing trademarks and includes data on mark characteristics and designs, prosecution events,
ownership, classification, renewal history, foreign priority, and international registration.
International bulk data is more difficult to locate. Additionally, the United Kingdom
government has released bulk trademark data.37
The USPTO has also released a bulk dataset of trademark assignments.38 This dataset
comprises documents that parties have filed relating to trademark assignments, securities
interests, releases of security interests, and other related documents. The USPTO dataset
includes information from 1952, and is updated annually. It should be noted that, as
for patents, recording trademark assignments and security interests with the USPTO is
not mandatory. Therefore, the dataset is only a subset of all trademark assignments and
security interests.
Google has also assembled various large databases of trademark information from
the USPTO, which are available for free download.39 The Google data was updated
through 2015 and the file format is eXtensible Markup Language (XML), which may
permit easier manipulation than the bulk USPTO data. One series of data files consists
of the images of trademark word mark, serial number, registration number, filing date,
registration date, goods and services, classification number(s), status code(s), design
search code(s) and pseudo mark(s) from April 7, 1884 to present. Another series of data
files includes the text of trademark applications for the same time period. Google also
provides trademark assignment data (1980–present), as well as Trademark Trial and
Appeal Board decisions (1955–present). Finally, Google provides a dataset with images
of all applications (not just those that mature into registrations) from a recent period,
2010 until present.
III. COPYRIGHTS
Unlike patents and trademarks, copyrights are not substantively examined by the govern-
ment. Copyrights are registered by the U.S. Copyright Office, provided all the paperwork
is properly completed and the required fee is remitted. Currently there is no requirement
that a creator of an expressive work register her copyright. Rights accrue even without
registration, and there is no central repository of copyrighted but unregistered works.
Thus, any investigation of current copyrights that relies upon registrations will be limited
by selection concerns. Registration provides certain benefits, including permitting the
owner to initiate a lawsuit and potentially authorizing special ‘statutory’ damages. Before
1989, and especially before the passage of the 1976 Copyright Act, registration played a
more important role.
37
Gov.UK. ‘Transparency data IPO: Trade mark data release’, www.gov.uk/government/pub
lications/ipo-trade-mark-data-release.
38
United States Patent and Trademark Office. 2017. ‘Trademark Assignment Dataset’, www.
uspto.gov/ip/officechiefecon/tm_assignments.jsp. For a discussion of the dataset, see Graham et al.
(2015).
39
The Google bulk download of USPTO trademark data can be found at Google. 2012.
‘USPTO Bulk Downloads: Trademarks’, www.google.com/googlebooks/uspto-trademarks.html.

Registrations and renewal information for works registered before 1978 are maintained
in physical copy in the Copyright Public Records Reading Room. The U.S. Copyright
Office is presently engaged in a Digitization and Public Access project, with the goal of
providing electronic access to the pre-1978 registrations.40 Although the pre-1978 regis-
trations and other documents are not available through the Copyright Office, there are
other limited sources for the information. For instance, some early information on books,
dramatic compositions, and other works is maintained by the University of Pennsylvania
Libraries.41
For works registered in 1978 or later, the Copyright Office maintains the Copyright
Catalog,42 an electronic searchable database. The Copyright Catalog permits searching by
author, title, keyword, and registration number, and several other fields.43
Many creative works are licensed through The American Society of Composers,
Authors and Publishers (ASCAP). In the U.S., ASCAP does not share much informa-
tion about its collection of copyrighted works. Analogous EU collecting societies often
publish more data.44 Martin Kretschmer and Ruth Towse maintain a wiki of empirical
studies relating to copyrights.45
Information about specific creative works is available. For example, Nielsen has data
on music and book sales, among other things. Nielsen information is expensive to obtain
and we do not believe that it provides free data to academics. Additional information
about music can be found on Discogs46 and Musicbrainz.47 The Internet Movie Database
has information on movies.48 Bowker49 has more information on books, as does Amazon.
IV. TRADE SECRETS AND OTHER INTELLECTUAL

PROPERTY
By definition, information protected by trade secret laws must be secret. It cannot be gen-
erally known or readily ascertainable. Consequently, any public database of trade secret
information would destroy the trade secret status of the information. Thus, we are aware
of no public directory. Information about trade secrets, when available, is typically found
in judicial decisions relating to claims of misappropriation of trade secrets. These cases,
some of which are filed in state courts, can be found on Westlaw, Lexis, BloombergLaw,
40
U.S. Copyright Office. ‘Digitization Project’, http://copyright.gov/digitization/index.html.
41
The Online Books Page: Copyright Registration and Renewal Records, http://onlinebooks.
library.upenn.edu/cce/.
42
U.S. Copyright Office. ‘Copyright Catalog’, http://cocatalog.loc.gov/cgi-bin/Pwebrecon.cgi.
43
Robert Brauneis and Dotan Oliar have assembled the post-1978 copyright registration data
and plan to make it available to researchers in bulk format in the near future.
44
For a list of some of these societies, see Wikipedia. 2017. ‘List of copyright collection socie-
ties’, http://en.wikipedia.org/wiki/List_of_copyright_collection_societies. Martin Kretschmer and
Ruth Towse have been very successful mining data from those sources.
45
www.copyrightevidence.org/evidence-wiki/index.php/Copyright_Evidence.
46
Discogs. 2017, www.discogs.com/.
47
MusicBrainz, https://musicbrainz.org/.
48
IMDb. 2017, www.imdb.com/.
49
Bowker, www.bowker.com/.

and other online litigation databases. As far as we know, other types of IP-like rights, such
as rights of publicity, performance rights, unfair competition, and so forth, are also only
catalogued in reported decisions.
V. LICENSING DATA
IP licensing data is generally limited to those licenses that are disclosed to the Securities
and Exchange Commission (SEC) by public companies as material transactions under the
securities laws. These licenses are available on the EDGAR database for no charge,50 as
well as Lexis and Westlaw, which allow more sophisticated searching than on the SEC’s
website.
Unfortunately, the set of publicly disclosed licenses is much smaller than the total
universe of licenses. Some private companies have collected the licenses in EDGAR, plus
have added other licenses available from other sources, and have made them available for
a fee. These companies include RoyaltySource,51 ktMine,52 RoyaltyStat,53 and Valuation
Resources.54 Royalty Range provides a collection of European licenses.55 The Association
of University Technology Managers provides information on university licenses.56
Moreover, there is scant publicly available information about letters that demand the
recipient to take a license. Chillingeffects has a limited database of patent demand letters,
as provided by some demand letter recipients.57
VI. IP LITIGATION
For federal cases, the Administrative Office of the Courts (AO) provides access to elec-
tronic dockets and related documents on its Public Access to Court Electronic Records
(PACER) system.58 Importantly, PACER only carries those dockets and related docu-
ments that have been electronically logged. We have found that starting in 1999, PACER
carries 95 percent or more of all cases, and nearly 100 percent of cases since the mid-2000s.
However, because many courts did not convert to electronic case filing until the mid-2000s
or later, case documents—other than orders—are widely available only for cases filed in
the mid-2000s or later (and vary by jurisdiction before then). Perhaps more problematic
than PACER’s limited selection of cases and documents are its limitations in searching
50
U.S. Securities and Exchange Commission. ‘EDGAR: Company Filings’, www.sec.gov/edgar/
searchedgar/companysearch.html.
51
RoyaltySource, www.royaltysource.com/.
52
Ktmine. ‘License Agreement Database’, www.ktmine.com/ip-data/license-agreements/.
53
RoyaltyStat. 2017, www.royaltystat.com/.
54
Valuation Resources. ‘Royalty Rates and License Fees’, http://valuationresources.com/Econo
micData/Royalty.htm.
55
Royalty Range, www.royaltyrange.com/.
56
The Association of University Technology Managers. ‘AUTM Licensing Activity Surveys’,
www.autm.net/resources-surveys/research-reports-databases/licensing-surveys/.
57
https://trollingeffects.org/letters.
58
Public Access to Court Electronic Records (PACER), www.pacer.gov/.

for cases. The ‘Case Locator’ allows search by court, case number, subject matter (e.g.,
patent, copyright, trademark, and party name) across all courts.59 However, there is no
means by which to search docket entry text, much less the text of documents filed with
the court. Moreover, PACER charges $0.10 per page of information, which can often add
up quickly, especially when users are downloading multiple documents.
For these reasons, several commercial and non-commercial services provide alternative
access to the PACER data. For instance, the RECAP project (PACER spelled backwards)
provides free access to select PACER dockets and documents.60 PACERPro61 and
DocketAlarm62 provide front-end user interfaces to PACER that allow searching across
cases and dockets by many different fields, including free-text and Boolean-style searches.
Both services generally provide documents already in their databases at no additional
charge and access to any PACER document for an additional, small fee. In a recent
development, the USPTO has released bulk raw data from patent litigation, including the
cases, attorneys, and docket entry descriptions.63
Other services have assembled large amounts of PACER data and resell it on a sub-
scription basis (sometimes with additional fees for downloading documents not on file).
Lexis Courtlink,64 Westlaw CourtExpress,65 and Bloomberg Law Dockets66 each offer
docket-related services with sophisticated searching across cases and related dockets, as
well as some information located in court documents (e.g., patent numbers in CourtLink).
Bloomberg Law Dockets will retrieve many documents from PACER without charge for
academic accounts. Other services which are IP specific—including DocketNavigator,67
Lex Machina, MaxVal IP,68 and RPX69—provide advanced searching across different
types of IP cases (generally since 2000). These services also offer daily email updates of
newly filed cases. Many of these services also provide information about pending Patent
Trial and Appeal Board (PTAB)70 and International Trade Commission (ITC) actions.71
59
PACER, https://pcl.uscourts.gov/search. For discussion of the accuracy of PACER’s nature
of suit field, see Sag (2013) and Kesan and Ball (2006, p. 260, n. 177).
60
Free Law Project. 2017. ‘RECAP Project — Turning PACER Around’, www.recapthelaw.
org/.
61
PacerPro. 2017, www.pacerpro.com/.
62
Docket Alarm. 2017, www.docketalarm.com/.
63
United States Patent and Trademark Office. 2017. ‘Patent Litigation Docket Reports
Data’, www.uspto.gov/learning-and-resources/electronic-data-products/patent-litigation-docket-
reports-data/. Patent numbers and case types for cases filed from 2003 to 2016 will also be
available in this dataset.
64
www.lexisnexis.com/en-us/products/courtlink-for-corporate-or-professionals.page.
65
Thomson Reuters Westlaw. 2017. ‘Dockets’, https://courtexpress.westlaw.com. A scaled-
down version of CourtExpress is available in Westlaw itself via its Westlaw Dockets service.
66
www.bloomberglaw.com/page/law_school#advanced-search/dockets.
67
Docket Navigator. 2015, http://home.docketnavigator.com/.
68
Litigation Databank. 2017, http://litigation.maxval-ip.com/.
69
RPX Insight. 2017, https://search.rpxcorp.com/users/sign_in.
70
PTAB data is also available for no charge at https://ptabdataui.uspto.gov/#/documents.
71
ITC data is also available for no charge at United States International Trade Commission.
2017. ‘Electronic Document Information System (EDIS)’, https://edis.usitc.gov/edis3-external/app.
The ITC also provides summary statistics about its investigations at United States International
Trade Commission, https://usitc.gov.

Some of the services, such as Westlaw, Lexis, and Bloomberg, also make available docket
and related information in select state court cases.
A major limitation of all of these services is that there is no means to download data in
bulk for academic analysis, such as empirical study with STATA, Excel, or other statistical
packages. Docket Navigator provides some downloading capabilities in Excel format on
a free basis for academics. The University of Michigan’s ICPSR service provides limited
federal case information and related coding from the AO for no charge.72 However, ICPSR
recently changed its policy and now requires institutional review board (IRB) approval for
its litigation data, and makes data usage subject to confidentiality restrictions. The Federal
Judicial Center (FJC) recently began providing the same information for free without
restriction.73 Moreover, the reliability of some of the AO data has been questioned.74 A
private consulting group of patent law professors and economists, Academic Expert Group
(AEG), has assembled a large collection of PACER data and related documents on IP litiga-
tion cases, which it makes available for bulk download to academics for research purposes
for a small fee.75 Unlike the ICPSR data, no IRB approval is needed, though the AEG
generally requires that the data be kept confidential other than for peer review purposes.
Direct access to PACER and its documents provides a much larger universe of docu-
ments than the sets of opinions available on Westlaw, Lexis, and Bloomberg. However,
one of the authors has conducted an informal study of summary judgment patent orders
available on PACER and has found that at least since 2010, it appears Westlaw carries
nearly all of the orders that appear on PACER in electronic format. A related issue is the
so-called ‘Rule 36’ orders from the Federal Circuit, which are mere summary affirmances
with no further information.76 In order to know which issues were appealed in these cases,
it is necessary to review the briefs. Although Westlaw has these briefs from 2004 onward,
they do not consistently appear prior to that date. Because the Federal Circuit did not add
electronic case filing until a few years ago, it is necessary to consult at the Federal Circuit
the paper versions of missing briefs filed prior to 2004.
Another problematic coverage issue in these databases is the set of reported litigated
patent numbers. Researchers have relied on Derwent’s ‘LitAlert’ service (available in
Westlaw and Lexis) to provide reported litigated patents.77 This data is derived from
reports sent by district courts to the USPTO of pending litigation. Unfortunately, in
separate work we are undertaking to assemble litigated patent numbers, we have found
that many districts do not regularly report to the USPTO. Thus, the Derwent data sub-
stantially underreports litigated patents. As noted earlier, some services provide searching
by patent number against data derived from PACER (e.g., Courtlink and RPX), but these
72
www.icpsr.umich.edu/icpsrweb/ICPSR/series/72.
73
For lawsuits filed after 1988, the data is available at www.fjc.gov/research/idb/interactive/
IDB-civil-since-1988.
74
See, for example, Beebe (2006, pp. 1652–54) and Hadfield (2004).
75
Please contact the authors of this chapter for further information about AEG’s data.
76
For a discussion of Rule 36 affirmances see Schwartz (2008). The Federal Circuit adopted
Rule 36 in 1989. See Dunner et al. (1995). Even before 1989, there were typically a few Federal
Circuit summary affirmances each year, which also require consulting the briefs to ascertain the
appealed issues.
77
Clarivate Analytics, Westlaw, http://ip-science.thomsonreuters.com/support/patents/dwpiref/
hosts/westlaw/.

data are not available for bulk download. We plan to complete our litigated patent number
database and make it available for research purposes in the near future.
Another area of interest is data about non-practicing entities (NPEs), pejoratively termed
‘patent trolls’. The two main sources of data for researchers has been data from Patent
Freedom78 and RPX.79 Unfortunately, these companies have not allowed general release of
their bulk data for independent verification by other researchers, and subsequent work by
one of the authors along with other researchers has shown that both entities tend to be fairly
aggressive in labeling entities as NPEs.80 This alternative coding of NPEs (for 2010 and
2012) classifies litigants as one of operating companies, individual inventors, failed operat-
ing companies/failed startups, universities, patent holding companies, patent aggregators,
technology development companies, and IP subsidiaries of operating companies. The
full data set is available for full download at the NPE Data website.81 The Stanford NPE
Litigation Dataset, led by Shawn Miller, classifies a random sample of patent plaintiffs since
2000.82 Other data about business entities can be acquired from Compustat,83 Hoovers,84
VentureXpert,85 Wharton Research Data Services,86 and Dun & Bradstreet.87
There are some sources that have coded the underlying data in order to provide more
information and, sometimes, statistics about IP litigation. As mentioned earlier, the AO
data available at the ICPSR service provides additional, hand-coded information, about
IP cases (though, again, some have questioned its reliability). Another site of interest
is PatStats, run by Paul Janicke and the University of Houston School of Law,88 which
provides statistics about patent law issues at the district court and Federal Circuit levels
for cases since 2000.
Other entities provide reports with statistics, such as the Federal Circuit,89 the FJC,90
the Government Accounting Office,91 PricewaterhouseCoopers,92 RPX,93 Lex Machina,94
78
Patent Freedom, www.patentfreedom.com/.
79
RPX. 2017, www.rpxcorp.com/.
80
Some of RPX’s data is now searchable at its portal, but is not available for download in bulk.
See RPX Insight. 2017, https://search.rpxcorp.com/users/sign_in.
81
www.npedata.com.
82
https://law.stanford.edu/projects/stanford-npe-litigation-dataset/. For a description of the
dataset, see Miller (2017).
83
For academics, Compustat is available on the Wharton Research Data Services Wharton
Research Data Services. 2017, https://wrds-web.wharton.upenn.edu/wrds/index.cfm.
84
D&B Hoovers. 2017, http://hoovers.com.
85
VentureXpert is a Thomas Reuters product.
86
Wharton Research Data Services. 2017, https://wrds-web.wharton.upenn.edu/wrds/.
87
Dun & Bradstreet. 2017, www.dnb.com/.
88
Patstats.org, www.patstats.org/.
89
United States Court of Appeals for the Federal Circuit, www.cafc.uscourts.gov/the-court/
statistics.
90
United States Courts, www.uscourts.gov/Statistics/JudicialBusiness/2013.aspx.
91
U.S. Government Accountability Office. 2013. ‘Intellectual Property: Assessing Factors That
Affect Patent Infringement Litigation Could Help Improve Patent Quality’, www.gao.gov/products/
GAO-13-465.
92
PWC. 2017. ‘2017 Patent Litigation Study’, www.pwc.com/us/en/forensic-services/publications/
patent-litigation-study.jhtml.
93
RPX. 2017. ‘Reports’, www.rpxcorp.com/reports/.
94
Lex Machina. 2015, https://lexmachina.com/category/reports/.

Docket Navigator,95 and LegalMetric.96 A very useful report is the American Intellectual
Property Law Association survey, which estimates fees and expenses for a variety of IP
activities, from prosecution to licensing and litigation.97
The underlying cases, especially cases decided by the various courts of appeals, can
be obtained from numerous sources, most notably Westlaw, Lexis, and BloombergLaw.
There are several indexed and coded data sets of reported decisions of IP cases that are
publicly available. For instance, William Mitchell Law School provides a trademark
litigation search engine,98 and the University of Georgia provides a patent litigation
datafile.99 These data sets are useful and required substantial effort on the part of those
who assembled the datasets. Researchers should be aware that since these datasets were
compiled from reported cases from Westlaw, Lexis, United States Patents Quarterly
(USPQ), or BloombergLaw, they constitute only a subset of all litigated cases.
USPQ, in particular, may be particularly unrepresentative of the population of litigated
cases (Rantanen, 2016). USPQ has apparently reported all or substantially all of the his-
toric appellate decisions relating to patents, except for summary affirmances. However, its
coverage of district court decisions is less complete, with some districts and judges largely
absent. Also, like all judicial decisions relating to litigation, judgments on the merits (i.e.,
bench trials and grants of summary judgment) are more likely to be enshrined in a reported
decision than other rulings (i.e., denials of summary judgments). Jury verdicts are also
missing from reported decisions, because they are decided by the jury without a decision by
the judge, unless there is a separate judicial ruling concerning judgment as a matter of law.
Depending upon the research question presented, these selection issues may be paramount.
Westlaw and Lexis appear to presently be comprehensive. However, we believe that
even as recently as 2009, both Westlaw and Lexis were missing a non-trivial number
of unpublished opinions. In the 1999–2000 timeframe, they were missing many more
unpublished opinions. Furthermore, while Westlaw and Lexis both have the complete set
of ‘published’ opinions (those available in F. Supp, F. Supp.2d., F.2d, F.3d, etc.), at least
before 2010, they had somewhat different collections of unpublished opinions. Because
the completeness of these databases has increased over time, researchers should exercise
caution when using them to study time trends.
Data on specific aspects of IP litigation is available on an ad hoc basis. There are too
many of these individual datasets to list them exhaustively here. To provide a small sample
for illustrative purposes, an enormous series of datasets relating to technology standards,
patent pools (including membership and licensees) and industry consortia is available
at the Searle Center of Northwestern Law School.100 Rudi Bekkers, Christian Catalani,
95
Docket Navigator. 2015. ‘Year in Review’, www.docketnavigator.com/document/special/Doc
ket%20Navigator%20Year%20in%20Review%202015.pdf.
96
Legal Metric. 2016, www.legalmetric.com/.
97
American Intellectual Property Law Association. 2017, www.aipla.org/learningcenter/library/
books/econsurvey/Pages/default.aspx.
98
http://app.mitchellhamline.edu/trademark.
99
The UGA Patent Litigation Datafile, http://people.terry.uga.edu/jlturner/patentlitigationdata/.
100
Northwestern Pritzker School of Law. ‘Data on Technology Standards, Industry Consortia,
and Innovation’, www.law.northwestern.edu/research-faculty/searlecenter/innovationeconomics/da
ta/technologystandards/.

Arianna Martinelli, Timothy Simcoe, and Cesare Righi maintain the Disclosed Standard
Essential Patents Database.101 Recent data on fee shifting in patent infringement lawsuits
has been released.102 Professor Barton Beebe has released data on likelihood of confusion
cases in federal trademark litigation, fair use decisions in copyright law, and trademark
dilution cases.103
Our discussion so far has centered on U.S. IP litigation. Data on international IP litiga-
tion is much more difficult to access, especially for American researchers. For example,
DARTS-IP makes available information on international IP litigation for a fee.104
VII. CONCLUSION
A substantial amount of data is available for patents, copyrights, and trademarks,
domestically and internationally, as well as domestic IP litigation. Unfortunately, much
of this data is not accessible for bulk download for research purposes, sometimes making
it time-consuming or costly to acquire. Less data is available for international IP litigation
and licensing, and little to no data—other than reported decisions—is available for trade
secrets and other IP rights. These limitations have made it difficult for researchers to
perform comprehensive empirical studies. We hope that this chapter has illuminated some
of the pockets of data available to researchers, and these pockets will expand over time
so as to provide a wide array of comprehensive data to answer fundamental questions in
the field of IP.
REFERENCES
American Intellectual Property Law Association. 2017, www.aipla.org/learningcenter/library/books/econsurvey/

Pages/default.aspx.
Beebe, Barton. 2006. ‘An Empirical Study of the Multifactor Tests for Trademark Infringement’, 94 California
Bloomberg Law. 2017. www.bloomberglaw.com/patent_search.
Bowker, www.bowker.com/.
Clarivate Analytics, Westlaw, http://ip-science.thomsonreuters.com/support/patents/dwpiref/hosts/westlaw/.
Clarivate Analytics. 2017. ‘Derwent Innovation’, http://info.thomsoninnovation.com/.
Crouch, Dennis. 2015. ‘Data: Counting Fee Shifting Cases’, https://patentlyo.com/patent/2015/05/counting-
shifting-cases.html.
D&B Hoovers. 2017, http://hoovers.com.
darts-ip, www.darts-ip.com/world/.
de Rassenfosse, Gaétan, Dernis, Hélène and Boedt, Geert. 2014. ‘An Introduction to the Patstat Database with
Example Queries’, 47 Australian Economic Review 395–408.
101
http://ssopatents.org/
102
For data from 2015 see Crouch (2015). Older data is also available from Research on
Innovation, www.researchoninnovation.org/data.html.
103
www.bartonbeebe.com/. The site includes the data used in ‘An Empirical Study of U.S.
Copyright Fair Use Opinions, 1978–2005’, 156 Penn. L. Rev. 549 (2008); ‘The Continuing Debacle
of U.S. Antidilution Law: Evidence from One Year of Trademark Dilution Revision Act Case
Law’, 24 Santa Clara Comp. & High Tech. L.J. 449 (2008); ‘An Empirical Study of the Multifactor
Tests for Trademark Infringement’, 95 Cal. L. Rev. 1581 (2006).
104
darts-ip, www.darts-ip.com/world/.

Discogs. 2017, www.discogs.com/.

Docket Alarm. 2017, www.docketalarm.com/.
Docket Navigator. 2015, http://home.docketnavigator.com/.
Docket Navigator. 2015. ‘Year in Review’, http://home.docketnavigator.com/year-review/.
Dun & Bradstreet. 2017, www.dnb.com/.
Dunner, Donald R., Jakes, J. Michael and Karceski, Jeffrey D. 1995. ‘A Statistical Look at the Federal Circuit’s
Patent Decisions: 1982–1994’, 5 Federal Circuit Bar Journal 151–80.
European Patent Office. 2017. ‘Espacenet Patent search’, http://worldwide.espacenet.com/.
European Patent Office. 2017. ‘Patstat’, www.epo.org/searching/subscription/raw/product-14-24.html.
European Policy for Intellectual Property, www.epip.eu/datacentre.php.
Frakes, Michael D. and Wasserman, Melissa F. 2013. ‘Does Agency Funding Affect Decisionmaking?: An
Empirical Assessment of the PTO’s Granting Patterns’, 66 Vanderbilt Law Review 67–125.
Frakes, Michael D. and Wasserman, Melissa F. 2017. ‘Is the Time Allocated to Review Patent Applications
Inducing Examiners to Grant Invalid Patents? Evidence from Microlevel Application Data’, 99(3) The Review
of Economics and Statistics 550–63.
Free Law Project. 2017. ‘RECAP Project—Turning PACER Around’, www.recapthelaw.org/.
Free Patents Online, freepatentesonline.com.
Google Patents, https://patents.google.com/.
Google. 2012. ‘USPTO Bulk Downloads: PAIR Data, https://www.google.com/googlebooks/uspto-patents-
pair.html.
Google. 2012. ‘USPTO Bulk Downloads: Patent Assignment Text’, https://www.google.com/googlebooks/
uspto-patents-assignments.html.
Google. 2012. ‘USPTO Bulk Downloads: Patent Classification Information’, https://www.google.com/google
books/uspto-patents-class.html.
Google. 2012. ‘USPTO Bulk Downloads: Patent Maintenance Fees’, https://www.google.com/googlebooks/
uspto-patents-maintenance-fees.html.
Google. 2012. ‘USPTO Bulk Downloads: Patent Petition Decisions’. https://www.google.com/googlebooks/
uspto-patents-petitions.html.
Google. 2012. ‘USPTO Bulk Downloads: Trademarks’, https://www.google.com/googlebooks/uspto-tradema
rks.html.
Google. 2017. ‘About Google Patents’, https://support.google.com/faqs/answer/2539193.
Gov.UK. ‘Search for a trade mark’, www.gov.uk/search-for-trademark.
Gov.UK. ‘Transparency data IPO: Trade mark data release’, www.gov.uk/government/publications/ipo-trade-ma
rk-data-release.
Government of Canada. 2017. ‘Canadian Trademarks Database’, www.cipo.ic.gc.ca/app/opic-cipo/trdmrks/
srch/bscSrch.do.
Graham, Stuart J.H., Marco, Alan C. and Miller, Richard. 2015. ‘The USPTO Patent Examination Research
Dataset: A Window on the Process of Patent Examination’, https://ssrn.com/abstract=2702637.
Graham, Stuart J.H., Marco, Alan C. and Myers, Amanda F. 2015. ‘Monetizing Marks: Insights from the
USPTO Trademark Assignment Dataset’, https://papers.ssrn.com/sol3/papers.cfm?abstract_id=2430962.
Hadfield, Gillian. 2004. ‘Where Have All the Trials Gone? Settlements, Nontrial Adjudications and
Statistical Artifacts in the Changing Disposition of Federal Civil Cases’, 1 Journal of Empirical Legal
Studies 705–34.
Hall, Bronwyn H., Jaffe, Adam B. and Trajtenberg, Manuel. 2001. ‘The NBER Patent Citation Data File:
Lessons, Insights and Methodological Tools’. National Bureau of Economic Research, Centre for Economic
Policy Research Discussion Paper Series, No. 3094.
http://app.wmitchell.edu/trademark.
http://patentadvisor.com/.
http://rosencrantz.berkeley.edu/batchsql/downloads.
www.bartonbeebe.com/.
www.bna.com/bloomberglaw-dockets/.
www.npedata.com.
www.patentcore.com/.
https://courtlink.lexisnexis.com.
https://ptabtrials.uspto.gov/.
www.chillingeffects.org.
www.icpsr.umich.edu/icpsrweb/landing.jsp.
IMDb. 2017, www.imdb.com/.
Kesan, Jay P., and Ball, Gwendolyn G. 2006. ‘How Are Patent Cases Resolved—An Empirical Examination of
the Adjudication and Settlement of Patent Disputes’, 84 Washington University Law Review 237–312.
Ktmine. ‘License Agreement Database’, www.ktmine.com/ip-data/license-agreements/.

Legal Metric. 2016, www.legalmetric.com/.

Lex Machina. 2015, https://lexmachina.com/category/reports/.
Lu, Qiang, Myers, Amanda F. and Beliveau, Scott. 2017. ‘The USPTO Patent Prosecution Resaerch Data:
Unlocking Office Action Traits,’ USPTO Economic Working Paper No. 2017-10, https://papers.ssrn.com/
sol3/papers.cfm?abstract_id=3024621.
Litigation Databank. 2017, http://litigation.maxval-ip.com/.
Marco, Alan C., Carley, Michael, Jackson, Steven and Myers, Amanda F. 2015. ‘The USPTO Historical Patent
Data Files: Two Centuries of Innovation’, USPTO Economic Working Paper, No. 2015-1, www.uspto.gov/
sites/default/files/documents/USPTO_economic_WP_2015-01_v2.pdf.
Miller, Shawn P. et al., 2017. ‘Introduction to the Stanford NPE Litigation Dataset’, https://law.stanford.edu/
publications/introduction-to-the-stanford-npe-litigation-dataset/.
MusicBrainz, https://musicbrainz.org/.
National Bureau of Economic Research. ‘Patent Data Project’, https://sites.google.com/site/patentdataproject/
Home.
Northwestern Pritzker School of Law. ‘Data on Technology Standards, Industry Consortia, and Innovation’,
www.law.northwestern.edu/research-faculty/searlecenter/innovationeconomics/data/technologystandards/.
OECD. 2017. ‘Intellectual property (IP) statistics and analysis’, www.oecd.org/science/inno/oecdpatentdatabases.
htm.
Pacer, https://pcl.uscourts.gov/search.
PacerPro. 2017, www.pacerpro.com/.
Patent Freedom, www.patentfreedom.com/.
Patstats.org, www.patstats.org/.
Public Access to Court Electronic Records (PACER), www.pacer.gov/.
Public Citizen. 2017. ‘How to File a FOIA Request: A Guide’, http://citizen.org/Page.aspx?pid=458.
PwC. 2017. ‘2017 Patent Litigation Study’, www.pwc.com/us/en/forensic-services/publications/patent-litigation-
study.jhtml.
Rantanen, Jason. 2016. ‘Empirical Analysis of Judicial Opinions: Methodology, Metrics, and the Federal
Circuit’, 49 Connecticut Law Review 227–91.
ReedTech. 2017. ‘Public PAIR Patent Access Download’, http://patents.reedtech.com/Public-PAIR.php.
ReedTech. 2017. ‘USPTO Data Sets’, http://patents.reedtech.com/pgog.php.
Research on Innovation, www.researchoninnovation.org/data.html.
Royalty Range, www.royaltyrange.com/.
RoyaltySource, www.royaltysource.com/.
RoyaltyStat. 2017, www.royaltystat.com/.
RPX. 2017, www.rpxcorp.com/.
RPX. 2017. ‘Reports’, www.rpxcorp.com/reports/.
Sag, Matthew. 2013. Empirical Studies of Copyright Litigation: Nature of Suit Coding. Loyola University Chicago
School of Law Research Paper, No. 2013-017, https://papers.ssrn.com/sol3/papers.cfm?abstract_id=2330256.
Schwartz, David L. 2008. ‘Practice Makes Perfect? An Empirical Study of Claim Construction Reversal Rates
in Patent Cases’, 107 Michigan Law Review 223–84.
The Association of University Technology Managers. ‘AUTM Licensing Activity Surveys’, www.autm.net/resou
rces-surveys/research-reports-databases/licensing-surveys/.
The Online Books Page: Copyright Registration and Renewal Records, http://onlinebooks.library.upenn.edu/
cce/.
The Patent Box. 2017, www.thepatentbox.com/.
The UGA Patent Litigation Datafile, http://people.terry.uga.edu/jlturner/patentlitigationdata/.
Thomson Reuters Westlaw. 2017. ‘Dockets’, https://courtexpress.westlaw.com.
TrademarkNow, www.trademarknow.com/.
U.S. Copyright Office. ‘Copyright Catalog’, http://cocatalog.loc.gov/cgi-bin/Pwebrecon.cgi.
U.S. Copyright Office. ‘Digitization Project’, http://copyright.gov/digitization/index.html.
U.S. Government Accountability Office. 2013. ‘Intellectual Property: Assessing Factors That Affect Patent
Infringement Litigation Could Help Improve Patent Quality’, www.gao.gov/products/GAO-13-465.
U.S. Securities and Exchange Commission. ‘EDGAR: Company Filings’, www.sec.gov/edgar/searchedgar/
companysearch.html.
United States Court of Appeals for the Federal Circuit, www.cafc.uscourts.gov/the-court/statistics.html.
United States Courts, www.uscourts.gov/Statistics/JudicialBusiness/2013.aspx.
United States Department of Justice. 2011. ‘FOIA.Gov: Make a Request’, www.foia.gov/report-makerequest.
html.
United States International Trade Commission, https://usitc.gov.

United States International Trade Commission. 2017. ‘Electronic Document Information System (EDIS)’,
https://edis.usitc.gov/edis3-external/app.
United States Patent and Trademark Office. ‘Trademark Electronic Search System (TESS)’, http://tess2.uspto.
gov/.
United States Patent and Trademark Office. 2017. ‘Bulk Data Products’, www.uspto.gov/learning-and-resources/
electronic-bulk-data-products.
United States Patent and Trademark Office. 2017. ‘Historical Patent Data Files’, http://www.uspto.gov/learning-
and-resources/electronic-data-products/historical-patent-data-files.
United States Patent and Trademark Office. 2017. ‘Patent Application Full Text and Image Database’, http://appft.
uspto.gov/netahtml/PTO/search-adv.html.
United States Patent and Trademark Office. 2017. ‘Patent Application Information Retrieval’, http://portal.uspto.
gov/external/portal/pair.
United States Patent and Trademark Office. 2017. ‘Patent Litigation Docket Reports Data’, www.uspto.gov/
learning-and-resources/electronic-data-products/patent-litigation-docket-reports-data/.
United States Patent and Trademark Office. 2017. ‘Portal Applications’, https://ppair.uspto.gov/authenticate/
AuthenticateUserLocalEPF.html.
United States Patent and Trademark Office. 2017. ‘Statistics’, www.uspto.gov/learning-and-resources/statistics.
United States Patent and Trademark Office. 2017. ‘Trademark Assignment Dataset’, www.uspto.gov/ip/officechie
fecon/tm_assignments.jsp.
United States Patent and Trademark Office. 2017. ‘Trademark Case Files Dataset’, www.uspto.gov/learning-and-re
sources/electronic-data-products/trademark-case-files-dataset-0.
United States Patent and Trademark Office. 2017. ‘US Patent Full-Page Images’, http://patft.uspto.gov/netahtml/
PTO/patimg.htm.
United States Patent and Trademark Office. 2017. ‘USPTO Patent Full-Text and Image Database’, http://patft.
uspto.gov/netahtml/PTO/search-adv.htm.
Valuation Resources. ‘Royalty Rates and License Fees’, http://valuationresources.com/EconomicData/Royalty.
htm.
Wharton Research Data Services. 2017, https://wrds-web.wharton.upenn.edu/wrds/.
Wharton Research Data Services. 2017, https://wrds-web.wharton.upenn.edu/wrds/index.cfm.
Wikipedia. 2017. ‘List of copyright collection societies’, http://en.wikipedia.org/wiki/List_of_copyright_collec
tion_societies.
World Intellectual Property Organization. ‘Intellectual Property Statistics’, www.wipo.int/ipstats/en/.
World Intellectual Property Organization. 2017. ‘Global Brand Database’, www.wipo.int/branddb/en/.
World Intellectual Property Organization. 2017. ‘Patentscope’, https://patentscope.wipo.int/search/en/search.
jsf.
Legislative materials
Freedom of Information Act, 5 U.S.C. § 552.

1976 Copyright Act, Pub. L. No. 94-553, 90 Stat. 2541.

PART II
EMPIRICAL STUDIES
RELATING TO PATENTS

Section A
Metrics

2. Patent citation data in social science research:
overview and best practices* 105
Adam B. Jaffe * and Gaétan de Rassenfosse‡**

†
Contents
I. Introduction
II. Citations as an Indicator of Invention Attributes
A. Forward Citations and Technological Impact
B. Backward Citations and Reliance on Previous Technology
C. Technological Distance and Diversity
D. Linkage to Science
E. Economic Value
III. Citations as an Indicator of Knowledge Flows
A. Geographic Dimension of Knowledge Flows
B. Temporal Dimension of Knowledge Flows
C. Validation Studies
IV. Citations as Links in Knowledge or Innovation Networks
A. Mapping Patents, Individuals, Institutions, and Regions
B. Mapping of Technological Trajectories
V. Pitfalls and Best Practices in Use of Citation-Based Indicators
A. Office Effects
B. Time and Technology Field Effects
C. Examiner Effects
D. Strategic Effects
VI. Conclusion
References
Appendix
Knowledge flows […] are invisible; they leave no paper trail by which they may be measured and
tracked, and there is nothing to prevent the theorist from assuming anything about them that she
likes.
Paul Krugman (1991)
* We thank David Schwartz for suggesting this project. We received helpful comments from
Bronwyn Hall, Dietmar Harhoff, Sadao Nagaoka, and Beth Webster, as well as participants at the
2015 Workshop on the Economics of Intellectual Property in Northwestern University. Jan Kozak
provided valuable research assistance. The authors are solely responsible for all opinions or errors.
†
* Motu Economic and Public Policy Research, Level 1, 97 Cuba St, Wellington 6011 New
Zealand; Queensland University of Technology: Te Pūnaha Matatini Centre of Research Excellence.
‡
** Brandeis University, 415 South St, Waltham MA 02453 USA; Queensland University of
Technology.
20

Patent citation data in social science research 21
I. INTRODUCTION
Eugene Garfield is one of the pioneers of the study of citation data. In his 1955 article,
Garfield proposes to build a citation index for scientific articles in order to make it pos-
sible for “the conscientious scholar to be aware of criticisms of earlier papers.” He further
explains, “even if there were no other use for a citation index than that of minimizing
the citation of poor data, the index would be well worth the effort required to compile
it” (p. 108). It turns out that citation indices have been used in a variety of ways and for
a variety of purposes. Two of the most notable uses are to assess the attributes of the
idea embedded in a scientific article and to track its diffusion through time, space and
technology domains. In fact, Garfield (1955) foresaw these two uses as he described the
citation index as an “association-of-ideas index” (p. 108) and explained that the citation
index may “help the historian to measure the influence of the article—that is, its ‘impact
factor’” (p. 111).
While the analogy with the broader field of bibliometrics may seem obvious, patent
citations differ from scientific citations in substantial ways. Citations in patents are the
results of a highly mediated process that involves multiple parties: the inventor, the patent
attorney, and the patent examiner (Meyer, 2000). These parties have different incentives
for citing publications and may do so at different times and in different sections of the
patent document (Cotropia, Lemley, and Sampat, 2013). Much of the empirical research
relies on US citations, but there are important differences across jurisdictions in citation
rules and practice.1 This creates interesting opportunities for research on non-US data,
but also suggests a degree of caution in thinking about the global implications of results
based solely on US data.
The widespread use of patent citations in social science research can be traced to the
availability of patent statistics in digitally readable form in the late 1970s.2 Zvi Griliches
(1979), in his important manifesto for research on R&D and productivity growth, sug-
gested that the frequency with which patents from different industries cite each other
could be used as a measure of the technological proximity of industries. An early strand
of research on patent citations was the work of Francis Narin and his associates at CHI
Research, Inc. (Carpenter, Narin, and Woolf, 1981; Carpenter and Narin, 1983; Narin
and Noma, 1985; Narin, Noma, and Perry, 1987). An influential early demonstration of
the potential utility of patent citation data in economic research was the PhD research of
Griliches’ student Manuel Trajtenberg (Trajtenberg, 1990a, 1990b). The use of patent cita-
tion data has grown dramatically over the last two decades, as illustrated in Appendix A.
What makes citations potentially useful is that they convey information about the
cumulative nature of the research process, as well as information about the consequences.
1
The present survey also discusses evidence on citations at the European Patent Office
whenever available.
2
The earliest reference that we found is Clark (1976). It presents statistics on the obsolescence
of US Patent and Trademark Office patents using citation data. Garfield (1966) discusses the use of
patent citation searches to say something about the significance of a patent, but it does not present
any systematic analyses or statistics. Kuznets (1962) did not specifically discuss citations, but did
emphasize that patent documents are a rich and deep source of information on the inventive pro-
cess, and urged that this richness be exploited in addition to researchers’ simply counting patents.

While some inventors and research organizations pursue patents for motives of prestige or
internal tracking of research success, most patent applications are made with the goal of
securing commercial advantage, or at least preserving options for pursuit of commercial
advantage. Another virtue of patent data for social science research is that patents reside
in a non-market-based technological classification system, allowing one to place patents,
inventors and organizations in technology space in a way that is not derived from sales or
other economic data that one may be trying to relate to invention.3 Furthermore, the clas-
sification scheme is hierarchical so that technology categories can be very fine or relatively
broad as desired. This feature, and others, has been combined with patent citation data
to provide powerful indicators.
This chapter provides an overview of the major uses of such data and the issues that
arise in such research. Other authors have previously discussed the use of patent statistics
in social science research (e.g., Griliches, 1990; Lerner and Seru, 2015), and Gay and Le
Bas (2005) provide a brief overview of the use of patent citations to measure invention
value and knowledge flows. However, we are not aware of a broad survey on the use of
patent citation data.4 In order to identify the papers to include in this survey, we started
from a limited number of references that we were aware of and complemented those using
a keyword-based search on Google Scholar. We then expanded this core of references by
looking at cited and citing references. Ultimately, we kept the most influential articles,
either in terms of the number of citations received or in terms of relevance of the find-
ings. The majority of papers are published in economics, management, and information
science journals.
Conceptually, we classify research using patent citations into two broad groups. One
research line uses a variety of citation-based statistics to characterize the inventions, in
terms of the magnitude and nature of their impact, as well as the nature and magnitude
of the departure that they represent relative to the existing pool of knowledge. This work
is discussed in the next section. The other research line focuses on the citations themselves,
using them as proxies for knowledge linkages across inventors in order to explore the
nature of knowledge flows and the factors that affect those flows. This research is discussed
first with regard to relatively simple metrics of knowledge flow in Section III, and then with
respect to attempts to map interactions in a more complex network framework in Section
IV. We then provide some brief comments on practical difficulties and pitfalls in using
citations data in Section V. Section VI concludes with opportunities for future research.
II. CITATIONS AS AN INDICATOR OF INVENTION

ATTRIBUTES
There is no agreed-upon model of inventions and the inventive process, which leads to
some ambiguity in how citation metrics are interpreted. Nonetheless it is possible to
3
Jacob Schmookler pointed out that in a patent subclass “Dispensing of semi-solid materi-
als,” he found a patent for a manure spreader and another for a toothpaste tube (Schmookler,
1966).
4
Jaffe and Trajtenberg (2002) reprint 12 of the key papers on patent citations by them and
their co-authors.

identify two broad aspects of the process that underlie citation-based inferences. First,
we can think of all possible technologies as mapping onto a high-dimensional technology
space, such that a given invention can be located in that space, and a patent represents
the right to exclude others from marketing products that impinge upon a specified region
(or regions) of that space. Second, the invention process is cumulative, that is, inventions
build on those that came before and, in turn, facilitate those that come after. In this
“geometric” interpretation, the patent claims delineate the metes and bounds of the
region of technology space over which exclusivity is being granted, while the citations
indicate previously marked off areas that are in some sense built upon by or connected to
the invention being granted.
Thus the citations that appear in a patent (its “backward” citations) inform us about
the technological antecedents of the patented invention. A patent that contains many
citations corresponds to an invention with many antecedents; a patent whose citations
are to technologically diverse previous patents has diverse antecedents; a patent whose
citations are to old patents corresponds to an invention with old antecedents, and so
forth. Conversely, the citations received by a patent from subsequent patents (“forward”
citations) inform us about the technological descendants of the patented invention. A
patent that is never cited was a technological dead end. A patent with many or technologi-
cally diverse forward citations corresponds to an invention that was followed by many or
technological diverse descendants.
Note that the discussion so far is entirely definitional. We have said nothing about
the possibility of causal connections between these different attributes of inventions, or
between any of these attributes and the private or social value of the invention. Ultimately,
we are interested in whether, for example, patents with relatively few technological
antecedents are more or less likely to spawn multiple lines of research, or whether patents
that generate many or diverse technological descendants correspond to inventions that
generate large social benefits. It is in large part to be able to say something about these
questions that citation metrics have been developed. In a very broad sense, citation
analysis is predicated on an expectation that the extent and nature of an invention’s
antecedents tell us something about the novelty or “radicalness” of the invention, and the
extent and nature of its descendants tell us something about both its technological impact
and its economic value. But different authors propose or use different characterizations
of citation information to elucidate these ideas.
In practice, writers are not always clear on the underlying concept that a given metric is
intended to measure, and given metrics are used in different contexts as proxies or indica-
tors for different concepts. In some cases, researchers postulate a relationship between
a given citation metric and an underlying concept, and then test hypotheses about the
concept taking that relationship as a given. In other cases, researchers attempt explicitly
to validate the extent to which a given metric reflects a particular underlying conceptual
attribute of inventions. We will consider these different approaches below in the context of
specific papers, but for expositional purposes it is useful to consider five broad categories
of approaches:
1. counts of forward citations as an indicator of subsequent technological impact;

2. counts of backward citations as an indicator of the extent of reliance on previous
technology;

3. characterization of both backward and forward citations in terms of technological

diversity and technological distance;
4. examination of references to non-patent literature as an indicator of science linkage;
and
5. use of citations as an indicator for private and social value.
We consider each category in turn.
A. Forward Citations and Technological Impact
Using the number of forward citations as a measure of technological impact of a patented

invention can be motivated by direct analogy to the larger and pre-existing bibliometric
literature starting with Garfield (1955). Nonetheless, Trajtenberg, Henderson and Jaffe
(1997) undertook to demonstrate the validity of this (and other) metrics by comparing
the citation rate to university patents and corporate patents, based on a maintained
assumption that university patents are more “basic” and hence have, on average, greater
technological impact. To incorporate the cumulative nature of invention into the metric,
they proposed that the importance of an invention be characterized by the number of
forward citations received, plus a fractional weight multiplied by the number of citations
received by those citing patents. That is, important patents are those that are cited a
lot and are cited by patents that are themselves relatively highly cited.5 The authors
showed that importance by this definition is indeed higher for university patents than for
corporate patents, using a sample of patents assigned to US corporations, matched by
patent class and grant date to patents assigned to US universities. In addition, they discuss
qualitatively the highest-importance patents in their sample, and argue that the citing
patents can be seen as technological descendants, and these highly “important” patents
are indeed subjectively very important in their respective fields.
More recently, taking advantage of improvements in computing power, scholars have
taken into account the whole stream of citations. For example, Lukach and Lukach (2007)
have proposed computing importance by the PageRank score of patents. This method is
directly inspired from Google’s “random surfer” model and takes into account the fact
that different citations weigh differently depending on the importance of the citing docu-
ments (Brin and Page, 1998). However, the authors are not able to validate their ranking
using external measures such that the conditions under which the PageRank method is
more appropriate than a straightforward citation count are unclear. This approach is a
natural extension of earlier work and begins to move this line of analysis towards the
“innovation network” formulation discussed later in the text.
Albert, Avery, Narin, and McAllister (1991) provide a validation study of the use of
forward citations as an indicator of impact. They reported a strong correlation between
5
The authors report “forward importance” as the number of citations received plus .5 times
the number of citations received by the citing patents, and undertook sensitivity analysis varying
this weight between .25 and .75. Extending this throughout the citation tree involves a geometri-
cally declining weight—for example, if patent E cites patent D which cites patent C which cites
patent B which cites patent A, we might consider patent B to contribute 1 to the importance of A,
patent C .5, patent D .25 and patent E .125.

the citation intensities of 77 Kodak silver halide patents and expert evaluations of techni-
cal impact and importance of the patents. Narin (1995) showed that patents that have
attained the legal status of pioneering patents in the United States, as well as other promi-
nent patents appearing in such patent office publications as “Hall of Fame” patents, are
very highly cited. Czarnitzki, Hussinger and Schneider (2011) relate a group of “wacky”
patents to control groups and test the extent to which commonly used metrics are able to
identify wacky patents from patents in the control group. Wacky patents are selected by
an employee of the World Intellectual Property Organization “for their futile nature, as
they do not involve a high-inventive step or only marginally satisfy the ‘non obviousness’
criterion” (p. 131). They find that the number of forward citations is a good predictor
of importance. However, other measures such as originality and generality (discussed
below) were higher for wacky patents. Another interesting confirmation of patent cita-
tions as indicative of technological impact is Benson and Magee (2015). They identify 28
“technological domains” (e.g., “Solar Photovoltaics” or “Genome Sequencing”) in which
it is possible to identify a specific metric of the technological state of the domain (e.g.,
watts/$ for Solar Photovoltaics). They take the exponential rate of improvement of these
metrics across domains and across time as the dependent variable in regressions on vari-
ous citation metrics of patents in the technology domain. They find that forward citations
are positively related, and the average age of backward citations negatively related, to the
rate of improvement of the technology over the subsequent ten-year period.
B. Backward Citations and Reliance on Previous Technology
While it seems clear that important inventions generate more forward citations, the oppo-
site may hold for backward citations. That is, more trivial inventions are more extensively
rooted in what has come before, while more basic inventions are less incremental in nature
and thus have fewer identifiable antecedents (Trajtenberg, Henderson, and Jaffe, 1997).
Another way to think of this is that a patent will, to some extent, tend to cite other patents
all the way back along the inventive trajectory upon which it lies. Patents that are near the
beginning of a trajectory are in this sense more basic, and may be expected to make fewer
backward citations because they have less historical background.
Empirical evidence is rather inconclusive. Trajtenberg, Henderson, and Jaffe (1997)
find that university patents (presumably more important than the average patent) do
make fewer citations and cite patents that are themselves less highly cited. However, von
Wartburg, Teichert, and Rost (2005) provide a different view. They correlate a measure
of backward citations with expert ratings on the technological value added (in the form
of technical scoring tables) of 107 patents related to four strokes internal combustion
engines. Their backward citations measure counts first and second-generation citations
received. They obtain a statistically significant correlation coefficient of 0.38, implying
that patents with higher technological value added build on more references. Liu, Hseuh,
Lawrence, Meliksetian, Perlich, and Veen (2011) propose a more in-depth analysis of
backward references and patent value. They correlate the number of backward references
with the probability that a patent will stand up in court and find a statistically strong
positive association. Overall, it is unclear whether the number of backward citations
captures patent importance.

C. Technological Distance and Diversity
As noted above, one of the basic virtues of patent data is that they provide a non-market-
based technological classification system for inventions. Looking at the way in which
citations span the technology space defined by the classification scheme is a natural
way to characterize the technological complexion of both an invention’s roots and its
impacts. Broadly speaking, there are two major aspects to be considered, whether looking
forward or backward. One is pure distance: how technologically different are the patents
connected by a citation link—for example, does a drug patent cite other patents for
compounds in the same chemical class, or patents on other chemicals, or mechanical or
electronic patents? The other is breadth or diversity: independent of whether that drug
patent generally cites other patents that are close to or far from itself, are they all bunched
together in technology space, or are they dispersed far from each other?
Trajtenberg, Henderson, and Jaffe (1997) implement a measure of technological distance
using a three-level representation of the US Patent and Trademark Office (USPTO) patent
classification system. The lowest level used is the three-digit original patent class (e.g., Electric
Lamp and Discharge Devices); the next level is the set of two-digit c ategories (e.g., Electrical
Lighting); the highest level is six very broad fields (e.g., Electrical and Electronic). The
authors axiomatically set two patents in the same patent class at distance 0; two that are in
different classes but the same category at distance .33; two that are in different categories but
the same broad field as distance .66; and two that are not even in the same field as distance 1.
They then calculate the average distance over both forward and backward citations for each
patent in the university and corporate samples. As expected, they found that the forward
citations received by university patents came, on average, from farther away in technology
space, although the difference was small and not always statistically significant. For back-
ward citations, there was no consistent pattern—that is, university patents did not systemati-
cally cite earlier patents that were, on average, technologically more distant by this metric.
To measure technological dispersion or diversity, Trajtenberg, Henderson, and Jaffe
(1997) proposed 1 minus the Herfindahl-Hirschman Index of concentration of the cita-
tions across patent classes—that is, 1 minus the sum of squared shares of citations in each
class. This metric is equal to zero if all citations are in the same class, and it approaches
unity as the citations are spread thinly across all classes. The authors dubbed this metric
of diversity “generality” when applied to forward citations, and “originality” when
applied to backward citations.6, 7 They conjectured that both measures should be larger
6
For a small number of citations, it is clear that this measure is heavily influenced by the number
of citations—for example, a patent receiving only two citations cannot possibly have generality
greater than .5. Whether or not this is a problem is largely a matter of interpretation; in some sense
it is meaningful to say that a patent receiving only two citations cannot have a very diverse impact. A
different interpretation is that every invention has a latent or unobserved generality that is randomly
realized in the citations it happens to receive. Under this formulation, the distribution of citations
across patent classes is multinomial, and the observed generality or originality is a biased estimator
of the true parameter. Bronwyn Hall has derived a formula to correct for this bias (Hall, Jaffe and
Trajtenberg, 2001); it produces a significant correction for patents with just a few citations.
7
Ziedonis (2004) has built on this idea to construct a measure of the fragmentation of owner-
ship rights to a firm’s complementary patents. Backward citations are stratified by assignee instead
of technology class.

for more basic inventions, and therefore expected to be larger for university patents than
for corporate patents. This hypothesis was borne out in the data for generality measure,
but not for originality.
A concept related to generality is that of “General Purpose Technology” (GPT). GPTs
are conceived as technologies that subsequently connect to many different application or
development technologies to allow multiple lines of technology innovation and diffusion.
Frequently mentioned examples are the electric motor in the late 19th and early 20th cen-
turies, and digital information technology in the late 20th century. Hall and Trajtenberg
(2006) use data from a selected sample of 780 most highly cited patents that were granted
by the USPTO in the years 1967–99 to construct generality, number of citations, and
patent class growth, for both cited and citing patents, intended to identify GPTs in their
early stages. The paper finds that highly cited patents differ in almost all respects from the
population of all patents (they take longer to be issued; have twice as many claims; are
more likely to have a US origin; are more likely to be assigned to a US corporation; are
more likely to have multiple assignees; have on average higher citation lags; have a higher
generality; and are in patent classes that are growing faster than average). The paper
concludes that the identified measures, although promising, give contradictory messages
when taken separately and that it is not obvious how to combine those measures to choose
a sample of GPT patents.8 The fundamental difficulty is that we don’t have measures of
how general-purpose a technology is other than broad conceptions of GPT technologies.
Thus, while it seems plausible that general-purposeness would be reflected in citation
patterns, it is hard to pin such patterns down or test their validity.9
Youtie, Iacopetta, and Graham (2008) found that nanotechnology patents from
1990–93 were more general than computer patents and much more general than drug
patents, and interpreted this result as evidence that nanotechnology is an emerging GPT.
Moser and Nicholas (2004), however, found that electricity patents from the 1920s were
less general and less highly cited than chemical and mechanical patents from the same
period, suggesting that the relationship between the characteristics that make a technol-
ogy a GPT and other characteristics of inventions is complex.
Another concept related to technological distance and diversity is that of a “radical” or
“breakthrough” invention. Ahuja and Lampert (2001) propose that radical inventions are
simply the top 1 percent of patents ranked on citations received in a given year. Dahlin and
Behrens (2005) adopt a more sophisticated approach. They conceive a “radical” invention
within a given technology domain (tennis rackets, in their application) to be one that
recombines previous technology elements in a new and different way, but which is then
imitated and so spawns subsequent patents that combine technology elements in a manner
8
Hall and Trajtenberg (2006) explain that the generality measures suffer from the fact that
they treat citations from patents in patent classes different from the cited patents in the same way,
although some patent classes are very different and some are closely related. They suggest that the
future research could construct a weighted generality measure, with weights inversely related to
the overall probability that one class cites another class. To the best of our knowledge, no one has
implemented such an approach.
9
Hall and Trajtenberg (2006) also show that a disproportionate share of the patents in
the extreme upper tail of the distribution for generality and total forward citations in the period
1967–99 are IT patents, suggesting that these metrics may be indicative of a GPT.

substantially similar to the radical invention. They construct a measure of the “overlap”
in the respective sets of patents cited by two different patents, and show that the radical
inventions (oversized and wide-body rackets, in their application) had little overlap with
previous or contemporary patents, but significant overlap with patents that came after.
D. Linkage to Science
As discussed, patents contain references to non-patent documents, the overwhelming

majority of which are scientific papers. On this basis, the number of non-patent backward
citations made by a patent, or the fraction of backward citations that these non-patent
citations represent, has been explored as a metric of the closeness of linkage between an
invention and scientific research.10
Collins and Wyatt (1988) looked at citations to scientific papers from 366 genetics pat-
ents granted from 1980 to 1985 in order to trace linkages from basic research to genetics
technology. The United States had the highest number of papers cited in patents, followed
by the United Kingdom, Japan, Germany, and France. These figures were compared to
the total output of genetics papers for those countries, showing some differences, which
were interpreted as indicating that the United Kingdom produced more papers that were
useful in developing patented technology than Germany, France or Japan. The number of
citations from patents received per paper was highest for the United Kingdom, followed
by the United States and Germany.
Callaert, Van Looy, Verbeek, Debackere, and Thijs (2006) characterize non-patent
references in a sample of patents at the USPTO and the European Patent Office (EPO)
from 1991–2001. Non-patent references are found in 34 percent of USPTO patents and
38 percent of EPO patents, comprising about 17 percent of all references (patent and
non-patent combined). For both the USPTO and EPO, more than half of non-patent
references are journal references. Of the remaining non-patent references, many can be
considered scientific in the broader sense (as they consist of conference proceedings,
books, databases or other non-journal scientific publications), or technology related. The
paper reports that at the USPTO at least, 42 percent of non-journal non-patent refer-
ences can be considered scientific in broader sense, and 40 percent relate to technological
information. For the EPO sample, these figures are 77 percent and 20 percent, respectively.
Tijssen (2002) provides a note of caution on the use of non-patent references. He found
no relationship between the number of non-patent references and the inventor-reported
dependence on science in a small (< 100) sample of Dutch patents from 1998–99. Li,
Chambers, Ding, Zhang, and Meng (2014) qualify this finding. They argue that non-self-
citations to scientific papers are a noisy measure of science linkage, but that applicant
self-citations to scientific papers are indeed informative of science linkage. Roach and
Cohen (2013) matched patent citations to survey reports from R&D lab managers in
the United States, with particular focus on the extent to which patent citations capture
knowledge flows to commercial R&D from publicly funded research. They find that
10
Lemley and Sampat (2012, footnote 12) find that the vast majority of references to non-
patent prior art at the USPTO come from applicants, not examiners, potentially making these a
relevant measure of science dependence.

patent citations reflect codified knowledge. However, citations miss the reliance on private
and contract-based science, as well as basic research. (The discussion in the section on
citations as a measure of knowledge flows considers further whether non-patent refer-
ences are an indicator of science dependence.)
E. Economic Value
As noted above, the (public or private) economic value of an invention is a distinct

concept from its technological impact. Citations are, first and foremost, an indicator of
technological impact. But it turns out that forward citation intensity is, in fact, correlated
with economic value. There are, however, several different concepts of economic value.
First, we can in principle think of the (gross) social value of an invention—that is, the
total producers’ and consumers’ surplus associated with its use. In some cases this gross
social value may be much greater than the net value, for which we would subtract off the
lost rents that may be suffered by previous technologies made wholly or partially obsolete.
The gross social value is greater than the private value—that is, the value to the owner of a
patented invention; the net social value may be either greater or less than the private value,
depending on the magnitude of the “rent stealing” effect. For any of these concepts, we
can distinguish the value of the invention and the value of the patented invention, which
differ by the value of the legal protection afforded by the patent grant. In practice, these
different value concepts may or may not be distinguishable, and proxies for value are often
used whose mapping onto these different value concepts may be ambiguous.
An early strand of research on citations and economic value was the work of Francis
Narin and his associates seeking to develop indicators based on patent data of companies’
competitiveness or technological strength. Carpenter, Narin, and Woolf (1981) showed
that inventions identified in the Industrial Research Institute IR100 Awards are much
more highly cited than a random sample of matched patents. Narin, Noma, and Perry
(1987) found that the average citation frequency of a company’s patent portfolio was
associated with increases in firms’ profits and sales among publicly traded pharmaceutical
companies.
Trajtenberg (1990b) calculated the social welfare gains associated with successive
generations of computed tomography scanners by estimating hedonic demand functions
for the attributes. He then showed that the number of citation-weighted patents associated
with each generation was statistically predictive of the magnitude of welfare gains, while
the raw or unweighted count of patents was not correlated with surplus (sample of about
500 patents). This suggests that the gross social value of these inventions is associated
with the citation intensity of the associated patents. Interestingly, the unweighted patent
counts were correlated with the level of R&D expenditure. He interpreted these findings
as suggesting that the number of patents is associated with the magnitude of research
effort, but not indicative of research success. Counting citation-weighted patents then
combines the scale of effort with a measure of such success and yields a measure of
effective research output.
Moser, Ohmstedt, and Rhode (2014) identified specific improvements in hybrid
corn and gathered data on the magnitude of the yield improvement they allowed. They
interpret this as measuring the “inventive step” associated with the patent; but as the
measurement is in the use domain rather than strictly in the technology domain, it

seems more closely related to social value than to inventive step, per se. They found that
there is indeed a strong correlation between yield improvements and citation intensities.
Interestingly, they find that there are a small number of early patents that are routinely
cited in almost all patents in the field. Excluding these citations enhances the correlation
between yield and citation frequency.
Hall, Jaffe, and Trajtenberg (2005) consider the relationship between citation intensity
and the private value of patents by relating citation-weighted patents to the market
value of the firm. They confirm that citation weighting greatly improves the information
content of patent counts in terms of predicting market value. In addition, they find that
citations from future patents assigned to the same firm as the original patentee have
a larger associated market value than citations from others.11 They also find that a
disproportionate share of the value associated with patents is associated with a very small
number of highly cited patents. Finally, they find that forward citations are associated
with increases in market value at the time a patent is initially granted, suggesting that to
a significant extent market participants can anticipate the eventual value of inventions at
this early stage, and those expectations are (on average) then confirmed by subsequent
citations.
Lanjouw and Schankerman (2001) provide indirect evidence of the relationship
between citations and value, by assuming that patents that are litigated are, on average,
more valuable than those that are not, and comparing the citation patterns of litigated
patents with a control sample of non-litigated patents. They find that the probability of
litigation rises with the number of claims and the number of forward citations per claim,
while declining with the number of backward citations per claim. Allison, Lemley, Moore,
and Trunkey (2003) undertake a similar approach. Consistent with expectations, they find
that litigated patents are more highly cited. Interestingly, they find that litigated patents
also have more backward citations.
Harhoff, Scherer, and Vopel (2003) obtained estimates from patent holders of the
private value of 772 patents with a 1977 German priority date, and that were maintained
to full term. They then examined how that reported value correlated with publicly observ-
able indicia of patent value, including patent citations (and also the number of four-digit
International Patent Classification (IPC) codes and family size). They found that both
the number of forward citations and the number of backward references to the patent
literature are significantly correlated with patent value (see also Harhoff, Narin, Scherer,
and Vopel 1999). Interestingly, they also found that the number of citations made to non-
patent literature was predictive of value, particularly in drug and chemical patents. They
noted that the predictive value of backward citations (both patent and non-patent) is quite
useful, as this information is available at time of patent grant, while forward citations must
be awaited.12 It is unclear theoretically why backward citations are predictive of value. For
non-patent references, it is plausible that in some fields inventions linked to science are less
11
Since Trajtenberg (1990b) showed that total citations are correlated with social returns,
the finding that self-citations have a stronger effect on market value than other citations suggests
that self-citation is associated with the extent of appropriation of the social returns by the original
patenting firm.
12
Similarly, international family size is a measure predictive of value that is knowable soon
after patent application.

incremental and hence more valuable. For backward patent citations, it may reflect some
tendency for bigger, more complex patents to make more backward citations and also be
more valuable on average. In addition, the positive correlation between the number of
backward citations and value may simply arise from the fact that applicants have stronger
incentives to search for prior art for more important patents (Sampat, 2010).
Gambardella, Harhoff, and Verspagen (2008) undertook a similar survey of inventors
listed in patent applications at the EPO. They found that the number of forward citations
is by far the best predictor of reported value, but that the fraction of the variance in
reported value explained by any or all of the metrics was relatively low, consistent with a
view of citation-weighted patents as an indicator of value, but one with substantial noise.
Nicholas (2008) looked at patents granted to US corporations between 1910 and 1939,
and identified the citations to those historical patents from the period 1976–99. He found
that about 15 percent of the patents from the 1910s received at least one citation from
the recent patents, rising to almost 30 percent for those from the 1930s. He then went on
to show that citation-weighted patents constructed in this way are correlated with firm
market value. Thus, patents that are still cited after 40 to 60 years are more valuable than
those that are not. What we cannot know from this exercise (since early citations have
not been captured) is the extent to which valuable patents are simply more highly cited
at all lag durations, or whether there is greater persistence in the sense that the rate of
obsolescence is lower.
Bessen (2008) related the value of patents, as indicated by both renewal information
and firm financial data, to a number of patent characteristics, including forward citations
received. He estimated that each additional citation is associated, on average, with an
increase in value of about 1 percent. Nonetheless, the relationship is very noisy, so that
even among very highly cited patents, a significant fraction appears to be of little value; 37
percent of the patents in the top decile in citation intensity from 1991 were not renewed.
Recent work by Abrams, Akcigit, and Popadak (2015) also suggests an overall positive
correlation between forward citations and patent value, but with an inverted-U-shaped
relationship in which value falls at high citation rates. This finding is provocative, but it is
unclear how robust it is, given the highly selected nature of the sample and the fact that
the value of individual patents was estimated as the value of patent portfolios divided by
the number of patents in the portfolio.
The next section moves away from work focused on citations as indicators of invention
characteristics, and discusses the use of citation data to capture geographic and temporal
dimensions of the innovation process.
III. CITATIONS AS AN INDICATOR OF KNOWLEDGE FLOWS
A. Geographic Dimension of Knowledge Flows
Jaffe, Henderson, and Trajtenberg (1993) took on the challenge identified by Krugman
(1991) on the invisibility of knowledge flows. They suggested that patent citations could
be used as a kind of “paper trail” that could allow knowledge flows to be measured and
tracked. They took a sample of patents from universities, large firms and other firms, and
identified all of their citations. They then found, for every citing patent, a corresponding

“control” patent, issued at the same time and in the same primary US patent class as the
citing patent, and compared the frequency with which citing patents were geographically
proximate to the cited patents with the frequency with which the control patents were
proximate. Looking at metropolitan statistical areas, states and countries, and eliminating
citations that were “self-citations” from the same firm, they showed that citations were
indeed more likely to be proximate. For example, at the level of metropolitan areas, 7–9
percent of citations (depending on the nature of the cited patents) were from the same
area, while only 1–4 percent of the control patents were, and the differences were highly
significant statistically.
Thompson and Fox-Kean (2005) criticize the Jaffe, Henderson and Trajtenberg meth-
odology. They argue that selecting control patents based on the primary patent class of
the citing patents is too rudimentary to capture the heterogeneity of technology. Patents
in the same main patent class may be in different subclasses with inherently different tech-
nologies, and patents are assigned to multiple classes, again introducing heterogeneity not
captured by the main patent classification. In response, Henderson, Jaffe, and Trajtenberg
(2005) agree that it is possible that finer technological controls might be appropriate,
but they point out that slicing things too finely minimizes the possibility for identifying
knowledge flows across subclasses. Ultimately, the question comes down to the robustness
of the localization effect under different identifying assumptions.
A number of other authors have similarly used citation data to measure knowledge
flows. Almeida and Kogut (1997) compare the patent citations of small and large semi-
conductor firms, and find that the citations made by small firms are more geographically
localized. Hicks, Breitzman, Olivastro, and Hamilton (2001) show that US companies’
citations to university patents exhibit geographic localization, particularly to patents of
nearby public universities. Almeida and Kogut (1999) examine citation patterns among
semiconductor firms in the United States, including data on both the firms and the inven-
tors. They show that a significant fraction of the geographic localization of the citations
can be traced to specific engineers who move among firms, but are more likely to move
to another nearby firm than to one that is farther away. Sonn and Storper (2008) show
that, despite improvements in communications technologies, geographical localization
has been increasing over time.
Thompson (2006) compares the extent of localization in citations listed by the inventor
to those added by the examiner. He finds localization at both the metropolitan area and
state levels in both the examiner and inventor citations. Inventor citations are found to
be about 20 percent more likely to match the country of origin of the citing patent than
are examiner citations. In a similar vein, Alcácer and Gittelman (2006) estimate the prob-
ability that a citation is generated by an examiner or an inventor, conditional on a set of
variables that are frequently employed in the knowledge spillover literature. They find
that examiner citations introduce bias for some variables only (e.g., self-citations). They
find no evidence that the degree of geographic proximity between citing and cited patents
differs for inventor and examiner citations.
A subtler pitfall in the use of citations to track knowledge flows relates to the interven-
tion of law firms in the drafting of the patent document. Wagner, Hoisl, and Thoma
(2014) show that patents by firms which rely on external agents are more likely to cite
documents that are part of the law firm’s knowledge repository. They take this result as
evidence that law firms help overcome localization. However, a blunter interpretation is

that external agents include citations that the firms were not aware of, further increasing
the noise in patent citation data.
Maurseth and Verspagen (2002) used data on citations among European patents to
construct a region-by-region citations frequency matrix. They then looked at numerous
variables to explain these frequencies. Geographical distance has a negative and substan-
tial impact on knowledge flows. Controlling for distance, knowledge flows are greater
between regions located within one country than between regions located in separate
countries. The country effect remains even if regions share the same language, though
sharing a language increases the amount of knowledge flows between two regions by up
to 28 percent. The study also suggests that knowledge flows are industry specific, and
regions’ technological specialization is an important determinant of their technological
interaction.
B. Temporal Dimension of Knowledge Flows
Caballero and Jaffe (1993) and Jaffe and Trajtenberg (1999) developed a structured
model of knowledge diffusion across space and time. They postulate that two competing
forces dominate the citation process. Over time, knowledge gradually diffuses, so that the
number of people potentially citing a given patent increases exponentially with time. But
the relevance or usefulness of a bit of knowledge becomes obsolete, leading to a counter-
vailing exponential depreciation in the likelihood of citations. The parameters of these
two exponential functions can be estimated econometrically. If allowed to vary across
different technologies, different kinds of research organizations, and different geographic
locations, they then capture the rates of diffusion in different areas across organizations
and across space. Jaffe and Trajtenberg show, for example, that the geographic localization
of citations diminishes as time passes, and also that obsolescence (as captured by declining
citation rates) is more rapid in electronic technologies than in chemical and mechanical
technologies.
Bacchiocchi and Montobbio (2009) used this double-exponential function to look
at knowledge flows from universities and public research organizations compared to
flows from corporate patents in six countries: France, Germany, Italy, Japan, the United
Kingdom, and the United States. They found that technology embodied in patents from
universities and public research organizations diffuses more rapidly than that of firms.
The diffusion rates are relatively homogenous across technological fields, but vary across
countries: rapid in the United States and Germany, less so in France and Japan.
Mehta, Rysman, and Simcoe (2010) have criticized this diffusion model on the ground
that the age of a citation is computed as the citation year minus the application year,
leading to an identification problem. Because citations received by a patent are rare
before it is issued, the authors propose to use the lag between application year and
grant year as a source of exogenous variation. They find that the citation peak occurs
earlier than suggested by the double-exponential function. However, their method does
not alter differences in the mean citation ages across industries. They conclude that the
double-exponential function provides a good approximation to the non-parametric age
distribution.

C. Validation Studies
Jaffe, Trajtenberg, and Fogarty (2000) report a survey of inventors to test the extent to
which citations in those inventors’ patents correspond to the inventors’ perceptions of
how their inventions depended on earlier knowledge, and how the rate of citation relates
to inventors’ own perceptions of impact or importance. They find that citations are a valid
but noisy indicator of knowledge flows: the likelihood of reported knowledge impact is
significantly higher (both quantitatively and statistically) when a citation link exists, but
a significant fraction of citations (perhaps as high as one-half) do not correspond to any
reported knowledge link.
Duguet and MacGarvie (2005) tested the validity of patent citations as a measure of
knowledge flows using data from French firms on their patents and citations, combined
with survey responses regarding sources of knowledge. The total number of backward
citations was correlated with survey answers about R&D and innovation, but this cor-
relation was weakened by controlling for the number of patents held by the citing firm.
Backward citation rates of French firms can reflect their R&D activities (if the technology
is obtained from firms located in the EU), or purchases of equipment goods (if the source
is located outside the EU). In general, it can be understood that backward citations are
correlated with learning through R&D collaboration, licensing of foreign technology,
mergers and acquisitions and equipment purchases.
In their analysis, Roach and Cohen (2013, discussed above) found evidence of both
“errors of omission” (reported knowledge flows with no corresponding citations) and
“errors of commission” (observed citations with no corresponding reported knowledge
flows). They conclude that despite these sources of measurement error, patent cita-
tions are likely to reflect meaningful aspects of knowledge flows from public research.
Interestingly, they found that references in patents to non-patent publications (primarily
scientific literature) are a better indicator of knowledge flow than are citations in com-
mercial patents to the patents of universities and other public labs (cf. Tijssen, 2002).
The next section discusses a third category of citation data research, in which the focus
shifts to using citation links to understand and characterize networks.
IV. CITATIONS AS LINKS IN KNOWLEDGE OR INNOVATION

NETWORKS
A natural way of representing citation data is in the form of a network. Researchers have
used concepts from network theory to grasp the way the innovation system is structured
and the way knowledge is formed. A first group of studies seeks to map key components
of the innovation system (patents, individuals, institutions and regions). A second group
of studies uses the network of citations to map technological trajectories. We review these
two applications in turn.
A. Mapping Patents, Individuals, Institutions, and Regions
Huang, Chiang, and Chen (2003) rely on patent citation data to map Taiwan’s electronic
industry. The researchers identify USPTO patents belonging to 58 relevant Taiwanese

companies as well as the citations made by these patents. They identify the strength of
the relationship between companies by looking at the strength of bibliographic coupling.
Bibliographic coupling is a method proposed by Kessler (1963) that involves identifying
related documents through common cited references. The researchers then applied cluster
analysis to the data produced to identify major sectors of the Taiwanese electronic indus-
try. Although bibliographic coupling provides rich insights on the relatedness of patent
documents, more recent studies make better use of network analysis theory and tools.
Chen and Hicks (2004) study the citation “degree” distribution of 16 million citations
made to the 3 million USPTO patents granted in the period from 1963 to 1999. The degree
of a “node” (patent) is simply the number of “connections” (citations) received by the
node. They estimate that the distribution follows a power law with an exponent of 2.89,
which is very similar to the parameter obtained for scientific papers by Dorogovtsev and
Mendes (2002).13 The fact that the degree distribution of the patent citation network
follows a power-law is indicative of so-called scale-free networks, which can be seen as
networks characterized by large hubs through which knowledge flows.
Li, Chen, Huang, and Roco (2007) use a patent citation network to study the knowledge
transfer process between entities. In particular, they study the efficiency with which
knowledge transfers within the network compared to a random network. Their measure
of efficiency is the average path length between any pair of patents in the network. They
focus on USPTO nanotechnology patents in the period from 1976 to 2004. They find
that knowledge transfer across assignees in the citation network is more efficient than
knowledge transfer that would occur in a random network. Knowledge flow across
(assignee) countries is as efficient as a random network. However, knowledge flow across
technology fields is less efficient than knowledge flow that would occur in a random
network. In other words, technological distance is a greater barrier to knowledge flows
than geographic distance.
Hung and Wang (2010) examine the characteristics of the citation network formed
by radio-frequency identification (RFID) patents. They find that the network can be
characterized as a “small-world” network—that is, a network in which most nodes can be
reached from every other by a small number of steps. They also find that the network has a
power-law connectivity distribution and exhibits preferential connectivity behavior. That
is, a few key patents have a very large number of connections and the majority of patents
have few connections. The authors conclude that only a limited number of patents play
a key role in diffusing RFID technology. This approach provides a more system-based
way of thinking about knowledge flows than simply counting citations: key patents
not only are highly cited patents, but also connect and integrate different technological
trajectories.14 More detailed analyses of technological trajectories using citation network
are described in the next section.
13
Cf. Huang, Huang, Chang, Chen, and Lin (2014), who provide evidence that the distribu-
tion of patent citations is more concentrated than the distribution of citations in scientific articles.
14
Hu, Rousseau, and Chen (2012) provide another study on the importance of patents using
their positions in the citation network. Other applications include Liu and Shih (2011), who use the
network formed by patents to improve patent classification.

B. Mapping of Technological Trajectories
Scholars have recently used citation networks to identify technological trajectories that
led to the advent of major technological breakthroughs. The main trajectory, or search
path, is the sequence of links and nodes that is central to the development of a technology.
It represents the main flow of ideas in the development of a technology. The method was
pioneered by Hummon and Doreian (1989) on a citation network of scientific papers
describing the development of DNA theory. This approach shifts the focus from the
nodes of the network (looking at individual patents) to the connections that these nodes
form. It allows the identification of key patents through their structural connectivity in
the network. Technologically important patents should belong to the main paths of the
citation network and/or locate at particularly critical junctions within those paths.
Mina, Ramlogan, Tampubolon, and Metcalfe (2007), Verspagen (2007), and Fontana,
Nuvolari, and Verspagen (2009) applied the method to patent citation networks. Mina
et al. (2007) use it to understand how medical knowledge emerges, grows and evolves.
They argue that the approach provides a dynamic view of innovation that recognizes the
long-term, path-dependent and complex nature of technology. Their case study is based
on treatment for coronary artery disease and covers 5136 USPTO patent documents
granted between 1976 and 2003. The authors seek to identify the main path and “islands”
of the network. Islands are small clusters of inventions whose internal connectedness is
relatively superior to the strength of their outward connections within the global network.
The authors argue that islands allow accounting for the variety of complementary and
competing areas of technical expertise that contributed to the advancement of the
technology. They report that the results form a consistent map of the major scientific and
technological trajectories in the domain.
Fontana, Nuvolari, and Verspagen (2009) study the structural connectivity of the
citation network formed by patents related to local area networks (LAN) technology.
Innovation in such a systemic technology has three main features. First, innovation is
distributed: it takes place at the level of individual components, but these components
all have to work together. Second, innovations in systems tend to be incremental and to
occur around well-established technical designs. Third, innovations also tend to occur
continuously. The authors argue that the classical approach of assessing the importance
of patents by counting the number of citations they have received may have drawbacks in
such systemic technologies. It may fail to identify concepts and principles that could act as
“focusing devices” for a sequence of inventive activities. By contrast, a structural analysis
of the citation network would allow the identification of inventions that have played a
major role in the evolution of LAN technology. They find that the main path they have
identified displays a coherent economic and engineering logic, consistent with qualitative
accounts of the evolution of the Ethernet standard.
One of the most interesting insights of the paper comes from the analysis of companies
owning patents that lie on the main path. No company is “dominant” in the sense of
claiming ownership of the majority of patents on the main path, which the authors take
as evidence that no company is strategically placed along the main path of knowledge
flow. Verspagen (2007) performs a similar analysis for citations among fuel cell patents. He
finds that there are dominant companies: a small number of organizations hold patents
belonging to the main path. Study of the ownership structure of technologies on the

main path provides a novel way of characterizing technology dominance. It is a promising

avenue for research in industrial economics and strategic management.
V. PITFALLS AND BEST PRACTICES IN USE OF CITATION-

BASED INDICATORS
We take the opportunity of this review to discuss potential pitfalls associated with patent
data. We focus on four key challenges.
A. Office Effects
Institutional differences across jurisdictions induce differences in citation practices across

offices. We briefly summarize two main differences in citation practices between the EPO/
Japan Patent Office (JPO) and the USPTO for illustrative purposes. More generally,
researchers should get a clear understanding of citation practices in the office of interest
before using citation-based indicators.15
A first difference is the “duty of candor” in US patent law. Failure to report known
relevant prior art may lead to subsequent revocation of the patent (inequitable conduct
doctrine). There is no duty of candor in European patent law, and applicants do not have
to submit a list of prior art. It follows that search reports at the EPO usually contain many
fewer references than USPTO search reports. In fact, according to EPO philosophy, “a
good search report contains all the technically relevant information within a minimum
number of citations” (Michel and Bettels, 2001, p. 189). In addition, since applicants at the
EPO do not bear the same responsibility to disclose prior art as applicants at the USPTO,
the citations come mostly from the examiner. This does not undermine their interpreta-
tion as indicators of impact or value; for example, Harhoff, Scherer, and Vopel (2003,
discussed above) find EPO citations to be predictive of value. It does suggest that EPO
citations might be less indicative of knowledge flows, although we are not aware of any
empirical analysis of this question comparable to the survey work of Jaffe, Trajtenberg,
and Fogarty (2000). In Japan, the patent law was revised in 2002 and imposed on appli-
cants the obligation to disclose prior art. Although non-compliance with the disclosure
requirement bears less severe consequences than in the United States, the reform led to a
substantial increase in prior art disclosure by applicants. Takahiro, Nagaoka, and Naito
(2015) find that about 8 percent of citations came from applicants in the years following
the reform, compared to around 4 to 5 percent before the reform.
A second important difference with the USPTO is that EPO patent examiners classify
documents cited in particular citation categories (Schmoch, 1993). A document that
shows essential features of the invention or questions the inventive step of these features
if taken alone is marked with the letter “X”. A document that questions the inventive step
if combined with another document is marked with the letter “Y” (hence “Y” citations
never occur singly). The letter “A” marks a document that shows the general state of the
15
For example, researchers interested in EPO citations should read the “Guidelines for
Examination in the European Patent Office” available on the EPO website.

art. According to Schmoch (1993, p. 195), a patent document can be highly cited because
it comprises “a good description of the prior art from a didactic point of view.” The
classification provides opportunities for finer analyses. One may want to exclude class
“A” citations for assessing the inventive step of patents, but class “A” citations are relevant
for measuring technological proximity of patents. Additional classification codes exist;
see Webb, Dernis, Harhoff, and Hoisl (2005) for a discussion. Examiners at the JPO also
classify citations into categories. In particular, they flag whether citations are used as
grounds for rejection, similar to “X” and “Y” citations at the EPO, or whether they are
used for assessing the application but do not serve as a basis for rejection, similar to “A”
citations at the EPO (Goto and Motohashi, 2007).
The classification into categories opens the door to original uses of citation data. For
example, von Graevenitz, Wagner, and Harhoff (2011) identify patent thickets at the EPO
using X and Y citations. Their measure identifies constellations in which three firms each
own patents that block patent applications of the other two firms (so-called triples). The
authors show that density of triples in complex technology areas has risen steadily since
the early 1980s, whereas the density of triples has been constant in discrete technology
areas. Guellec, Martinez, and Zuniga (2012) use “X” and “Y” citations together with
administrative information on the patent examination process (withdrawal and grant
events) to identify defensive patents—that is, patent applications used to pre-empt others
from getting their patents granted. Palangkaraya, Webster, and Jensen (2011) posit that
patents with a higher inventive step will generate more “X” and “Y” citations, and use this
information to proxy for the probability of grant ex-ante.
Beyond institutional differences in the use of citations, researchers have also illustrated
the presence of home bias in citation practices. Bacchiocchi and Montobbio (2010)
analyze the geographic distribution of cited documents for a set of 657 151 equivalent
patents filed at the EPO and the USPTO. In theory, distributions should be similar, since
they refer to the same invention. They find that the frequency of US-cited patents at the
USPTO exceeds 65 percent, while the frequency at the EPO is less than 40 percent. That
examiners have a tendency to cite local documents does not come as a surprise.16 However,
it illustrates an important limitation of the use of citations for assessing cross-border
knowledge flows.
B. Time and Technology Field Effects
The number of citations received by a patent increases as time passes such that there are
strong cohort effects. This issue can be dealt with in a straightforward manner by counting
citations received in a fixed time interval (e.g., citations received up to five years after
grant). A more serious concern is the increase over time of citations made per patent. Hall,
Jaffe, and Trajtenberg (2001) report that the average patent issued in 1999 made over twice
as many citations as the average patent issued in 1975 (10.7 versus 4.7 citations). Although
this issue does not affect the comparison of patents within a cohort, citation inflation
16
For example, there is a substantial cost to including non-English references at the USPTO.
When using a foreign language reference in a rejection, examiners should provide a translation of
the entire document.

makes it challenging to compare patents across cohorts. Analogously, citation practices

and the intensity of activity vary by technology fields, so that what constitutes a high
citation rate in one field may be modest or small for another field.17 The authors discuss
two econometric techniques to deal with citation inflation and varying intensities by
field: scaling citation counts by “dividing them by the average citation count for a group
of patents to which the patent of interest belongs”; and identifying the multiple biases
on citation rates via econometric estimation. Marco (2007) provides a recent illustration
of the latter technique. He argues that by estimating a hazard rate based only on factors
that are correlated to citation inflation rather than value, residuals can be used to measure
latent patent value. For example, the ratio of observed citations to predicted citations
may represent a proxy of patent value. Such an approach is an important step forward,
although it is difficult to identify factors that are truly exogenous to value.
A broader question, which has received little coverage in the literature, relates to dif-
ferences in patenting and citation practices across technology fields. We know that the
propensity to patent differs across fields (Cohen, Nelson, and Walsh, 2000), and that
the relevance of patent data as innovation indicator therefore also varies across fields
(e.g., Danguy, de Rassenfosse, and van Pottelsberghe, 2014). However, to the best of our
knowledge, no study has investigated in a systematic manner how differences across fields
affect the relevance of patent citation data.
C. Examiner Effects
Cockburn, Koru, and Stern (2002) show that there is substantial examiner heterogene-
ity—for example, in terms of variations in tenure at the USPTO and in the average
approval time per issued patent. Such heterogeneity translates into variations in outcomes
of the examination process—such as in the volume and pattern of citations made.18
Lemley and Sampat (2012) demonstrate the presence of an examiner effect, in the sense
that more experienced examiners cite less prior art. Alcácer and Gittelman (2009) paint a
picture of examiner-added citations across key strata of patent data. They report that the
proportion of citations added by examiners is higher for patents: by foreign applicants to
the USPTO; by applicants with a large patent portfolio; and by applicants in electronics,
communications, and computer-related fields. Criscuolo and Verspagen (2008) perform
a similar analysis for EPO patent data. They show that the share of inventor citations
declined from about 14 percent in 1985 to 9 percent in 2000. In addition, there is sub-
stantial variation across fields. More than 20 percent of citations in organic chemistry
patents were added by the inventor, while for information technology patents this share
is 4 percent.19
17
Technology fields are tracked using the patent office classification systems. Historically,
the United States has maintained its own classification (USPC), while other offices use the IPC.
The USPTO has recently introduced a Cooperative Patent Classification based on the IPC, and is
phasing out the USPC.
18
Alcácer and Gittelman (2006) estimate that examiners insert two-thirds of citations on the
average patent, and 40 percent of all patents have all citations added by examiners.
19
There are few country-specific studies. See Azagra-Caro, Mattsson, and Perruchas (2011) for
Spanish evidence.

Examiner intervention may bias the information content of citations. It may undermine
the use of citations as a measure of knowledge flow, since the inventors may not have even
been aware of the patents cited by examiners at the time of invention. However, examiner
citations may be taken as a valid reflection of technological and economic value. In this
spirit, Hegde and Sampat (2009) show that examiner citations have a much stronger
relationship with renewal probability (a measure of private value) than the number of
applicant citations.
D. Strategic Effects
Variations in the number of examiner-added citations may also come from differences
in applicants’ incentives to search for or disclose prior art. Recent research suggests that
citing prior art (or not) is a strategic decision. Atal and Bar (2010) study firms’ incen-
tives to search for unknown prior art. Although applicants at the USPTO have a duty
to disclose what they know, they have no duty to search for prior art and may be better
off by remaining ignorant. The authors show theoretically that firms search more when
R&D investment (a proxy for innovation quality) and patenting costs are higher. Sampat
(2010) provides empirical data on when applicants search for prior art. He shows that
applicants contribute more prior art for their more important inventions. He also shows
that applicants are more likely to search for prior art in fields where individual patents
are important for appropriating returns from R&D (chemicals and drugs) and less likely
to do so in industries where firms tend to accumulate patent portfolios for other strategic
reasons (computers and communications, electronics and electrical, and mechanical).
Lampe (2012) focuses on applicants’ decision to disclose known prior art. He identifies
“voluntary withholding” of citations to prior art material by looking only at citations
that were present on prior patents issued to the same firm. He estimates that applicants
withhold between 21–33 percent of relevant citations. The rate is higher for firms applying
for computer and electronic patents (25–42 percent) and lower for firms applying for drug
and chemical patents (8–22 percent). More generally, Lampe finds that the likelihood
of citation is positively correlated with proxies of patent value (number of claims and
forward citations) and negatively correlated with the size of applicant patent portfolios.
VI. CONCLUSION
The use of patent citation data in social science research has exploded in the last two
decades. As just one indication, the frequency of appearance of the term “patent citation”
in scientific documents listed in Google Scholar increased tenfold between 2000 and
2014 (Appendix A). As is often the case, this increase reflects increases in both supply
and demand. On the supply side, the digitization of the patent office records, combined
with the increased power of computers to analyze them, makes analyses possible today
that simply could not have been undertaken 25 years ago. The number of scientific docu-
ments referencing the National Bureau of Economic Research patent citation data file
is likewise continuously increasing (Appendix A). On the demand side, intangible assets
are increasingly seen as a source—some would argue the dominant source—of economic
returns. By definition, intangible assets are hard to track and measure, and so researchers

interested in diverse questions about knowledge accumulation and diffusion, innovation,

firm strategy and regional economic growth seek measures that convey information about
the sources and consequences of these assets.
Neither of these trends is likely to reverse, so interest in measures of this kind is likely
to continue to grow. Recent developments in computational linguistics may allow for
construction of measures that are conceptually related to citations, but use all of the
information contained in the patent text rather than relying solely on the links between
patents that are explicitly identified via citation. It is now possible, for example, to identify
connections between a patent and its antecedents by measuring the frequency with which
important words are used in both patents, to measure novelty by identifying patents that
use a certain technical term or combination of words in a particular phrase for the first
time, and to measure impact by counting the number of subsequent patents that use such
a phrase (e.g., Packalen and Bhattacharya, 2015). Younge and Khun (2015) use more
advanced techniques to develop a text-based pairwise similarity comparison of any and
every two patents at the USPTO. These new approaches have not yet been subjected to
the kind of validation that has demonstrated the economic significance of citations, but
because they utilize more information, they offer the promise of a valuable broadening
and deepening of the research possibilities.
A more mundane, but equally important task is to further validate citation indicators.
This applies to both established and novel indicators, at both the USPTO and other
offices. For example, it is unclear whether the count of backward citations proxies for
patent importance. Even the link between forward citations and economic value, one
of the most established and used indicators, is not well understood. In a similar vein,
little research exists on technology field differences on the relevance of patent citation
data. The need for validation studies will grow more pressing as new indicators are being
developed and more patent offices make their data available. Similarly, legislative changes
affect citation practices in non-trivial ways, and conclusions drawn using data from one
time period are not necessarily valid in another time period. This calls for a continuous
assessment of the validity of citation indicators.
Another exciting area of research is the further application of network theory and
analysis tools to the patent citation network. For example, the identification of key
technologies and actors on the main knowledge path promises to greatly improve our
understanding of industry dynamics and the knowledge creation process. A limitation of
current research in the area is the insularity of two communities of scholars. Studies by
scholars using advanced network analysis tools offer little practical implications, whereas
studies by scholars looking at real-world implications use quite basic network analysis
tools. A promising way forward is to better integrate the technical and the practical aspects
of network analysis.
Finally, researchers realize that the patent citation generation process is complex, but
more work needs to be done to understand it. The complexity of the patent citation
generation process is a blessing and a curse. Whereas it may distort the reality in an
undesirable fashion, it may also provide a window into the incentives faced by inventors,
patent attorneys and examiners and serve as a source of econometric identification.
The example of examiner-added citations is a case in point. Whereas citations made by
examiners arguably weaken the measurement of knowledge flows, they also strengthen
the measurement of patent value.

REFERENCES
Abrams, David, Ufuk Akcigit, and Jillian Popadak. 2015. “Patent value and citations: creative destruction or
strategic disruption?” University of Pennsylvania Law School PIER Working Paper 13-065.
Ahuja, Gautam, and Curba Lampert. 2001. “Entrepreneurship in the large corporation: a longitudinal study
of how established firms create breakthrough inventions”, 22:6–7 Strategic Management Journal 521–543.
Albert, M., D. Avery, F. Narin, and P. McAllister. 1991. “Direct validation of citation counts as indicators of
industrially important patents”, 20:3 Research Policy 251–259.
Alcácer, Juan, and Michelle Gittelman. 2006. “Patent citations as a measure of knowledge flows: the influence
of examiner citations”, 88:4 The Review of Economics and Statistics 774–779.
Alcacer, Juan, Michelle Gittelman, and Bhaven Sampat. 2009. “Applicant and examiner citations in US patents:
an overview and analysis”, 38:2 Research Policy 415–427.
Allison, John, Mark Lemley, Kimberley Moore, and R. Derek Trunkey. 2003. “Valuable patents”, 92 Georgetown
Law Journal 435.
Almeida, Paul, and Bruce Kogut. 1997. “The exploration of technological diversity and the geographic localiza-
tion of innovation”, 9 Small Business Economics 21–31.
Almeida, Paul, and Bruce Kogut. 1999. “Localization of knowledge and the mobility of engineers in regional
networks”, 45:7 Management Science 905–917.
Atal, Vidya, and Talia Bar. 2010. “Prior art: to search or not to search”, 28:5 International Journal of Industrial
Azagra-Caro, Joaquín, Pauline Mattsson, and François Perruchas. 2011. “Smoothing the lies: the distinctive
effects of patent characteristics on examiner and applicant citations”, 62:9 Journal of the American Society
for Information Science and Technology 1727–1740.
Bacchiocchi, Emanuele, and Fabio Montobbio. 2009. “Knowledge diffusion from university and public research:
a comparison between US, Japan and Europe using patent citations”, 34:2 The Journal of Technology Transfer
169–181.
Bacchiocchi, Emanuele, and Fabio Montobbio. 2010. “International knowledge diffusion and home-bias effect:
do USPTO and EPO patent citations tell the same story?”, 112:3 The Scandinavian Journal of Economics
441–470.
Benson, Christopher, and Christopher Magee. 2015. “Quantitative determination of technological improvement
from patent data”, 10:4 PLoS ONE e0121635.
Bessen, James. 2008. “The value of US patents by owner and patent characteristics”, 37:5 Research Policy
932–945.
Brin, Sergey, and Lawrence Page. 1998. “The anatomy of a large-scale hypertextual web search engine”, 30
Computer Networks and ISDN Systems 107–117.
Caballero, Ricardo, and Adam Jaffe. 1993. “How high are the giants’ shoulders: an empirical assessment of
knowledge spillovers and creative destruction in a model of economic growth”, in NBER Macroeconomics
Annual 1993, Volume 8 (pp. 15–86). Cambridge, MA: MIT Press.
Callaert, Julie, Bar Van Looy, Arnold Verbeek, Koenraad Debackere, and Bart Thijs. 2006. “Traces of prior art:
an analysis of non-patent references found in patent documents”, 69:1 Scientometrics 3–20.
Carpenter, Mark, and Francis Narin. 1983. “Validation study: patent citations as indicators of science and
foreign dependence”, 5:3 World Patent Information 180–185.
Carpenter, Mark, Francis Narin, and Patricia Woolf. 1981. “Citation rates to technologically important
patents”, 3:4 World Patent Information 160–163.
Chen, Chaomei, and Diana Hicks. 2004. “Tracing knowledge diffusion”, 59:2 Scientometrics 199–211.
Clark, C. V. 1976. “Obsolescence of the patent literature”, 32:1 Journal of Documentation 32–52.
Cockburn, Iain, Samuel Kortum, and Scott Stern. 2002. “Are all patent examiners equal?: the impact of
characteristics on patent statistics and litigation outcomes”, National Bureau of Economic Research Working
Paper 8980.
Cohen, Wesley, Richard Nelson, and John Walsh. 2000. “Protecting their intellectual assets: appropriability
conditions and why U.S. manufacturing firms patent or not”, National Bureau of Economic Research Working
Paper 7552.
Collins, Peter, and Suzanne Wyatt. 1988. “Citations in patents to the basic research literature”, 17:2 Research
Policy 65–74.
Cotropia, Christopher, Mark Lemley, and Bhaven Sampat. 2013. “Do applicant patent citations matter?”, 42:4
Criscuolo, Paula, and Bart Verspagen 2008. “Does it matter where patent citations come from? Inventor vs.
examiner citations in European patents”, 37:10 Research Policy 1892–1908.
Czarnitzki, Dirk, Katrin Hussinger, and Cédric Schneider. 2011. “‘Wacky’ patents meet economic indicators”,
113:2 Economics Letters 131–134.

Dahlin, Kristina, and Dean Behrens. 2005. “When is an invention really radical?: Defining and measuring
technological radicalness”, 34:5 Research Policy 717–737.
Danguy, Jérôme, Gaétan de Rassenfosse, and Bruno van Pottelsberghe de la Potterie. 2013. “On the origins of
the worldwide surge in patenting: an industry perspective on the R&D patent relationship”, 23:2 Industrial
and Corporate Change 535–572.
Dorogovtsev, S., and J. Mendes. 2002. “Evolution of networks”, 51:4 Advances in Physics 1079–1187.
Duguet, Emmanuel, and Megan MacGarvie. 2005. “How well do patent citations measure flows of technology?
Evidence from French innovation surveys”, 14:5 Economics of Innovation and New Technology 375–393.
Fontana, Roberto, Alessandro Nuvolari, and Bart Verspagen. 2009. “Mapping technological trajectories as
patent citation networks. An application to data communication standards”, 18:4 Economics of Innovation
and New Technology 311–336.
Gambardella, Alfonso, Dietmar Harhoff, and Bart Verspagen. 2008. “The value of European patents”, 5:2
European Management Review 69–84.
Garfield, Eugene. 1955. “Citation indexes for science: a new dimension in documentation through association
of ideas”, 122:3159 Science 108–111.
Garfield, Eugene. 1966. “Patent citation indexing and the notions of novelty, similarity, and relevance”, 6:2
Journal of Chemical Documentation 63–65.
Gay, C., and C. Le Bas. 2005. “Uses without too many abuses of patent citations or the simple economics
of patent citations as a measure of value and flows of knowledge”, 14:5 Economics of Innovation and New
Goto, Akira, and Kazuyuki Motohashi. 2007. “Construction of a Japanese patent database and a first look at
Japanese patenting activities”, 36:9 Research Policy 1431–1442.
Griliches, Zvi. 1979. “Issues in assessing the contribution of research and development to productivity growth”,
10:1 Bell Journal of Economics 92–116.
Griliches, Zvi. 1990. “Patent statistics as economic indicators: a survey”, 28:4 Journal of Economic Literature
1661–1707.
Guellec, Dominique, Catalina Martinez, and Pluvia Zuniga. 2012. “Pre-emptive patenting: securing market
exclusion and freedom of operation”, 21:1 Economics of Innovation and New Technology 1–29.
Hall, Bronwyn, Adam Jaffe, and Manuel Trajtenberg. 2001. “The NBER patent citation data file: lessons,
insights and methodological tools”, National Bureau of Economic Research Working Paper 8498.
Hall, Bronwyn, Adam Jaffe, and Manuel Trajtenberg. 2005. “Market value and patent citations”, 36:1 RAND
Hall, Bronwyn, and Manuel Trajtenberg. 2006. “Uncovering GPTs using patent data”, in C. Antonelli, D. Foray,
B.H. Hall, W.E. Steinmuller, eds., New Frontiers in the Economics of Innovation and New Technology—Essays
in Honor of Paul A. David. Cheltenham, UK; Northhampton, MA: Edward Elgar Publishing.
Harhoff, Dietmar, Francis Narin, Frederic Scherer, and Katrin Vopel. 1999. “Citation frequency and the value
of patented inventions”, 81:3 Review of Economics and Statistics 511–515.
Harhoff, Dietmar, Frederic Scherer, and Katrin Vopel. 2003. “Citations, family size, opposition and the value
of patent rights”, 32:8 Research Policy 1343–1363.
Hegde, Deepak, and Bhaven Sampat. 2009. “Examiner citations, applicant citations, and the private value of
patents”, 105:3 Economics Letters 287–289.
Henderson, Rebecca, Adam Jaffe, and Manuel Trajtenberg. 2005. “Patent citations and the geography of
knowledge spillovers: a reassessment: comment”, 95:1 American Economic Review 461–464.
Hicks, Diana, Tony Breitzman, Dominic Olivastro, and Kimberly Hamilton. 2001. “The changing composition
of innovative activity in the US—a portrait based on patent analysis”, 30:4 Research Policy 681–703.
Hu, Xiaojun, Ronald Rousseau, and Jin Chen. 2012. “A new approach for measuring the value of patents
based on structural indicators for ego patent citation networks”, 63:9 Journal of the American Society for
Information Science and Technology 1834–1842.
Huang, Mu-Hsuan, Li-Yin Chiang, and Dar-Zen Chen. 2003. “Constructing a patent citation map using
bibliographic coupling: a study of Taiwan’s high-tech companies”, 58:3 Scientometrics 489–506.
Huang, Mu-Hsuan, Wei-Tzu Huang, Cheng-Ching Chang, Dar-Zen Chen, and Chang-Pin Lin. 2014. “The
greater scattering phenomenon beyond Bradford’s law in patent citation”, 65:9 Journal of the Association for
Information Science and Technology 1917–1928.
Hummon, Norman, and Patrick Dereian. 1989. “Connectivity in a citation network: the development of DNA
theory”, 11:1 Social Networks 39–63.
Hung, Shiu-Wan, and An-Pang Wang. 2010. “Examining the small world phenomenon in the patent citation
network: a case study of the radio frequency identification (RFID) network”, 82:1 Scientometrics 121–134.
Jaffe, Adam, and Manuel Trajtenberg. 1999. “International knowledge flows: evidence from patent citations”,
8:1–2 Economics of Innovation and New Technology 105–136.
Jaffe, Adam, and Manuel Trajtenberg. 2002. Patents, Citations, and Innovations: A Window on the Knowledge
Economy. Cambridge, MA: MIT Press.

Jaffe, Adam, Manuel Trajtenberg, and Michael Fogarty. 2000. “Knowledge spillovers and patent citations:
evidence from a survey of inventors”, 90:2 American Economic Review 215–218; also published with addi-
tional detail as “The Meaning of Patent Citations: Report on the NBER/Case-Western Reserve Survey of
Patentees”, in Jaffe and Trajtenberg (2002), op cit.
Jaffe, Adam, Manuel Trajtenberg, and Rebecca Henderson. 1993. “Geographic localization of knowledge
spillovers as evidenced by patent citations”, 108:3 Quarterly Journal of Economics 577–598.
Kessler, M. 1963. “Bibliographic coupling between scientific papers”, 14:1 American Documentation 10–25.
Krugman, Paul. 1991. Geography and Trade. Cambridge, MA: MIT Press.
Kuznets, Simon. 1962. “Inventive activity: Problems of definition and measurement”, in The Rate and Direction
of Inventive Activity: Economic and Social Factors (pp. 19–52). Princeton, NJ: Princeton University Press.
Lampe, Ryan. 2012. “Strategic citation”, 94:1 Review of Economics and Statistics 320–333.
Lanjouw, Jean, and Mark Schankerman. 2001. “Characteristics of patent litigation: a window on competition”,
32:1 RAND Journal of Economics 129–151.
Lemley, Mark, and Bhaven Sampat. 2012. “Examiner characteristics and patent office outcomes”, 94:3 Review
Lerner, Joshua, and Amit Seru. 2015. “The use and misuse of patent data: issues for corporate finance and
beyond”, Harvard Business School Mimeo.
Li, Rui, Tamy Chambers, Ying Ding, Guo Zhang, and Liansheng Meng. 2014. “Patent citation analysis:
calculating science linkage based on citing motivation”, 65:5 Journal of the Association for Information Science
Li, Xin, Hsinchun Chen, Zan Huang, and Mihail Roco. 2007. “Patent citation network in nanotechnology
(1976–2004)”, 9:3 Journal of Nanoparticle Research 337–352.
Liu, Duen-Ren, and Meng-Jung Shih. 2011. “Hybrid-patent classification based on patent-network analysis”,
62:2 Journal of the American Society for Information Science and Technology 246–256.
Liu, Yan, Pei-Yun Hseuh, Rick Lawrence, Steve Meliksetian, Claudia Perlich, and Alejandro Veen. 2011.
“Latent graphical models for quantifying and predicting patent quality”, in Proceedings of the 17th ACM
SIGKDD International Conference on Knowledge Discovery and Data Mining 1145–1153.
Lukach, Ruslan, and Maryna Lukach. 2007. “Ranking USPTO patent documents by importance using random
surfer method (pagerank)”, available at SSRN 996595.
Marco, Alan. 2007. “The dynamics of patent citations”, 92:4 Economics Letters 290–296.
Maurseth, Per, and Bart Verspagen. 2002. “Knowledge spillovers in Europe: a patent citations analysis”, 104:4
Scandinavian Journal of Economics 531–545.
Mehta, Aditi, Marc Rysman, and Tim Simcoe. 2010. “Identifying the age profile of patent citations: new
estimates of knowledge diffusion”, 25:7 Journal of Applied Econometrics 1179–1204.
Meyer, Martin. 2000. “What is special about patent citations? Differences between scientific and patent cita-
tions”, 49:1 Scientometrics 93–123.
Michel, Jacques, and Bernd Bettels. 2001. “Patent citation analysis. a closer look at the basic input data from
patent search reports”, 51:1 Scientometrics 185–201.
Mina, A, R. Ramlogan, G. Tampubolon, and J. Metcalfe. 2007. “Mapping evolutionary trajectories: applica-
tions to the growth and transformation of medical knowledge”, 36:5 Research Policy 789–806.
Moser, Petra, and Tom Nicholas. 2004. “Was electricity a general purpose technology?”, 94:2 American
Moser, Petra, Joerg Ohmstedt, and Paul Rhode. 2014. “Patent citations and the size of patented inventions–
evidence from hybrid corn”. Available at SSRN 2641659.
Narin, Francis. 1995. “Patents as indicators for the evaluation of industrial research output”, 34:3 Scientometrics
489–496.
Narin, Francis, and Elliot Noma. 1985. “Is technology becoming science?”, 7:3–6 Scientometrics 369–381.
Narin, Francis, Elliot Noma, and Ross Perry. 1987. “Patents as indicators of corporate technological strength”,
16:2 Research Policy 143–155.
Nicholas, Tom. 2008. “Does innovation cause stock market runups? Evidence from the great crash”, 98:4
Packalen, Mikko, and Jay Bhattacharya. 2015. “New ideas in invention”, NBER Working Paper 20922.
Palangkaraya, Alfons, Elizabeth Webster, and Paul Jensen. 2011. “Misclassification between patent offices:
evidence from a matched sample of patent applications”, 93:3 Review of Economics and Statistics
1063–1075.
Roach, Michael, and Wesley Cohen. 2013. “Lens or prism? Patent citations as a measure of knowledge flows
from public research”, 59:2 Management Science 504–525.
Sampat, Bhaven. 2010. “When do applicants search for prior art?”, 53:2 Journal of Law and Economics 399–416.
Schmoch, Ulrich. 1993. “Tracing the knowledge transfer from science to technology as reflected in patent
indicators”, 26:1 Scientometrics 193–211.
Schmookler, Jacob.1966. Invention and Economic Growth. Cambridge, MA: Harvard University Press.

Sonn, Jung, and Michael Storper. 2008. “The increasing importance of geographical proximity in knowledge
production: an analysis of US patent citations, 1975–1997”, 40 Environment and Planning A 1020–1039.
Takahiro, Maeda, Sadao Nagaoka, and Yusuke Naito. 2015. “Effects of stronger disclosure rule on applicants’
behavior and on examination efficiency: Evidence from Japan”, Paper presented at EPIP 2015 conference at
the University of Glasgow.
Thompson, Peter. 2006. “Patent citations and the geography of knowledge spillovers: evidence from inventor-
and examiner-added citations”, 88:2 Review of Economics and Statistics 383–388.
Thompson, Peter, and Melanie Fox-Kean. 2005. “Patent citations and the geography of knowledge spillovers:
a reassessment”, 95:1 American Economic Review 450–460.
Tijssen, Robert. 2002. “Science dependence of technologies: evidence from inventions and their inventors”, 31:4
Trajtenberg, Manuel. 1990a. Economic Analysis of Product Innovation: The Case of CT Scanners. Vol. 160.
Trajtenberg, Manuel. 1990b. “A penny for your quotes: patent citations and the value of innovations”, 21:1
Trajtenberg, Manuel, Rebecca Henderson, and Adam Jaffe. 1997. “University versus corporate patents: a
window on the basicness of invention”, 5:1 Economics of Innovation and New Technology 19–50.
Verspagen, Bart. 2007. “Mapping technological trajectories as patent citation networks: a study on the history
of fuel cell research”, 10:1 Advances in Complex Systems 93–115.
von Graevenitz, Georg, Stefan Wagner, and Dietmar Harhoff. 2011. “How to measure patent thickets—a novel
approach”, 111:1 Economics Letters 6–9.
von Wartburg, Iwan, Thorsten Teichert, and Katja Rost. 2005. “Inventive progress measured by multi-stage
patent citation analysis”, 34:1 Research Policy 1591–1607.
Wagner, Stefan, Karin Hoisl, and Grid Thoma. 2014. “Overcoming localization of knowledge—the role of
professional service firms”, 35:11 Strategic Management Journal 1671–1688.
Webb, Colin, Hélène Dernis, Dietmar Harhoff, and Karin Hoisl. 2005. “Analysing European and International
Patent Citations: A set of EPO patent database building blocks”, OECD STI Working Paper 2005/09.
Younge, Kenneth, and Jeffrey Kuhn. 2015. “Patent similarity: a vector space model”, Miméo, Ecole polytech-
nique fédérale de Lausanne.
Youtie, Jan, Maurizio Iacopetta, and Stuart Graham. 2008. “Assessing the nature of nanotechnology: can we
uncover an emerging general purpose technology?”, 33:3 The Journal of Technology Transfer 315–329.
Ziedonis, Rosemarie. 2004. “Don’t fence me in: fragmented markets for technology and the patent acquisition
strategies of firms”, 50:6 Management Science 804–820.

APPENDIX
Figure 2.1A plots the yearly number of scientific articles listed in Google Scholar that
contain the term “patent citation” (solid line), and the number of articles citing the NBER
patent citation data file described in Hall, Jaffe, and Trajtenberg (2001) (dashed line). One
can reasonable assume that the latter group of articles forms a subset of the former group.
1000
Containing “patent citation”
900
Citing HJT
800
700
Number of Articles
600
500
400
300
200
100
0
1980 1985 1990 1995 2000 2005 2010
Publication Year
Note: HJT refers to Hall, Jaffe, and Trajtenberg (2001).
Figure 2.1A Number of scientific articles listed in Google Scholar

3. Patent value
John R. Allison* 20
Contents
I. Introduction
II. Social Value and Private Value
III. How Does Patent “Quality” Fit into the Equation?
IV. The Focus on Private Value
V. Private Value and Patent Portfolios
VI. Identifying Private Patent Value
A. Is There a Litigation-Value Link?
B. Possible Indicators of Private Patent Value
1. Prior art references
2. Patent claims
3. Citations received (“forward citations”)
4. Payment of maintenance or renewal fees
5. Number of countries in which patent protection is obtained
(“patent family” size)
6. Number of patent classifications
7. Claim breadth and words, words, words
8. Patent examination characteristics
9. Licensing
10. Transfers of ownership
11. Litigation outcomes
VII. Directly Measuring Private Value
References
I. INTRODUCTION
What is a patent worth? Can we actually estimate the value of particular patents, or is
merely identifying which patents may have value the best we can do? Indeed, can we even do
that? Given that receiving a patent from the U.S. Patent and Trademark Office (USPTO)
requires a meaningful investment, why don’t all patents have at least some value? Is value
best identified or estimated by examining standalone patents, groups of them, or both?
Can characteristics of patents themselves be employed usefully in identifying patents that
have value? That is, do we know one when we see it? And if so, which characteristics?
* The Mary John and Ralph Spence Centennial Professor, McCombs Graduate School of
Business, University of Texas at Austin.
47

Is there a link between the fact of patent litigation and patent value? And is there any
association between litigation outcomes and patent value? Can information outside the
patent document be useful in identifying patents of value? More fundamentally, what
do we mean in the first place when we refer to patent value? Are there different types of
value? And, value to whom: the owner of the patent or patents in question or the public
at large? Drawing from research in both law and economics performed by many scholars
during the past 25 years, this chapter addresses questions such as these.
II. SOCIAL VALUE AND PRIVATE VALUE

A patent may have “social value,” that is, value to society as a whole in the sense of
positively affecting the well-being of some meaningful portion of the population.1 It may
have “private value,” or economic value to its owner. Any patent that places into the public
repository of knowledge an invention that is truly novel and nonobvious, and that fully
describes the invention such that a skilled person in the field can make and use it, should
necessarily have some social value, even if the invention is not ultimately commercialized.
Much of the normative public discourse on patents is couched in terms of “patent qual-
ity” rather than patent value. Because the concept of “patent quality” essentially refers to
conformity with patentability requirements, it correlates to some degree with both social
and private value, although factors other than or in addition to quality will affect both
types of value. It seems very likely that a patentee will care more about patent quality
when he or she perceives private economic value. Psychologist and Nobel laureate Daniel
Kahneman (2011), among others, has observed that inventors as a class, like entrepre-
neurs, tend to be quite optimistic, thus leading to the fact that large numbers of patents
have no value to anyone. Their optimism is not excessive or unwarranted in every case,
however, so the value perceptions of patent applicants should correlate at some level with
actual private value. Moreover, there is probably a self-fulfilling prophecy at work in at
least some cases, given that perceptions of future value can lead to substantial investment
in obtaining and enforcing stronger patents.
Although the private and social value of patents should correlate positively with one
another at varying levels, and both should be associated with patent quality, the story
cannot end there. A valid patent covering a pioneering invention should, it would seem,
bring more value to the public than one covering a barely nonobvious incremental
improvement on what others had already done. If the claims of such a patent are drafted
broadly enough to take advantage of the invention’s pioneering status, private value to the
owner should be comparably high as well.
Some questions about social value will border on the metaphysical, such as whether,
other things being equal, a valid patent on a device or method in the medical field
that helps save or improve lives has inherently more social value than an equally valid
patent on a manufacturing process or machine that improves the efficiency with which
semiconductor chips can be fabricated. On the other hand, what if the money saved on
1
There is no practical means of knowing whether any invention with or without a patent
constitutes a Pareto improvement.

Patent value 49
chip fabrication is invested in the successful development of software or a new integrated

circuit design that aids in the invention of a genetic engineering technique which helps
stave off the onslaught of Alzheimer’s disease?
Empirical work by legal academics and economists on patent value cannot be neatly
separated into studies of social value or private economic value. Although both concepts
are notoriously difficult to pin down, social value appears to be the more challenging of
the two to identify or measure.
A 2012 study (Barry and Delcamp) attempts to disentangle social and private value,
and is discussed at length here because it serves to highlight some of both the strengths
and weaknesses found in much of the research on patent value. Many other studies in the
field suffer from the same or similar problems as this one. The authors seek to distinguish
private and social patent value in the context of comparing so-called “discrete” and
“complex” technologies. A discrete technology is one in which a marketable product is
protected by a single patent or a small number of them. Pharmaceutical drugs are the
clearest example. Pharmaceuticals, which are created using the technologies of either
traditional chemistry or biotechnology, are also the field in which patents are generally
agreed to be the most economically valuable to their owners. Complex technologies, on
the other hand, are represented by marketable products protected by a large number of
patents, sometimes forming so-called “patent thickets.” Carl Shapiro (2001) has aptly
defined a patent thicket as “a dense web of overlapping intellectual property rights that a
company must hack its way through in order to actually commercialize new technology.”
A modern smart phone is the archetypical manifestation of a patent thicket because the
typical multifunctional phone incorporates a very large number of individual software
and hardware technologies, many of which are individually protected by patents.2 There
are thus many more risks that a competitor will infringe on at least one of these patents
when seeking to commercialize its own technology products.
To distinguish between discrete and complex technologies, the authors employ cat-
egories of patents that are determined by USPTO classifications. Although economists
have relied extensively on the patent classification system to define technology categories,
neither USPTO classifications nor the closely related International Patent Classifications
(IPCs) actually define technology areas at all. The USPTO classes begin at a level too
broad to be useful for academic researchers (e.g., “television”), followed by subclasses
that focus on detailed functions and often operate at too low a level of abstraction to
be of use for research purposes. At a high level, “music,” “fire extinguishers,” “animal
husbandry,” “brakes,” “dentistry,” and the like are not technology categories. At a
low level, “music—stringed—picking devices,” “fire extinguishers—chemical pressure
generating—automatic,” “animal husbandry—milkers—methods of milking—milking
station arrangements—with traveling platforms,” “brakes—vehicle—hub or disc—motor
vehicles,” “dentistry—orthodontics—bracket—having means to secure arch wire—
separable securing means—ligature wire” are likewise not technology areas.
The USPTO classification system and the IPCs were designed to assist patent e xaminers
and other searchers in finding relevant prior art patents, and not to provide a tool for
2
Some of the firmware code’s original expression in a smartphone is also rendered propri-
etary by copyright.

defining technology categories for research purposes. They can be somewhat useful for
some kinds of broad distinctions such as that between pharmaceuticals (an industry, not
a technology) and computer hardware (also an industry, or part of an industry, and not
a technology); but even then one will develop data sets with large data comparability
problems. In addition to being fundamentally unsuited for identifying technology fields
for research, the system is replete with errors, and its classes are often confoundingly
entangled in loops. The placement of patents into patent classifications by the USPTO
and other examining agencies does reveal distinctions among the many possible subjects
of inventive activity, but it reveals nothing approaching the kind of coherent, science-
based typology needed by social science research.3 The use by some researchers of the
number of different patent classifications into which a patent was placed as an independ-
ent variable in efforts to assess patent value is discussed later.
The Barry and Delcamp study properly uses renewal data revealing payment of patent
maintenance fees to the USPTO at post-issuance years four, eight, and 12 as a proxy for
private patent value.4 If one were to choose a single factor to suggest private patent value,
3
The only currently known way to place patents within technology classifications in a way that
avoids large Type I and Type II errors and that is not dependent on the error-prone patent clas-
sification system is to (1) develop concise definitions of technologies through experience over time,
and (2) carefully study the claims and written descriptions of patents to place them within these
definitions. This method, which has been employed by the author of this chapter numerous times
in published studies, has its own problems, including the fact that it is remarkably time consuming
and cannot feasibly be used for more than several thousand patents on any single project. This
approach also is not replicable unless another researcher is willing to devote a very large amount of
time to the approach. Attempts by this author working with others to automate the method with
software have not proved successful.
If one wishes to create a data set of more than several thousand patents for a research study,
the patent classification system is the only available means. It is strongly recommended that, if
the patent classification system is so employed, one or more additional filters also be employed to
diminish the rate of Type I and Type II errors, such as including patents issued to only a specified
set of owners that normally only patent in particular technology categories. Such an approach
is likely to diminish substantially the number of Type 1 errors, but Type II errors will remain
abundant, necessitating care in not overclaiming for the results of the study. An example of this
approach is found in Graham and Mowery (2003). There, the authors initially identified a group
of prepackaged software firms, and then identified the patent classifications for patents owned by
those companies before collecting a larger set of patents from those classifications. This approach
is undeniably better than using only the classifications, and produces a data set with few non-
software patents and thus with relatively few Type I errors. But Type II errors may still abound with
such an approach unless the researchers carefully draw boundaries for the inferences they draw.
The author of this chapter has studied both USPTO classifications and IPCs that purportedly
hold only data processing patents, and has discovered relatively large numbers of patents that do
not actually cover the manipulation of data. The Graham and Mowery approach represents a
useful methodology for developing a data set of patents in a particular field if one wishes to subject
the field to some type of patent-value analysis as long as any claims for the resulting findings are
carefully limited to what was actually studied, in this case only patents on this type of software
and not to software patents as a whole, and as long as Type II errors are expressly acknowledged
along with an analysis of how such errors might affect the results. The majority of software patents
have been issued to large hardware manufacturers rather than to pure software firms (Allison and
Mann, 2007).
4
The patent offices in a number of other nations require that renewal fees be paid annually.

Patent value 51
this is probably the most direct, given that it represents a decision by an owner to invest
in keeping its patent in force rather than allowing it to lapse. It is not without shortcom-
ings, however, as will be discussed below. The study supplements its identification of
patents having private value with a dummy variable for any patent that had been cited
in a lawsuit found in the Lex Machina intellectual property litigation database.5 There is
indeed something different about most patents that have been asserted in litigation when
they are compared with patents that have not been—a topic that is also discussed in more
detail later in this chapter.
The authors then use various indicators as a group to identify social value: citations
received (what some call “forward citations”), references to prior art patents (what some
call “backward citations,” which do not include references to prior publications other than
patents), number of claims, the size of so-called “patent families” (a misnomer for the
number of countries in which a patent owner has obtained patent protection on the same
invention), and what economics researchers have called “generality” and “complexity”
indices. The generality index purports to measure the degree of dispersion across the
technology areas of the prior art patents that patents cite as references, and the complexity
index putatively measures the degree of technology distribution across the patents that
cite the patent in question (“citations received,” or “forward citations”). The generality
and complexity metrics are based on the patent classification system, however, and thus
have limited value in identifying technology fields for research needs (Cohen and Merrill,
2003; Allison and Mann, 2007).
As discussed in more detail below, some of the patent value indicators the authors
chose—including citations received, prior art patent references, number of claims, and the
number of countries in which patents have been acquired on the same invention—have
been employed by a number of other scholars in both law and economics to signal value.
We will see that these and a few other metrics do signal something about the relative
importance of certain patents, but it is incorrect to associate them with social value. To
the extent that these variables suggest value, it is private economic value, just as payment
of maintenance fees clearly does. It is not social value. The number of claims in a patent,
for instance, bears no relation whatever to the value that a patent or a set of them has to
society as a whole.
III. HOW DOES PATENT “QUALITY” FIT INTO THE

EQUATION?
Most scholarship on patent value has not attempted to distinguish between social and
private value, but has instead simply focused on unspecified “value,” or in a number of
instances on what researchers have referred to as “patent quality.” Quality is not the same
thing as value, but at a conceptual level the two are clearly interrelated, quality arguably
being a subset of value. On the other hand, it is definitely possible for a patent or set of
patents to possess quality without having much private or social value. This statement
5
See https://lexmachina.com/. The authors of the study refer to the Stanford IP litigation
database, the name of the entity before Stanford spun it off as a private for-profit entity.

assumes that quality refers to the likelihood that a patent will be held valid if challenged
in USPTO administrative proceedings or infringement litigation, which seems to be a
reasonable definition of the term quality. A patent’s claims might undoubtedly cover an
invention that is novel and nonobvious, and the patent’s written description and drawings
(its “specification”) might clearly provide adequate enablement and written description
support for these claims, but the patent may still have little or no economic value if, for
example, (1) the claims are too narrowly drawn (i.e., the language of the claims is very
specific) to encompass what others are making, using, or selling; (2) the claims can easily
be circumvented by others who produce noninfringing alternatives; (3) the claims cover
an invention that is obsolete even while the patent remains in force; or (4) the claims cover
an invention for which there is no market demand in the first place.6 That said, however,
on average, a patent showing a high likelihood of validity is also likely to be more valuable
to its owner, and possibly to society, than one possessing deficiencies probably leading to
invalidity. Perhaps one can say that quality in many cases is a necessary but insufficient
condition for value.7
IV. THE FOCUS ON PRIVATE VALUE
The “value” that has been the subject of most studies is clearly private value, the economic
value that a patented invention has to its owner. It is unsurprising that most research on
patent value has either explicitly or implicitly focused on the private variety because even
though the notion of private value is a slippery one, any kind of social value assessment
is far more so.
Moreover, most scholarship has not attempted to measure the actual monetary value of
particular patents. Instead, what many view as the “black art” of quantitative valuation
of specific patents or groups of them has largely been left to the privacy of licensing,
sale (assignment) and other negotiations. Work by Schankerman (1998) did attempt to
quantitatively estimate the average value of patents in France across different technology
fields by using data on patent renewal rates.8 The paper offers a good theoretical defini-
6
Saying “undoubtedly” and “clearly” in connection with patent validity on any grounds is
an overreach, because there is always some degree of uncertainty attending the likely validity of
a patent until it has been judicially determined to be “not invalid.” Even then, the validity of the
patent in question can be challenged on other grounds if the owner sues a different defendant for
infringement.
7
One cannot dismiss the possibility, though, that a standalone patent that has a low prob-
ability of being found valid in litigation, and thus is of low “quality,” can nevertheless have some
form of value to its owner, as when the patent is part of a larger portfolio that its owner can use
as a bargaining chip in litigation, or even that a troll might use successfully to extract a so-called
“nuisance settlement” for some amount less than the accused infringer’s estimated litigation costs.
The latter phenomenon is, of course, largely a product of the exceptionally high cost of modern
patent litigation.
8
Patent “renewal” fees in Europe are payable each year beginning with the third year from the
original filing of the patent application, Article 86(1) European Patent Convention, www.epo.org/law-
practice/legal-texts/html/epc/2013/e/ar86.html; whereas patent “maintenance” fees in the U.S. are due
in the fourth, eighth, and twelfth year after patent issuance, www.uspto.gov/learning-and-resources/
fees-and-payment/uspto-fee-schedule#Patent%20Maintenance%20Fee.

Patent value 53
tion of the value of the patent right itself as “the incremental returns generated by holding
a patent on the invention, above and beyond the return that could be earned by using a
second-best means.” Because it is impossible to identify, let alone measure, the value that
a particular invention would have if not protected by patents but instead shielded from
competition as a trade secret or protected solely by private contracts such as licenses,
work on private patent value in legal scholarship has not attempted to identify the value
of only the patent right itself, but has instead sought to recognize patents with value from
the combination of the underlying invention and the patent right.9
In the technology management literature, Ernst, Legler and Lichtenthaler (2010)
sought to theoretically approximate the incremental value of the patent right beyond any
value inherent in the invention itself by employing a Monte Carlo simulation.10 Although
the authors suggest possible empirical studies to follow up on their analysis, one is hard
pressed to envision a workable empirical test of the difference between the value of patent
protection and the value of alternative protection for the invention such as the use of trade
secret law and contractual provisions.
V. PRIVATE VALUE AND PATENT PORTFOLIOS
Before turning in detail to an examination of factors that have been employed to identify
patented inventions suspected of possessing private economic value, an aside on patent
portfolios is warranted. Most research in law and economics has busied itself with the task
of identifying the fact of value for standalone patents or broad groups of them rather than
attempting the far more difficult job of suggesting which patents may have value primarily
or solely because of their incremental contribution to a coherent private portfolio of
patents or to a “thicket” of patents that covers some complex device such as a smartphone.
Parchomovsky and Wagner (2005) expounded a theory patent of portfolio value in sig-
nificant part to help explain the so-called “patent paradox.” The paradox—observed from
a large survey by Levin, Klevorick, Nelson, and Winter (1987)—is that, although R&D
executives at businesses of all sizes in all industries except pharmaceuticals view secrecy
and “first-mover advantages” (i.e., lead time) as much more effective for appropriating
returns on R&D investments than patents, corporations nevertheless continued patenting
9
Also, Schankerman’s study determined technology sectors by the use of patent classifica-
tions. Using this flawed method for defining technologies is perhaps the best one can do when
studying more than a few thousand patents, but any results based on the use of these classifications
must be viewed very cautiously.
10
The authors summarize their study and conclusions as follows:
We rely on Monte Carlo simulations with data from a case study in a large chemical firm to
estimate patent value according to our model. In the simulation analyses, we compare an R&D
project with patent protection and the same project without patent protection. The difference of
the values of the two projects is the surplus in profit that may be expected from having a patent
covering the project. This surplus is regarded as the value that is directly attributable to the
patent. The results of the simulation analyses indicate that the development costs and expected
net cash flows of a patent-protected project are higher than of an unpatented project. The higher
net cash flows outgrow the increased development costs, and patent value is positive. However,
this value is smaller than the overall project value of the patent-protected R&D project.

relentlessly and at accelerating rates. A later large survey of R&D executives by Cohen,
Nelson, and Walsh (2000) led to the undeniably correct conclusion that a large portion
of patents, especially those issued to large companies, were not acquired for the purpose
of protecting particular technology products, but instead were obtained for defensive
purposes (Goldenberg and Linton, 2012).11
Regarding portfolio value, the statement by Parchomovsky and Wagner (2005) that
“We find that for patents, the whole is greater than the sum of its parts: the true value
of patents lies not in their individual worth, but in their aggregation into a collection of
related patents – a patent portfolio,” is clearly true for many patents, especially for those
owned by large companies that are far more likely than individuals, small companies,
or universities to obtain large numbers of patents on related technologies for any of
several purposes. As noted by Wesley Cohen and his colleagues (2000), and as generally
known by many patent practitioners, large companies develop extensive patent portfolios
for defensive purposes so that they might better defend against infringement claims by
competitors in a take-off on the existential “mutually assured destruction” model of
nuclear armament. They likewise build portfolios to create more freedom of movement
for their R&D investments, to keep rivals at a technological distance, and sometimes
to build thickets of patents that enable complex pieces of hardware. Individual patents
or small sets of them that sometimes originate with the same original application are
typically more valuable to small entities of different types (Allison, Lemley, Moore, and
Trunkey, 2004).
Although single patents are more likely to have value to small entities,12 patent portfo-
lios are often important even to small entities for some of the same reasons that they are
important to large ones, except that they are likely to have smaller collections of patents
presenting fewer opportunities for cross-licensing to short-circuit litigation.13 However,
patents issued to individuals and small entities are more likely than those granted to large
entities to be involved in infringement litigation, in part because the large portfolios of
the latter provide them with more to trade (Lanjouw and Schankerman, 2003). The value
of patent portfolios of any size to any size or type of entity must naturally vary across
technology and industry sectors, as well as across firms and within firms. Relative to single
patents, relatively large portfolios should be more valuable in the software technology field
and in the information technology industry more generally than in technologies such as
chemistry and biotechnology and industries such as pharmaceuticals, because technical
11
An unrelated use of the term “patent paradox” has been to describe the asserted phenom-
enon of stronger patent laws with longer durations allowing greater profit to the inventor, but at the
same time discouraging related innovation as the protection for the underlying technology becomes
broader and the duration is longer (Goldenberg and Linton, 2012).
12
The USPTO defines a small corporation as one with fewer than 500 employees. A
company operating a non-labor intensive business may, of course, have very substantial assets
and sales while still being classed as a small entity for discounted USPTO fee payment purposes.
The USPTO also includes individuals and universities within its definition of a small entity.
Universities do not, of course, engage in manufacturing and thus do not have a need to cross-
license, although the manufacturers to which they grant licenses (typically exclusive licenses)
usually are manufacturers and may have the same need to use a patent portfolio defensively as
any other practicing entity.
13
Universities, of course, rarely engage in cross-licensing because they do not manufacture.

Patent value 55
advances tend to be smaller and more incremental in the former than in the latter. The
previous statement also is a partial reflection of the distinction between discrete and
complex technologies, to use the words of some researchers. Also, research has observed
that litigated patents as a group were issued to a larger portion of small entities than were
unlitigated patents (Allison, Lemley, Moore, and Trunkey, 2004).
Satisfactory methods for measuring, or indeed even identifying, the special value that
patent portfolios can have because the value of the whole is greater than the sum of its
parts, or the value that any particular patent might contribute to a portfolio that is beyond
any value that such patent might have on its own, have not been developed. Scholars have
identified large groups of patents that seem to have more value than other large groups of
patents, as with research on litigated versus unlitigated patents (Allison, Lemley, Moore,
and Trunkey, 2004), but the sets of litigated and unlitigated patents identified have not
consisted of any particular owner’s coherent portfolio.
Reitzig (2003) did examine the characteristics of a single company’s portfolio of
patents, his study focusing on 127 German or European patents and applications having
filing dates between 1979 and 1999, all but a few having German priority. The patents
and applications were owned by a single “large” semiconductor firm. In a 1999 survey
conducted by the author, a technology unit within the firm identified the patents and
applications as being held solely as “bargaining chips” for cross-licensing negotiation
purposes. None of the issued patents had been asserted in litigation. Each patent or
application was evaluated by four separate teams of experts from within this unit of the
patent-owning firm, each team consisting of between one and four engineers and one
marketing expert.14 Using an ordinal scale for each patent or application, these teams
assessed several characteristics such as novelty, technical advance, quality of technical
information disclosure, patent age, and difficulty that would be encountered by competi-
tors in attempting to “invent around.”
The teams then evaluated, again on an ordinal scale, the present value of each patent
or application as a function of how interested the owning firm would be in buying the
patent were it owned by the firm’s strongest competitor either at the time of application
filing or at the time of the interview by the study’s author. The author found correlations
between patent characteristics and value estimates at a medium level on a one-to-seven
scale, the greatest positive correlation being that between the age of patents and their value
estimates. The latter finding revealed that, in this set of semiconductor patents owned by a
single large firm, greater age was correlated with greater value to the firm, with the caveat
that the average age of patents in the data set was only about five years. The primary
value of this study appears to be its confirmation of several predictors of value that seem
quite intuitive. The paper’s finding regarding patent age is less intuitive, and probably
would not continue to hold true as the patents in question became even older because of
the increasing probability of obsolescence or the development by rivals of noninfringing
alternatives. The study is also notable as an unusual and well-conceived attempt to assess
the value of a single firm’s coherent patent portfolio, although, as the author candidly
admits, the generalizability of the findings is uncertain.
14
If a team was unable to reach consensus on how to evaluate a patent, the opinion of the
most senior engineer in the group was used.

VI. IDENTIFYING PRIVATE PATENT VALUE
A. Is There a Litigation-Value Link?
Relatively early research in economics and law associated litigation propensity with
private patent value, under the view that the extremely high costs of patent enforcement
would generally lead patent owners to sue for infringement only in the case of patents they
deem to be important (Lanjouw and Schankerman, 1999). “Importance” is surely associ-
ated in some way with private value, although this value is likely to take various forms. It
may be value associated with protecting technology embodied in a marketable product,
creating obstacles for competitors that may sue the patent owner for infringement of the
competitors’ patents, establishing a reputation for actively enforcing patents to ward off
other firms from getting too close to an owner’s protected technological territory, or even
the ability of a patent assertion entity (a subtype of “non-practicing entity” (NPE)) to
extract rents by means of settlements for less than the defendant’s projected litigation
costs. Whereas the first three versions of value to patent owners are either wealth creating
or at least not purely redistributive, the latter is like to be mostly redistributive in many
cases, especially in those instances in which the patent assertion entity plays no role in pro-
viding remuneration to the original inventors of the patented technology. Several patent
characteristics, including the number of references to patent and nonpatent (“printed
publications”) prior art, number of claims, number of related patents issued from the
same original application,15 and number of countries in which a patent is obtained on
the same invention in different countries (sometimes referred to by the misnomer “patent
family”) all require significant investment by patent applicants. Research shows that
these characteristics not only are associated with litigation propensity and independent
measures of patent value, but also tend to be correlated with one another. The correla-
tion among some of these characteristics suggests a perception among certain patent
applicants that their invention will have value to them, leading to greater investment in an
attempt to obtain and enforce stronger patents (Allison and Tiller, 2003).16
There clearly is something about patents asserted in infringement litigation that sets
them apart from patents that are not. Research by several scholars has found litigated
patents to have different characteristics than unlitigated ones; or perhaps it is more accu-
rate to say that litigated patents on average have characteristics that are greatly amplified
in comparison with contemporaneously issued or filed unlitigated ones.17 For example,
15
The number of patents issued to a firm on closely related technologies, which in the U.S.
will often spring from the same original patent application, is a metric that is better described as a
“patent family.”
16
The authors found a strong correlation between number of claims and number of prior art
references.
17
The authors found that, compared with unlitigated patents, contemporaneously issued
litigated ones have significantly more patent and nonpatent prior art references, have more claims
(and notably, more independent claims), are cited more often by later patents, are characterized by
more technology areas per patent, and spent more time in prosecution. Allison and Sager (2007)
refuted criticisms by another researcher of the statistical methods used in Allison, Lemley, Moore,
and Trunkey (2004). Allison, Tiller, and Zyontz (2012) likewise found litigated patents to have more
independent claims and citations received than unlitigated ones.

Patent value 57
Allison, Lemley, Moore, and Trunkey (2004) discovered that, when compared with
contemporaneously issued unlitigated patents, litigated patents have significantly more
prior patent and nonpatent prior art references, claims (including, more importantly, a
larger number of independent claims),18 and more technology areas per patent. Litigated
patents also were found to be cited more by later patents. Moreover, patents that are
asserted in litigation multiple times possess characteristics that are even more amplified
than those of patents that are the subject of litigation only once (Allison, Lemley, and
Walker, 2009). Comparing patents that were litigated eight or more times during an
eight-year period between the beginning of 2000 and the end of 2007 with those having
been litigated only once during the same period, the authors found that when compared
with once-litigated patents, the “most-litigated” ones cited significantly more patent and
nonpatent prior art, received significantly more citations from later patents—“forward
citations” (and notably, significantly more self-citations), had more technology areas per
patent, had more pre-litigation ownership changes, and were younger (Allison, Lemley,
and Walker, 2011).19 The fact that litigated patents, including those litigated many times,
tend to be relatively young may also suggest private value because the younger the patent,
the fewer the number of noninfringing alternatives there are likely to be, and the lower
the probability of invention obsolescence is likely to be (Allison, Lemley, Moore, and
Trunkey, 2004; Allison, Tiller, and Zyontz, 2012).20
B. Possible Indicators of Private Patent Value
Researchers have identified several patent characteristics as suggesting patent quality and
value, several of them having a clear association with litigation. Several research efforts
have identified what might be termed a “litigation link”—an association between the fact
of litigation of patents and quality or value.
1. Prior art references

One proxy for patent quality and value that scholars have identified is the number of prior
art references together with, to the extent practicable, some assessment of the types and
informational content of those references and their sources. Prior art is objective evidence
of what has been done previously in the relevant technology field. The most important
types of prior art that can be gleaned from patents themselves are references to prior
U.S. and foreign patents and to prior printed publications of various types (often called
18
The difference in the number of total claims between litigated and unlitigated patents is
driven by the difference in the number of independent claims, not dependent ones.
19
Like Allison, Lemley, Moore, and Trunnkey (2004), Allison, Lemley, and Walker (2009) did
not employ patent classifications to identify technology fields, but rather used manually created and
coded technology categories based on much trial and error in multiple studies. In a follow-up study,
the authors also found that the owners of the most-litigated patents overwhelmingly lost their
patent lawsuits, which is likely an artifact of technology area (most were software patents, especially
those in the software communications field) and type of owner (the owners of the most-litigated
patents were predominantly patent aggregators, or “trolls”) (Allison, Lemley, and Walker, 2011).
20
Allison, Tiller, and Zyontz (2012) found that both software business method patents and
patents in a randomly selected set of contemporaneously issue patents from the general population
had an average age of 4.5 years when litigation was instituted.

“nonpatent prior art”). It is intuitively appealing to view the quantity and, to the extent
we can measure it, the quality of prior art cited in patents as indicators of patent quality.
Both should correlate with (1) the seriousness of the patent applicant’s effort to identify
previous inventions and distinguish the new invention from prior ones, and (2) the rigor
and thoroughness of the USPTO’s examination. Furthermore, the most common basis
for judicial invalidation of patents in the U.S. is prior art that had not been considered by
the USPTO (Allison and Lemley, 1998).
This is not to say that the number and quality of references are perfect indicators of
patent quality. For example, the fear of being charged with inequitable conduct leads
some applicants to load up their applications with virtually any prior art reference they
have come across, regardless of relevance.21 However, patentees’ ex ante perceptions about
value should have some reliability, because patentees at the time of filing and prosecution
have a considerable amount of value-relevant information about the market in which they
will deploy the claimed invention. In a large data set of patents, it is reasonable to think
that patents disclosing more prior art references reflect a level of effort on the part of
the applicant and examiner that is higher on average than the effort involved in patents
that disclose fewer references. For the applicant, greater effort also means greater invest-
ment. Moreover, other indicia revealing greater value perceptions and investment in the
patenting enterprise, such as the number of claims and subsequent self-citations, correlate
strongly with the number of prior art references (Allison and Tiller, 2003; Allison, Tiller,
and Zyontz, 2012).22
Given the imperfection of inferences drawn from prior art, additional information
about patent value is needed to buttress those inferences. For example, suppose, as seems
likely, that patent owners care more about quality when they expect their patents to be
valuable. Research by Harhoff, Scherer, and Vopel (2003) found a significant association
between the number of references to prior patents and success in German opposition
proceedings. Such success should be a reasonable indicator of private value.23
Regarding the informational value of prior art references, the best one can do with
references to prior patents is to count them and identify the countries in which those
patents were issued. In a few instances, their country of origin could possibly provide
information about informational quality, but an effort to make such distinctions would be
quite speculative. Assessing the probable informational quality of references to nonpat-
ent prior art—printed publications—is more feasible, however, and may aid in a rough
assessment of patent quality and perhaps value. Allison and Tiller’s (2004) comparison
of Internet business method patents with patents from the general population, and later,
21
Under so-called “Rule 56,” USPTO rules require that a patent applicant disclose any
relevant, material prior art of which he or she is aware, which is sometimes referred to as the “duty
of candor.” 37 CFR § 1.56 (1977).
22
Allison and Tiller (2003) found a strong correlation between number of claims and number
of prior art references, as did Allison, Tiller, and Zyontz (2012).
23
Any inferences drawn from prior art references are complicated, however, by the fact that not
all prior art is cited by applicants. Examiners also add citations, and the USPTO began identify-
ing examiner-added references on the fact of patents in 2001. A 2008 working paper found that,
between 2001 and 2003, 40 percent of all applications included prior art cited by the examiner.
Alcácer, Gittelman, and Sampat (2008).

Patent value 59
Allison and Hunter’s (2006) comparison of business method patents in USPTO main
class 705 with those in secondary class 705, examined the source of every nonpatent prior
art reference, classified them into categories based on their estimated degree of objectiv-
ity and reliability, and then compared these classifications between data sets.24 The two
studies employed the same set of nonpatent prior art categories, except that the later study
separated academic and trade publications, as shown below:
(1) Academic Publications: This category represents publications of a type for which there is
an independent intermediating influence such as one or more editors or referees to increase
the probability of accuracy, reliability, and objectivity, and which are targeted primarily
at an academic, scholarly audience. Academic books, book chapters, journal articles, and
academic proceedings papers, which have been independently screened for accuracy and
objectivity, are the primary components of this category. Academic publications are likely
to be the most objective and reliable nonpatent prior art references because of the rigorous
peer-review process to which such publications are typically subjected.
(2) Trade Publications: This category includes trade books and chapters, trade journal articles,
and similar items. Trade publications are targeted primarily at a practitioner audience
rather than an academic one and report on developments in a field rather than create new
knowledge in that field as academic works are more likely to do. Like academic publica-
tions, trade publications are a type of nonpatent prior art for which there is an independent
intermediating influence such as one or more editors or referees to increase the probability
of accuracy and objectivity. Although these publications are quite unlikely to be subject
to the same degree of rigorous peer review as academic publications, they nevertheless
constitute prior art of relatively high quality and are a good reflection of the state of the
art at the time of publication.
(3) University Publications: This category includes publications from universities or consortia
of universities, such as those from university research labs, departments (such as computer
science, electrical engineering, information systems, business, etc.), individual faculty, and
graduate student theses/dissertations. Because these types of publications are developed
in an environment of objective academic inquiry, they typically will be prior art of good
quality although this quality is probably quite variable.
(4) Software: This category includes software programs and software documentation. These
are separated from other company- or industry-sponsored publications because of their
functional nature and obvious need for a high degree of accuracy and objectivity compared
with less functionally motivated company-sponsored prior art. Software and software
documentation therefore represent prior art of comparatively high quality.
(5) Patent-Related: This category includes published patent applications and patent office search
reports, such as PCT (Patent Cooperation Treaty) and EPO (European Patent Office) search
reports. Such publications are likely to be of highly variable quality as prior art. Published
patent applications are of uncertain quality as prior art because they have not yet been exam-
ined or otherwise tested. Published search reports are likely to be more objective and reliable
than published applications because of the involvement of independent search authorities.
(6) Government Documents: This category includes documents published by U.S. and foreign
governments and by international government organizations such as the World Intellectual
24
Allison and Tiller (2003) provided statistical comparisons between several categories of
nonpatent prior art categories in a population of software business method patents and a random
sample of contemporaneously issued patents from the general population. Allison and Hunter
(2006) provided statistical comparisons of several categories of nonpatent prior art in patents that
had been assigned to USPTO main class 705 and in secondary class 705 as part of an effort to
empirically evaluate the effect of the USPTO’s 2000 “second pair of eyes review” on the quality and
value of software business method patents.

Property Organization (WIPO), as well as websites sponsored by such entities. The category
does not include U.S. and foreign patent-related documents such as published patent
applications and search reports, which are treated separately because of their special nature.
The quality of government documents as prior art is likely to be extremely variable.
(7) Company/Industry Publications: This category includes press releases, websites, advertise-
ments, technical disclosure bulletins, and various other publications that were produced by
individual companies or industry groups and published with no independent intermediat-
ing influence to increase the probability of accuracy and objectivity. It does not include
software and software documentation, however, because these are sufficiently distinct from
and inherently more reliable than other types of publications from companies or industry
groups. After removing software and software documentation from the category, company-
and industry-sponsored publications overall cannot be treated as high quality prior art.
(8) Popular Press: This category includes not only newspapers, magazines, and other publica-
tions of general interest, but also news publications aimed at general business and legal
audiences. The relative quality of such publications varies greatly, but overall is relatively
low.
(9) Other: Includes sundry items such as individual webpages, but most references placed in
this category are those in which insufficient information was provided for determining what
the item really was, even after we conducted a web search of key names and terms in the
incomplete reference. One example is a reference to a partial title of an item, followed by
“found on the web on x date.”25
2. Patent claims

Like the number of prior art references, the number of claims in a patent relates intuitively
to private economic value. Following the detailed written description of the invention
in a patent, claims ideally identify the invention with linguistic precision and define the
patent owner’s property interest with similar precision.26 Having a patent attorney draft
more claims necessarily costs more money. Moreover, increasing the number of claims
in a patent can sometimes increase the universe of potential infringers and the likelihood
that the patent will be held to extend to competing products by providing the drafter
with more opportunities to use a variety of different language formats in claiming an
invention. Although researchers in economics had identified the total number of claims
as a predictor of litigation and value (Lanjouw and Schankerman, 1997), later empiri-
cal scholarship in law (Allison, Lemley, Moore, and Trunkey, 2004; Allison and Sager,
2007) found that the significant association between the number of claims and litigation
propensity is entirely driven by the number of independent claims, not dependent ones.27
Litigated patents have significantly more total and independent claims than unlitigated
25
Id. at 741–744.
26
Anyone with substantial experience in studying the text of patent claims, and the interpreta-
tion of them by federal courts, knows that the ideal cannot be realized.
27
Patent claims are either independent or dependent. As the term implies, an independent
claim stands by itself. An independent claim is usually relatively broad in its coverage by employing
more general language, which increases the universe of potential infringers. More general language
also translates into a higher probability of being invalidated for any of several reasons, including
claim indefiniteness and overlap with prior art. After each independent claim a patent usually
contains one or a series of dependent claims, each normally adding more specificity to a functional
element from the independent claim. Increased specificity narrows the technological reach and
decreases the number of potential infringers, but causes the dependent claim to be less vulnerable
to an invalidity finding. Thus, if an independent claim is invalidated in infringement litigation,

Patent value 61
ones (Allison, Lemley, Moore, and Trunkey, 2004). Moreover, Moore found that, in a
logistic regression model, the number of claims was a significant predictor of continued
payment of maintenance fees by owners (Moore, 2005).28 Payment of maintenance fees
as a value indicator is discussed below.
A further observation on independent claims is perhaps warranted. It is common
to find multiple independent claims in a patent in recognition of the fact that human
language is highly imperfect, even more so when it is used in an attempt to distinguish
the functional details of an invention from all other similar inventions. Patent drafters
often write multiple independent claims that cover the same invention, using different
verbal formats to increase the odds that a later product that is substantively identical or
similar to the disclosed invention will not “slip through the cracks.” The value of having
different independent claims on the same invention is aptly, albeit anecdotally, illustrated
by NTP, Inc. v. Research In Motion, Ltd. (2005), in which NTP successfully sued Research
In Motion, the maker of the BlackBerry personal communication device, for infringement
of several NTP patents. The United States Court of Appeals for the Federal Circuit
concluded that the BlackBerry functioned in a way that did not violate the independent
“method” (i.e., process) claim of the NTP patent, but did infringe the NTP patent’s
“system” independent claim.
3. Citations received (“forward citations”)

Another intuitive and empirically validated indicator of patent value is the number of times
that later patents cite a particular patent as prior art (“citations received” or “forward cita-
tions”). It is reasonable to expect a correlation between the number of citations received
and the relevance of the patent to continuing developments in the applicable technological
field.29 The number of citations received similarly relates to the likelihood that the patent
disclosed a fundamental development in the particular technology field, thus giving the
patent owner a valuable “head start.” Moreover, later citations to a patent by the owner
of the earlier patent (“self-citations”) suggest that the patent owner is building a group of
patents on closely related technological advancements, which suggests a greater likelihood
there may still be a finding of infringement if the defendant is making, using, or selling a product
that falls within the more specific confines of a dependent claim.
28
The number of claims also has been found to be significantly correlated with the number of
prior art references (Allison and Tiller, 2004).
29
One cannot simply use the raw number of citations received, however. As patents age, the
opportunity they have to be cited as prior art in other patents increases, although not in a linear
fashion. Thus, to properly use forward citations as a research metric, the absolute number of
citations received must be adjusted, or “standardized” to account for the different ages of patents
in a data set. A commonly used method of adjustment, suggested by Hall, Jaffe, and Trajtenberg
(2002), involves placing each patent in the data set into a cohort of other patents in the data set that
were issued during the same year. Thus, each cohort is one year, although cohorts of more than one
year could be used if necessary to solve a small numbers problem even though this would decrease
precision. The number of forward citations received by each patent is divided by the average
number of forward citations received by other patents in the same cohort. This gives the adjusted
number of forward citations for that patent in the data set. The process is repeated for every other
patent in the same cohort and then repeated for each patent in the other year cohorts. To obtain
the adjusted number of forward citations for an entire data set, the quantity of adjusted number of
forward citations received by all patents in the set is then averaged.

of commercial exploitation of the patent. Hence, it is no surprise that empirical studies

have found significant positive correlations between the number of forward citations and
firm market value, and between the former and likelihood of litigation.
Lanjouw and Schankerman (1999) identified citations received as a patent value indica-
tor, and a study by Hall, Jaffe, and Trajtenberg (2005) concluded that each additional
forward citation was associated with a 3 percent increase in firm market value. The same
Hall, et al. study found that “self-citations” were even more strongly associated with firm
market value than citations by others.30 In addition, Harhoff, Scherer, and Vopel (2003)
found that the number of citations received was a predictor of successful outcomes in
German patent opposition proceedings. Moser, Ohmstedt, and Rhodes (2015) discovered
a correlation between citations received and improvements in yields on patented hybrid
corn, the association being robust for alternative measures for improvements in the crop.
Other research shows that litigated patents received significantly more citations than
unlitigated patents (Allison, Lemley, Moore, and Trunkey, 2004), and that both external
citations received and the self-citations subset were significantly greater in patents that
had been litigated multiple times than in those litigated only once (Allison, Lemley, and
Walker, 2009).
Although some researchers have used a patent value metric calculated by dividing the
number of forward citations by the total number of claims, there appears to be no empiri-
cal or logical basis for doing so (Lanjouw and Schankerman, 2003).
Regarding both patent citations made and those received, as well as the litigation-value
link, recent but yet unpublished work by Torrance and West using eigenvector centrality
and hierarchical graphical approaches shows that litigated patents have vastly larger and
more complex networks of citations made and citations received than patents that have
not been litigated (Torrance and West, 2016).
4. Payment of maintenance or renewal fees

Intuitively, the willingness of patent owners to keep their patents alive by paying
“maintenance fees” (called “renewal fees” in Europe) should be an indicator of perceived
economic value. As shown by Kimberly Moore (2005), the failure of the owners of most
U.S. patents to pay the relatively modest maintenance fees necessary to keep their patents
30
Allison, Lemley, and Walker (2009) made the point that use of a relatively accurate decision
model for identifying self-citations is preferable to simply counting later citations having the same
assignee as other researchers had done. Because there are often multiple inventors on a patent, and
because ownership of a patent can change after issuance, there can be difficulty in identifying a
particular forward citation as a self-citation. Thus, Allison, Lemley, and Walker (2009) used the
following decision rule for identifying self-citations: A forward citation is a self-citation if either
(1) the owners of the main patent and the forward citation are the same, or (2) the owners are
different, the inventors in the main patent and the forward citation are the same, and there are no
co-inventors (i.e., no other inventors). To apply this decision rule, they had to examine the front
pages of each of the 3419 patents that constituted forward citations to patents in their two data
sets. Other researchers have not examined individual forward citations and thus used a blunter test
to identify self-citations: a forward citation in which the assignee (owner) of the patent is the same
in the cited and citing patents. See, e.g., Hall, Jaffe, and Trajtenberg (2002).
Like all citations received, self-citations must be adjusted to account for the varying ages of
patents in a data set.

Patent value 63
in force suggests that more than half of all patents are not worth even a few thousand
dollars a few years after they are issued. The intuition that payment of maintenance
fees relates to private patent value is buttressed by empirical research identifying strong
positive correlations between the indicators of value discussed above and maintenance
fee payments. Specifically, Moore concluded that patents with more claims and forward
citations are more likely to be maintained.
The payment of maintenance fees as a factor by itself could be somewhat limited as
a value indicator, however, if evidence were to show that, as is likely, corporate owners
with very large portfolios renew patents en masse in particular fields of activity where
they need to defend against competitive encroachments and against infringement claims
by competitors rather than giving careful consideration to the maintenance decision for
individual patents or small groups of them. The great majority of all U.S. patents are
owned by large companies as part of large portfolios. Undoubtedly, though, patents that
have been continued in force by fee payments should be more valuable than those that
have not.
5. Number of countries in which patent protection is obtained (“patent family” size)
Another indicator of value used in the existing literature is the number of countries in
which the owner has obtained patent protection on the same invention (often referred
to as a “patent family”).31 This makes sense because of the large expense of patenting
in multiple countries, and because patent protection is limited to the nation in which the
patent is granted, the number of different countries in which patents are obtained on the
same invention should correlate with the perception, and sometimes the reality, of wider
geographic markets for the patented technology. Putnam (1997) and Lanjouw, Pakes, and
Putnam (1998) made relatively early uses of what they referred to as the “size of the patent
family” as a value indicator. Lanjouw and Schankerman (1999) developed a patent quality
index that relied on the number of claims, prior art references, forward citations, and the
number of countries in which the patentee sought protection for the invention, finding,
inter alia, that the size of a patent family is significantly associated with the likelihood of
litigation. And Harhoff, Scherer, and Vopel (2003) found a significant association between
the number of countries in which patent protection was acquired and successful outcomes
in German opposition proceedings.
6. Number of patent classifications

A few researchers have sought to use the number of different USPTO or IPC classifica-
tions into which patents have been placed by the examining agency as a measure of
technological breadth and, consequently, of private patent value. A relatively early and
ingenious attempt to do so by Lerner (1994) examined patents issued to young biotechnol-
ogy firms in the Boston area and found that the number of four-digit IPCs per patent was
significantly associated with these firms’ success in obtaining financing. Lerner equated
31
Instead of identifying the number of countries in which a patent has been obtained on the
same invention, a better use of the term patent family is to denote the number of patents on closely
related technologies. In the U.S., which allows patents to issue from continuation applications,
multiple patents can issue from continuing applications of various types that all relate back to the
same original application.

the number of patent classifications per patent as a proxy for patent “scope” or “breadth.”
Using a sample of 535 financing rounds at 173 privately held venture-backed biotechnol-
ogy firms, Lerner found that one standard deviation increase in the average number
of four-digit IPCs was associated with a 21 percent increase in the firm’s value. Broad
patents, Lerner stated, are more valuable when substitutes in the same product class are
plentiful, a finding consistent with theoretical suggestions.32 However, later research has
not been able to generalize on Lerner’s findings in different contexts.33 Guellec and Van
Potterie (2000; 2002) found no association between the number of patent classifications
per patent and the success of patent applications in achieving issued patents. They found
that issued patents actually were characterized by a smaller average number of four-digit
IPCs than abandoned or rejected applications, which were presumably less valuable than
issued patents. In addition, using outcomes in German patent opposition proceedings as
a measure of private patent value, Harhoff, Scherer, and Vopel (2003) found no statistical
association between the number of four-digit IPCs and positive outcomes. Likewise,
Allison, Lemley, Moore, and Trunkey (2004) could find no positive association between
litigation likelihood and the number of USPTO or IPC classes.
Other things being equal, broader patents should indeed be more valuable, but the
number of different patent classifications is not a measure of patent breadth. The number
of different actual technology areas is, however, a significant predictor of litigation. If,
as many believe, the fact of litigation is associated with private patent value, the number
of actual technology areas per patent, when properly measured by studying the patents
themselves, should be a value indicator.
7. Claim breadth and words, words, words

Assuming validity, a broader patent is inherently more valuable than a narrower one
because patent breadth should translate into a larger universe of potential infringers.
The problem, of course, is that broader patents are more likely to be invalid than nar-
rower ones because they are more likely to overlap with prior art and fail the novelty or
nonobviousness requirements, suffer from fatal indefiniteness, or be noncompliant with
the enablement and separate written description disclosure requirements.
The fact that broader patents are more likely to be invalidated than narrower ones
causes the relationship between breadth and value to be murky, given the substantial
uncertainty associated with any prospect of invalidity before a court has definitively ruled.
Consequently, broader patents may or may not be more valuable relative to those that
reach less far technologically. The relationship between breadth and value will, generally
speaking, be idiosyncratic to a particular patent, or to a group of closely related patents.
Putting aside the increased invalidity risk posed by broad patent claims, it is necessary
to understand what is really meant by patent breadth. And when we say patent breadth,
we necessarily mean claim breadth. Such breadth is not attributable to the number of
total claims, as many have supposed. Patent breadth may sometimes be increased by the
32
As previously noted, the number of patent classifications per patent likely reveals something,
despite the fact that neither the USPTO nor the IPC classifications truly identify technology fields.
33
Lerner and others appear to believe that the classifications into which U.S. patents are
placed are the result of very careful decisions by patent examiners. Those having more familiarity
with the process, however, know this to be untrue.

Patent value 65
number of independent claims, but not necessarily so, and maybe not even usually so.
Breadth certainly is not the number of classifications into which examining authorities
have placed a patent or a set of patents. Patent breadth is increased, however, by the rela-
tive generality of the language within claims, particularly within independent claims. This
seems to be always the case, or at the least subject to only unusual exceptions. Language
generality within independent claims is thus the true measure of patent breadth. Exactly
when a word or phrase in a patent claim is relatively more general or specific, however, is
yet another question having what is likely to be a highly idiosyncratic answer.
Developing a consistently applicable metric for language generality has proved difficult,
precision almost certainly being unattainable. Some observers have regarded the average
word count per independent claim as a coarse estimator of breadth. In 2016, Oracle Corp.
proposed to the USPTO that a close additional review by an examiner and supervisory
examiner be given to patents with any allowed independent claims of fewer than 300
characters, or about four lines of text, because according to the company the USPTO
has almost never allowed independent claims this short. Oracle observed that its study of
patents granted to itself and a number of other large software and computer hardware
firms revealed that 1 percent or fewer of these firms’ issued patents included such a short
independent claim.
Malackowski and Barney (2008) observed, for example: “While claim breadth cannot
be precisely measured mechanically or statistically, counting the average number of
words per independent claim in an issued patent can serve as rough proxy if taken from
a sufficiently large, statistically relevant sample.” If no pretense of precision is adopted,
this is probably true. There is strong intuitive merit to the idea of an inverse relationship
between the word count of a patent claim and the breadth of its technological reach,
although testing the proposition will be fraught with uncertainty in light of the fact that
claim breadth is an inherently imprecise concept to begin with. And the immediately
following explanation by these authors carries things too far: “That is because each word
in a claim introduces a further legal limitation upon its scope.” As an absolute statement,
this is simply not true. A single word may or may not represent an additional restriction
on the technical reach of a patent claim.
A 2016 study found that the number of words in patent claims bears an inverse relation-
ship to the number of citations received by Japanese patents (Okada, Naito, and Nagaoka,
2016). That is, the shorter the claim measured by the number of words, the more adjusted
forward citations were received by Japanese patents in the data set. Evidence discussed
elsewhere in this chapter regarding such “citations received” as a demonstrated value
indicator lends additional credence to the idea that shorter independent claims tend to be
more valuable, again assuming validity.
8. Patent examination characteristics

Marco and Miller (2017) tested for associations between certain characteristics of the
USPTO patent examination process and litigation propensity. They separated their chosen
independent variables into logical groups identified as (1) application characteristics, (2)
examination characteristics, (3) patent characteristics, and (4) post-grant characteristics.
They also used a small set of patent value indicators, including the number of countries in
which patent protection had been achieved for the same invention (what some economists
term “size of the patent family”), the number of citations patents received in later patents

citing them as prior art (referred to by some economists as “forward citations”), the continued
payment of maintenance fees by patent owners to prevent patent expiration, and whether
patents are “standard-essential patents” (disclosed by a patent owner to a “standard-setting
organization”), in a technology field. These are well-chosen value indicators.
Among other findings, the authors found significant positive associations between the
likelihood of patent litigation and patent value indicators, and between patent litigation
and certain examination characteristics, namely: (1) the applicant going through the
appeals process for rejected applications within the USPTO; (2) the application being
examined by a relatively more senior examiner who has signature authority, in contrast
with a relatively more junior examiner without signature authority; (3) the number of
interviews between the applicant’s legal representative and the examiner; and (4) the
length of patent prosecution. Variables (1), (3), and (4), along with a few other independ-
ent variables associated with the examination process, most likely reflect the ability and
willingness of a patent applicant to invest more resources in the patenting effort, probably
because of a perception by the applicant of relatively greater importance or value. And
(4) is also most likely driven by the applicant having filed and pursued a greater number
of ancestor applications leading up to the application resulting directly in the patent in
question, as has been shown in previous research.
A major weakness of this study, however, is that data on patent litigation was obtained
from RPX, which is emphatically not an unbiased data source and the use of which by
researchers is difficult to understand. RPX is a private firm that, among other services,
assists patent infringement defendants with prior art that may assist in challenging the
validity of plaintiffs’ patents, provides patent infringement insurance, acquires patents
itself for licensing and litigation purposes (it is a patent troll), and collects data on patent
infringement litigation as part of the market intelligence service it supplies to clients.
There is no way to assess the methodological soundness of RPX-provided data or the
qualifications of those collecting the data. Thus, a researcher might justifiably decide not
to rely on conclusions drawn from such data.
9. Licensing
A few studies have focused on licensing in relation to patent value. A cogent argument
can be made that the mere existence of a licensing agreement suggests that the patent
or patents forming the subject of the agreement are, on average, more valuable than
unlicensed patents. It may be the case that a distinction is warranted between patent
licenses resulting from litigation and those reached without the filing of litigation. If the
distinction is made, however, a simple dichotomy between licenses achieved completely
apart from litigation and licenses entered into at some point after a patent infringement
lawsuit has been filed is unlikely to reveal much about the value of the subject patents.
Given that some infringement cases are instituted solely for the purpose of pressuring the
defendant to take a license, sometimes for a fee below the defendant’s expected litigation
costs, the stage of litigation at which licensing occurs will be an undeniable factor in any
use of licensing as a value indicator.
A study by Ruckman and McCarthy (2016) examined patent licensing in the bio
pharmaceutical industry, finding that indicators of patent value have less to say about
whether patents get licensed than do the characteristics of the patent owners themselves.
The prestige of the patent owning company and its experience with licensing activity

Patent value 67
were found to be associated with the fact of licensing to a much greater degree than
any particular characteristics of the patents themselves.34 There is some intuitive sense
to these findings, although the many unique traits of the pharmaceutical industry itself
militate against generalizing the findings without replication of this kind of study in other
industries and technologies.
Using a large proprietary data set of licensing transactions by so-called NPEs, Abrams,
Akcigit, and Popadak (2014) found an inverted U-shaped relationship between per-patent
licensing revenues and forward citations (citations received), with fewer citations at the
high end of the value metric than in the middle. Although the confidential nature of the
data’s source and collection methodology hampers an objective assessment of the study’s
reliability, there is no reason to believe that these authors would use a seriously flawed
data set. Licensing revenue clearly represents a direct measure of value for the kinds of
owners and the kinds of patents in this data set. It appears to be a fair guess that the patent
owners represented in these data are firms that own patents solely or at least primarily
for the purpose of licensing them or, failing that, suing on them. Without knowing more
about the data set, including which firms and which patents are included, as well as
how and why they were selected, one cannot be sure about whether the patent owners
represented in this data are all or mostly those that most observers would characterize as
“trolls,” or “patent assertion entities.” These concerns aside, the study clearly complicates
what previously had been the almost universal view of scholars that forward citations
are an indicator of patent value. At a minimum, the study reveals that the relationship is
not linearly simple (and not even monotonic, as the authors observe). Finally, one must
remember that, among those patents that possess any economic value, that value can take
a variety of forms other than generation of licensing revenue. The Abrams, Akcigit, and
Popadak study does, however, add meaningfully to the patent value literature.
10. Transfers of ownership

It also stands to reason that patents which have attracted post-issuance buyers could be
more valuable that those the ownership of which has remained the same since issuance.
Referring back to the litigation-value link, Chien (2011) found that patents ending up
in infringement litigation experienced significantly more ownership transfers through
assignment after the initial patent grant than those that had not been litigated. And
Allison, Lemley, and Walker (2009) revealed that patents which had been litigated mul-
tiple times during an eight-year time period were more likely to have been the subject of
post-issuance, pre-litigation assignments than patents litigated only once during the same
timeframe. On the other hand, Moore (2007) found that pre-litigation ownership transfers
had no significant effect on patent owner win rates in litigation.
11. Litigation outcomes

A final possible variable that may signal some information about patent quality or value is
the outcome of litigation in which infringement of a patent has been averred. Associating
34
It bears observing that there may be some reverse causation at work because the mere fact of
a license could signal to others that the subject patents have value regardless of whether they would
in the absence of the patent owning firm’s characteristics.

any type of variable with litigation outcomes is, however, exceedingly difficult, as seen
in Lanjouw and Schankerman (2003). Working from a data set of U.S. patent litigation
during 1978–99, they found that neither technology areas (improperly identified by patent
classifications), citations to prior patents (divided by number of total claims), citations
received (“forward citations,” also divided by the number of claims for some reason),
number of patents held by each litigating owner, nor the status of a patent owner as
an “unlisted” (primarily individuals and small companies) or “listed” (primarily large
companies) entity had any significant effect on settlement rate or on patent owner win
rate at trial.35 Regardless of whether such findings reflect reality, one cannot depend on
this study as authority for these propositions.
The authors acknowledged that their data on litigation, which was acquired from
commercial provider Derwent and the Federal Judicial Center, was incomplete because of
underreporting by federal district judges to the USPTO. Federal judges began reporting
patent cases to the USPTO in greater numbers after the end of this study, although judges’
omissions of some cases doubtless continue. Regardless, the authors report only trial
outcomes or settlement. Judges probably are more likely to report trial outcomes than the
many other possible results in litigation, so judicial underreporting may not be a serious
problem for this study. Using only trial decisions is a major problem, however.
Outcomes in litigation are complex, and even more so in patent infringement litigation
where multiple patents are often asserted in the same lawsuit, the same patent can be
litigated in more than one case, and judicial decisions are typically at the level of claims
within the patent rather than at the level of the entire patent. The authors did not appear
to be cognizant of the critical role that summary judgment plays in patent litigation,
especially on infringement where summary judgments of noninfringement are the most
common of the many possible outcomes. Also, many determinative, or at least crucial,
decisions are made on pre- and post-verdict motions for judgment as a matter of law, and
on appeal. Moreover, many nuances are attendant to each of these substantive decisions.
Most patent lawsuits are resolved on summary judgment and do not go to trial, and
patent owners and accused infringers face very different odds of successful outcomes in
summary judgment compared with trial. Patent owners lose a large percentage of cases
on summary judgment of noninfringement, but in the much smaller percentage of cases
that get to trial, patent owners win on both infringement and validity over 50 percent of
the time (Allison, Lemley, and Schwartz, 2014). There also is no indication in the paper
that cases such as those alleging false marking, design, or plant patents were excluded, as
should be done when a researcher is only interested in utility patents.
Also, the authors excluded from their data cases in which “the suit was dismissed
without a request from one of the parties,” but it is not at all apparent what this means.
Trial judges rarely dismiss cases or take any other action sua sponte, and no further
explanation was given.
Finally, one cannot depend on case outcomes reported by a busy judge who is com-
pletely uninterested in possible future uses in social science research of the data points
he or she creates. Outcomes reported by a commercial vendor likewise cannot be taken
35
Other researchers in economics management have been known to divide prior art references
and citations received by the total number of claims, which is a rather mystifying exercise.

Patent value 69
at face value because of the immense time needed to get the outcomes right. As noted,
outcomes in patent cases are complex, rarely being a simple win or loss for either party,
and accurately recording them is difficult even for one with extensive legal training and
intimate familiarity with patent litigation. The authors of the study acknowledged that
many patent infringement suits assert multiple patents, and for the purposed of collecting
data from the patents themselves they stated that they chose only the “main patent.”
There is no such thing in patent litigation when multiple patents are alleged to have been
infringed. Of the several patents commonly asserted in the complaint, some of them
frequently drop out of the case very early for unstated reasons. Some of them will drop
out during claim construction or immediately after the court’s claim construction order,
and some will remain. Summary judgment motions on validity and infringement by
both parties normally follow, and each of the grounds asserted by the accused infringer
for invalidity often will be the subject of separate decisions on summary judgment. And
patent owners do not simply sue for infringement “of the patent,” but instead sue for
infringement of particularly identified claims within a particular patent. An early-stage
decision on any of these can be a meaningful win or loss for either party.
With respect to the recording of settlements, in many instances cases are settled by
the parties without there being a docket entry so stating. A docket entry revealing a
consent judgment surely means a settlement. Voluntary dismissals are often the result of
settlements, but not necessarily so, and more research into the court documents may be
needed to determine with greater confidence whether the case was settled (Allison, Tiller,
and Zyontz).36 Stipulated judgments are often representative of settlements, but may be
the result of an earlier decision by one party to drop the suit if a subsequent decision
by the court causes its odds of winning to be dramatically reduced, as with a stipulation
by the patent owner that a decision of noninfringement on a particular patent claim
should later be rendered if the court adopts a specific interpretation of claim language.
Docket sheets are complicated and often very lengthy, sometimes with well over 500
entries to be deciphered, and court documents themselves often must be consulted to
make a relatively reliable outcome coding decision.
In Allison, Lemley, and Schwartz (2014), the authors who coded outcomes were law
professors with extensive patent litigation experience. They coded many specific litigation
outcomes in a data set of 945 cases filed in federal district courts in 2008 and 2009 that
were the subject of a decision on the merits by the end of 2013.37 The data was acquired
from Lex Machina and from manual coding of the patents that were the subject of the
litigation.38 From the cases provided by Lex Machina, the authors excluded lawsuits that
36
Even then, to be accurate one may have to refer to, or even create a separate category for,
“probable settlements.” Allison, Tiller, and Zyontz (2012) used categories for “clear” and “likely”
settlements, and tested first with only clear settlements and then with the combined categories.
37
Lemley and Schwartz were the outcome coders. A test of intercoder reliability revealed over
95 percent agreement between the coders.
38
Recently, the underlying documents, including motions and opinions, from district court
litigation became more readily available. Electronic filing requirements mean that the online filing
tool, Public Access to Court Electronic Records (PACER), has a nearly complete collection of
litigation documents from patent cases. Some scholars have taken advantage of PACER data to
analyze district court decisions. But the raw data provided by the Administrative Office of the

did not include a complaint either for infringement of a utility patent or for a declara-
tory judgment of noninfringement or invalidity of a utility patent. Thus, they excluded
inventorship and licensing disputes, patent law malpractice actions, and allegations of
design or plant patent infringement.
Any empirical study of patent litigation that looks at specific outcomes faces an
inevitable problem. To obtain reliable data on outcomes requires very careful manual
coding from docket sheets and the reading of complaints, motions, and judicial rulings.
The inherently time-consuming nature of the work limits the size of the data set when
the total number of cases is divided into specific decisions on various infringement and
validity issues on summary judgment, trial, and appeal. In this study, the numbers of
observations were sufficient for statistical analysis of a number of measured outcomes,
but not for others. The way in which the study was performed gave the authors confidence
that a high degree of reliability was achieved. However, results from any empirical study of
patent litigation, or indeed any form of litigation, will always be subject to the limitations
of several types of selection effects, which these authors discuss at length.
The most important findings of this study relating to patent quality and value were as
follows. (1) The number of patents asserted by the owner in the same lawsuit significantly
increased the patent owner’s odds of achieving a “definitive win,” or a win on all valid-
ity and infringement issues, on each patent (not just a better chance of winning on at
least one patent in the case). This finding may lend support to the idea that patents can
sometimes have more value as part of a portfolio, and that cumulative innovation matters.
On the other hand, it may simply be the case that the assertion of multiple patents in the
same lawsuit imposes a greater defense burden on the accused infringer. (2) Neither the
number of adjusted forward citations nor the number of prior art references (patent and
nonpatent) has any significant correlation with overall win rates, validity, or infringement
outcomes. Citations seem to tell us nothing about whether patents are valid or whether
they are infringed. The latter set of findings is both surprising and sobering, given the large
amount of time and effort that have been devoted to the use of these variables as patent
value indicators. Such negative findings are tempered, however, by the fact that litigation
outcomes are clearly affected by a number of undiscoverable or unmeasurable variables, as
shown by the low pseudo R2 for each logistic regression model in the study (each outcome
was a binary dependent variable and various patent and litigation characteristics were
independent variables). As the authors observe, who the attorneys, jurors, and judges are
matters, and other unidentified explanatory variables are doubtless at work as well.
United States Courts is notoriously error-prone, and it does a poor job of classifying outcomes.
Therefore, the authors of the study used Lex Machina.
Lex Machina provides convenient access to cleaned and verified PACER data for district
court patent litigation, which permitted us to evaluate all patent lawsuits. Lex Machina data offers
three primary benefits. First, it includes all lawsuits, even those without a decision available on
Westlaw or Lexis, so as not to overcount appellate decisions. Second, Lex Machina has cleaned
and evaluated the PACER data, eliminating many of the errors in the raw data. Finally, Lex
Machina has indexed the cases to identify all summary judgment rulings, trial events, and appeals.
LEX MACHINA, www.lexmachina.com. This being said, one nevertheless cannot rely on Lex
Machina’s reported outcomes, so the authors manually studied all docket sheets and relevant court
documents, which were available on Lex Machina. The authors of the study did not depend on the
outcomes reported by Lex Machina, however.

Patent value 71
Somewhat in alignment with the previous study, another in 2012 by a law professor
and an experienced in-house patent law practitioner found that patent validity deci-
sions were not significantly associated with most of the patent demographics that had
previously been found to predict the fact of litigation (Mann and Underweiser, 2012).
Although the study reveals an admirable degree of ingenuity and effort—the authors even
accomplishing a detailed investigation of the prosecution histories for a sizable sample of
the patents that were the subject of litigation in their data set—it is attended by several
flaws either that are necessarily encountered in studying litigation empirically or that were
avoidable. The most avoidable flaw is that the study examined only validity decisions on
appeal by the U.S. Court of Appeals for the Federal Circuit. The majority of the work in
patent litigation takes place at the district court level, and many important district court
decisions are not appealed. Moreover, unappealed district court decisions are subject to
different selection effects than those decided on appeal, and it is even possible that district
courts and the main patent appeals court may have somewhat different decision-making
tendencies. Allison, Lemley, and Schwartz (2014) found that district courts as a group
were significantly more likely to uphold the validity of patents than the Federal Circuit,
although the difference in selection effects at the trial and appeal levels means that one
should view this result with caution.
A second shortcoming that could have been avoided is that judicial rulings of infringe-
ment were not included, and for several reasons such decisions can be as important as
validity determinations. Many patent lawsuits are killed early in the process—for instance,
by the grant of summary judgment of noninfringement in favor of the defendant—
without any ruling on validity (Allison, Lemley, and Schwartz, 2014). These criticisms
may simply reveal the inevitable tradeoffs of any research effort such as this. The authors
delved into such great detail that adding infringement decisions may have made the effort
infeasible. The same might be said of including unappealed district court decisions, which
are substantially more numerous than appeals court decisions, but leaving out trial court
decisions has a serious negative impact on the value of the study. Any study of litigation
outcomes without infringement decisions and without unappealed district court decisions
is simply incomplete, however, no matter how difficult it would be to include them.
The authors essentially found that the fact of litigation and the outcomes in litigation
appear to be two very different phenomena in a value assessment. The authors found,
for example, that the number of total claims has no statistical association with judicial
rulings of patent validity (Mann and Underweiser, 2012).39 Even more surprising is the
result that the number of countries in which patent protection was obtained on the same
invention—the so-called “patent family”—was significantly but negatively associated with
patent validity in a comparison of patents ruled to be valid and invalid. The authors also
discovered a significant inverse relationship between the number of related patents issued
from ancestor applications in the same chain of continuations (a more descriptive use
of the term “patent family”), which others had related significantly to the fact of litiga-
tion (Allison, Lemley, Moore, and Trunkey, 2004). Notably, especially given that patent
39
Mann and Underweiser (2012) did not measure independent claims by themselves, which
are the cause of an association between claims and litigation. Singling out independent claims
requires a small program that will count total claims.

classifications do not identify or define technology categories, the authors discovered

a significant positive relationship between the number of patent classes per patent and
findings of validity, each additional class increasing the likelihood of validity by 9 percent.
Thus, the number of USPTO classifications or IPCs to which patents are assigned does
seem to signify something. It may be a proxy for some other kind of complexity that itself
is associated with patent validity. The authors themselves even admit that the assignment
of patents to multiple classifications could simply reflect classifications that are poorly
organized. And one cannot escape the fact that the USPTO classification system is both
convoluted and error-prone.
The authors also employed textual analysis software to construct a metric for denoting
the degree of alignment between a patent’s written description and its claims, which if
successful would show an association with a patent’s compliance with the enablement
and written description requirements of 35 U.S.C. § 112 (Allison and Ouellette, 2016).40
The measure, after ignoring articles and other common words, counted the number of
unique words in the written description that also appeared in the claims. The frequency
of these words was not measured, as this would cause longer written descriptions and
sets of claims to bias the results. The degree of alignment was treated as signifying likely
compliance with patent law’s disclosure requirements. Multiple regression revealed a posi-
tive correlation between the “distance” between the written description and the claims, as
measured by a smaller number of unique words in the written description also being found
in the claims, but only at a 90 percent confidence level (p<0.10).41 Also, it appears that the
authors employed the metric only with respect to overall validity or invalidity decisions
and not specifically to decisions on whether patents complied with the enablement and
written description mandates of Patent Act § 112. Although there are some problems with
the authors’ attempt to associate word conformity between the patent’s description and its
claims, further research is warranted on the use of measures such as this to assess likely
compliance with patent disclosure rules.42
Some final comments on the use of litigation outcomes as quality or value indicators
are warranted. First, one cannot ignore the inevitable limits on the empirical study of
specific litigation outcomes noted previously. Second, although some will view the ability
of a patent owner to successfully enforce a patent against an alleged infringer to be the
ultimate test of private value, patents have many other uses and litigation of them is an
extremely infrequent event, with only an estimated 1–2 percent of patents asserted in
court (Allison, Tiller, and Zyontz, 2012; Lemley, 2000). Can such rare events effectively
40
To be valid, a patent’s specification (written description and drawings) must both (1) enable
a person with ordinary skill in the art to make the claimed invention and put it into practice, and
(2) demonstrate that the inventor actually invented what is found in a particular claim.
41
The coefficient, which in logistic regression is a likelihood ratio, was very small (0.04), sug-
gesting a tiny practical effect.
42
In addition to apparently seeking to compare description-claims word conformity only
with validity generally and not with the specific patent disclosure requirements, they appeared
to recognize only enablement and claim definiteness as patentability requirements, and not the
so-called “separate written description requirement” that is fulfilled only when a patent’s specifica-
tion clearly reveals that the inventor(s) invented specifically what is found in the claims. Claim
definiteness is not a disclosure requirement, and should not be affected by any measure of word
commonality between the description and the claims.

Patent value 73
measure either the quality or the worth of patents? The very same question naturally
applies to all of the previous work on the link between the mere fact of litigation and
private patent value. There surely is something to it, but we can’t say how much. Then, too,
there are the many issues with selection effects and unmeasurable explanatory variables
that plague studies of both the fact and the outcome of litigation.
VII. DIRECTLY MEASURING PRIVATE VALUE
In the most direct attempt to measure patent value so far attempted, a 2013 study
used auction data in an effort to verify the reliability of certain previously suggested
patent value indicators. The authors analyzed the impact of patent characteristics on
the economic value of patents put on sale at auction, considering different typologies
of sellers and buyers (Odasso, Scellato, and Ughetto, 2013). Using a data set covering
all patent auctions held until the end of 2008 by intellectual property merchant bank
Ocean Tomo,43 the authors employed two measures of patent value drawn from the
auction process: the offer price (ex ante value set by the seller) and the closing price
(the market value determined by the buyers) of patent lots. They found a positive cor-
relation between forward citations and the economic value of patents (in terms of both
the lot offer price and the lot closing price), while the number of prior art references
had a significant positive impact only on the lot offer price. The number of countries in
which patent protection was achieved was positively associated with both the lot offer
price and the lot closing price. Some patent characteristics positively affected price only
in the case of buyers which were NPEs, including the number of claims and relatively
young patent age.
A study such as this, which measures actual patent transactions, holds promise as an
excellent means to measure private economic value in the case of those patents whose
worth resides in their capacity to generate licensing revenue, and if possible should be
repeated using other data sets, other contexts, and additional variables.
REFERENCES
Abrams, David S., Ufuk Akcigit, and Jillian Popadak. 2014. “Patent Value and Citations: Creative Destruction
or Strategic Disruption?” PIER Working Paper No. 13-065, University of Pennsylvania, Institute for Law &
Economics Research Paper No. 13-23. Available at https://papers.ssrn.com/sol3/papers.cfm?abstract_id=2351
809&rec=1&srcabs=2371600&alg=1&pos=10.
Alcácer, Juan, Michelle Gittelman, and Bhaven Sampat. 2008. “Applicant and Examiner Citations in US
Patents: An Overview and Analysis”, Harvard Business School Working Paper 09-016.
Allison, John R. and Lisa Larrimore Ouellette. 2016. “How Courts Adjudicate Patent Definiteness and
Disclosure”, 65 Duke Law Journal 610–95.
43
Ocean Tomo, which, according to the company, introduced the world’s first public auction
of patents and copyrights, is an intellectual property merchant bank that provides financial
products and services, including expert testimony, valuation, research, ratings, investments, risk
management, and transactions. See www.oceantomo.com/; https://en.wikipedia.org/wiki/Ocean_
Tomo#cite_ref-3.

Allison, John R. and Mark A. Lemley. 1998. “Empirical Evidence on the Validity of Litigated Patents”, 26
Allison, John R. and Ronald J. Mann. 2007. “The Disputed Quality of Software Patents”, 85 Washington
Allison, John R. and Starling D. Hunter. 2006. “On the Feasibility of Improving Patent Quality One Technology
at a Time: The Case of Business Methods”, 21 Berkeley Technology Law Journal 729–94.
Allison, John R. and Thomas W. Sager. 2007. “Valuable Patents Redux: On the Enduring Merits of Using Patent
Characteristics to Identify Valuable Patents”, 85 Texas Law Review 1769–98.
Allison, John R., Emerson H. Tiller, and Samantha Zyontz. 2012. “Patent Litigation and the Internet”, Stanford
Allison, John R., Mark A. Lemley, and David L. Schwartz. 2014. “Understanding the Realities of Modern
Patent Litigation”, 92 Texas Law Review 1769–801.
Allison, John R., Mark A. Lemley, and Joshua W. Walker. 2009. “Extreme Value or Trolls on Top? The
Characteristics of the Most-Litigated Patents”, 158 University of Pennsylvania Law Review 1–37.
Allison, John R., Mark A. Lemley, and Joshua W. Walker. 2010. “Patent Quality and Settlement Among Repeat
Patent Litigants”, 99 Georgetown Law Journal 677–712.
Allison, John R., Mark A. Lemley, Kimberly A. Moore, and Derek Trunkey. 2004. “Valuable Patents”, 92
Barry, Justus and Henry Delcamp. 2012. “The Private and Social Value of Patents in Discrete and Cumulative
Innovation”, 90 Scientometrics 581–606.
Chien, Colleen V. 2011. “Predicting Patent Litigation”, 90 Texas Law Review 283–329.
Cohen, Wesley M., Richard R. Nelson, and John P. Walsh. 2002. “Protecting Their Intellectual Assets:
Appropriability Conditions and Why U.S. Manufacturing Firms Patent (or Not)”, National Bureau of Econ.
Research, Working Paper No. 7552.
Ernst, Holger, Sebastian Legler, and Ulrich Lichtenthaler. 2010. “Determinants of Patent Value: Insights from
a Simulation Analysis”, 77 Technological Forecasting and Social Change 1–19.
Goldenberg, David H. and Jonathan D. Linton. 2012. “The Patent Paradox—New Insights Through Decision
Support Using Compound Options”, 79 Technological Forecasting and Social Change 180–85.
Graham, Stewart J.H. and David C. Mowery. 2003. “Intellectual Property Protection in the U.S. Software
Industry”, in Wesley M. Cohen and Stephen A. Merrill eds., Patents in the Knowledge-Based Economy.
Washington, D.C.: National Academies Press.
Grube, Christian. 2009. Measuring the Immeasurable: Valuing Patent Protection of Knowledge-Based
Competitive Advantages. Wiesbaden, Germany: Gabler (division of Springer).
Guellec, Dominique and Bruno Van Pottelsberghe de la Potterie. 2000. “Applications, Grants, and the Value of
Patent” (sic), 69 Economics Letters 109–14.
Hall, Bronwyn H., Adam B. Jaffe, and Manuel Trajtenberg. 2002. “The NBER Patent-Citations Data File:
Lessons, Insights, and Methodological Tools”, in Adam B. Jaffe and Manuel Trajtenberg eds., Patents,
Citations, and Innovations: A Window on the Knowledge Economy. Cambridge, MA: MIT Press.
Hall, Bronwyn H., Adam B. Jaffe, and Manuel Trajtenberg. 2005. “Market Value and Patent Citations”, 36
Harhoff, Dietmar, Frederic M. Scherer, and Katrin Vopel. 2003. “Citations, Family Size, Opposition and the
Value of Patent Rights”, 32 Research Policy 1343–63.
Kahneman, Daniel. 2011. Thinking, Fast and Slow. New York: Farrar, Straus and Giroux.
Lanjouw, Jean O. and Mark Schankerman. 1997. “Stylized Facts of Patent Litigation: Value, Scope and
Ownership”, National Bureau of Economic Research, Working Paper No. 6297.
Lanjouw, Jean O. and Mark Schankerman. 1999. “The Quality of Ideas: Measuring Innovation With Multiple
Indicators”, National Bureau of Economic Research, Working Paper No. 7345.
Lanjouw, Jean O. and Mark Schankerman. 2003. “Enforcement of Patent Rights in the United States”, in
Wesley M. Cohen and Stephen A. Merrill, eds., Patents in the Knowledge-Based Economy. Washington, D.C.:
National Academies Press.
Lanjouw, Jean O., Ariel Pakes, and Jonathan Putnam. 1998. “How to Count Patents and Value Intellectual
Property: The Uses of Patent Renewal and Application Data”, Journal of Industrial Economics 405–32.
Lemley, Mark A. 2000. “Rational Ignorance at the Patent Office”, 95 Northwestern University Law Review
1495–529.
Lerner, Josh. 1994. “The Importance of Patent Scope: An Empirical Analysis”, 25 RAND Journal of Economics
319–33.
Levin, Richard C., Alvin K. Klevorick, Richard R. Nelson, and Sidney G. Winter. 1987. “Appropriating the
Returns From Industrial R&D”, 3 Brookings Papers on Economic Activity 783–831.
Malackowski, James E. and Jonathan A. Barney. 2008. “What Is Patent Quality? A Merchant Banc’s
Perspective”, June Les Nouvelles: Journal of Licensing Executives Society International 123–34.

Patent value 75
Mann, Ronald J. and Marian Underweiser. 2012. “Patent Quality and Validity”, 9 Journal of Empirical Legal
Studies 1–32.
Marco, Alan C. and Richard D. Miller. 2017. “Patent Examination Quality and Litigation: Is There a
Link?” USPTO Economic Working Paper No. 2017-09. Available at https://papers.ssrn.com/sol3/papers.
Moore, Kimberly A. 2005. “Worthless Patents”, 20 Berkeley Technology Law Journal 1521–55.
Moore, Kimberly A. 2007. “Populism and Patents”, 82 NYU Law Review 69–111.
Moser, Petra, Joerg Ohmstedt, and Paul W. Rhode. 2015. “Patent Citations and the Size of the Inventive
Step—Evidence from Hybrid Corn”, National Bureau of Economic Research Working Paper No. 21443.
NTP, Inc. v. Research in Motion, Ltd., 418 F.3d 1282 (Fed. Cir. 2005).
Odasso, Christina, Giuseppe Scellato, and Elisa Ughetto. 2015. “Selling Patents at Auction—An Empirical
Analysis of Patent Value”, 24 Indus. and Corporate Change 417–38.
Okada, Yoshimi, Yusuke Naito, and Sadao Nagaoka. 2016. “Claim Length as a Value Predictor of a Patent”,
Institution of Innovation Research, Hitotsubashi University Working Paper No. 16-04. Available at www.
pubs.iir.hit-u.ac.jp/admin/ja/pdfs/file/1924.
Oracle Corp. 2016. “Topic Submission for USPTO Quality Case Studies (Feb.12, 2016)”. Available at www.
uspto.gov/patent/laws-and-regulations/comments-public/topics-submitted-quality-case-studies.
Parchomovsky, Gideon and R. Polk Wagner. 2005. “Patent Portfolios”, 154 University of Pennsylvania Law
Review 1–77.
Putnam, Jonathan. 1997. “The Value of International Patent Rights”, Working Paper. Available at www.competi
tiondynamics.com/wp-content/uploads/International-Patent-Rights.pdf.
Reitzig, Markus. 2003. “What Determines Patent Value? Insights From the Semiconductor Industry”, 32
Ruckman, Karen Elizabeth and Ian P. McCarthy. 2016. “Why Do Some Patents Get Licensed While Others Do
Not?” Indus. and Corporate Change DOI/10.1093/icc/dtw046 (forthcoming). Available at SSRN: https://ssrn.
com/abstract=2952557.
Schankerman, Mark. 1998. “How Valuable is Patent Protection? Estimates by Technology Field”, 29 RAND
Shapiro, Carl. 2001. “Navigating the Patent Thicket: Cross Licenses, Patent Pools, and Standard-Setting”, 1
Innovation Policy and the Economy 119–50.
Torrance, Andrew W. and Jevin D. West. 2016. “Are Litigated Patents More Important?”, Working Paper.
Available at www.ipsdm2016.com/.
Van Pottelsberghe de la Potterie, Bruno. 2002. “The Value of Patents and Patenting Strategies: Countries and
Technology Areas Patterns”, 11 Economic Innovation and New Technology 133–48.

Section B
Patent Institutions and Litigation

4. Empirical scholarship on the prosecution process
at the USPTO
Michael D. Frakes* and Melissa F. Wasserman**
Contents
I. Introduction
II. Evidence on the Effectiveness of the USPTO’s Examination Practices: Overview
III. Understanding the Examination Process
IV. Heterogeneity in Patent Office Outcomes
V. Evidence the USPTO is Deviating from its Mission
VI. Concluding Remarks
References
I. INTRODUCTION
Recent years have seen a surge in interest in intellectual property rights, in particular pat-
ents. This heightened level of interest has culminated in a consensus that the patent system
is functioning poorly. There is widespread agreement that the United States Patent and
Trademark Office (USPTO) allows too many invalid patents to issue that unnecessarily
drain consumer welfare, and that the USPTO’s backlog of patent applications prevents it
from providing timely review of applications. Additionally, there is growing concern that
patentability decisions at the USPTO are highly inconsistent across examiners—that is,
the USPTO’s decision to grant a patent application is driven not only by the merits of the
invention, but also by happenstance of the examiner to which the application is randomly
assigned.
While there is a substantial literature in economics bearing on the patent system, the
administrative process by which patent rights are initially established has received scant
attention. In the past decade a growing but nascent literature has emerged that has begun
to shed empirical light on the patent examination process. This chapter will provide a brief
overview of this literature, focusing only on studies that carry significant empirical com-
ponents, in addition to focusing only on studies of the U.S. patent system (for purposes
of tractability).1 Finally, we should note that the need for sound empirical guidance on
the administrative process of obtaining a patent is substantial. Concerns over the patent
examination process have in part prompted Congress to enact the first major patent
** Professor of Law and Economics at Duke University.

** Charles Tilford McCormick Professor of Law at the University of Texas at Austin.
*1 This summary of the literature is limited to those articles that were published (i.e., as work-
ing papers) as of January 2016.
77

reform act in over 60 years, spurred the Supreme Court to take a more active interest in
substantive patent law, and drove the USPTO to hold its first Patent Quality Summit.
Yet without sufficient empirical evidence as to which features of the USPTO distort its
decision making, policymakers have been left trying to reform the patent system without
understanding the root cause of the system’s pathologies.
II. EVIDENCE ON THE EFFECTIVENESS OF THE USPTO’S

EXAMINATION PRACTICES: OVERVIEW
In an ideal world, patent examiners would apply the patentability standards correctly on
average—that is, generally rejecting those applications that fail to meet the patentability
standards—and in a consistent manner. A number of studies have attempted to assess
the USPTO’s success in these regards and have demonstrated shortcomings across both
dimensions. Such shortcomings implicate both efficiency and equity concerns. Invalid
patents unnecessarily drain consumer welfare, stunt productive research and unreason-
ably extract rents from innovators. In addition to invoking concerns over an inequitable
process, inconsistent patentability determinations erode confidence in the USPTO by
creating the appearance of unfairness and arbitrariness, which could in turn diminish
incentives for innovation. In the next three sections we summarize empirical studies of the
patent examination process, dividing these studies into three broad categories: (1) those
examining the application process more generally; (2) those demonstrating heterogene-
ity in application outcomes across patent examiners; and (3) those demonstrating bias
toward granting by the USPTO.
III. UNDERSTANDING THE EXAMINATION PROCESS
A small but growing number of studies have sought to assess the basic nuts and bolts
of the patent application review process itself. Though these studies may fall short of
identifying causal pathways between certain features of the USPTO and distortions in
the review process, their findings may nonetheless serve as highly valuable inputs into
the broader policy debates surrounding the USPTO. Before turning to these studies, this
section briefly provides a background of the patent examination process.
Each year between 300 000 and 500 000 patent applications are filed at the USPTO.
Every patent application contains a specification, which describes the invention, and a set
of claims that defines the metes and bounds of the legal rights the applicant is seeking. In
addition, to satisfy applicants’ duty of candor under U.S. law,2 patent applications typi-
cally disclose to the USPTO “prior art”—that is, previous patents, patent applications, or
other publications that are material to the patentability of the relevant invention. Before
an application enters examination, it is routed to an Art Unit, a group of eight to 15
patent examiners who review applications in the same technological field. Upon arrival,
2
This does not include a duty to search for material information, but only a duty to disclose
material information of which an applicant is aware.

Empirical scholarship on the prosecution process at the USPTO 79
the Supervisory Patent Examiner (SPE) of that Art Unit randomly assigns the application
to a specific examiner.3 The assigned examiner then conducts her own prior art search
to supplement that disclosed by the patent applicant. After completing her search, she
assesses the patentability of the claims of the invention based on the criteria outlined in
the Patent Act and composes a “first office action” letter to the applicant that accepts
or rejects the claims. Although some applications are allowed in their entirety upon first
examination, more frequently, some or all of the claims fail to meet at least one of the
patentability requirements, as the examiner will detail in the first office action letter. The
applicant then responds by either amending the claims or disputing the rejection. After
the response, the patent examiner may issue a final rejection or allow the patent to issue.
Upon receiving a final rejection, an aggrieved patent applicant can either abandon the
application altogether, restart the examination process by filing a continuation applica-
tion, or appeal the denied application to the Patent Trial and Appeal Board.
While it may take several years from filing a patent application to a final patentability
decision from the USPTO, an examiner, on average, spends only 19 hours reviewing
a patent application, including reading the patent application, searching for prior art,
comparing the prior art with the patent application, writing a rejection, responding to
the patent applicant’s arguments, and often conducting an interview with the applicant’s
attorney. As a result, there is considerable discussion that patent examiners are not
given sufficient time to provide high-quality examination. Such deficiencies are poten-
tially problematic in light of the presumption of validity afforded to incoming patent
applications—that is, if a patent examiner fails to uncover and explicitly set forth reasons
as to why the application fails to meet the patentability requirements within the allotted
time for examination, she is legally expected to grant the patent.
Critical to the process just summarized above is the review of prior art by the examiner in
applying the patentability requirements, mainly the novelty and nonobviousness require-
ments. Several empirical studies have examined the role of prior art in this process. These
studies have exploited a 2001 change in procedure at the USPTO that made it possible to
measure examiner and applicant prior-art citations separately for the first time. Alcacer,
Gittelman, and Sampat (2009) explored the degree to which examiners contribute to the
prior art cited in issued patents. Examining all U.S. patents granted in 2001–03, they found
that examiners played a significant role in identifying prior art, adding 63 percent of all
citations on average and providing all of the citations for 40 percent of patents granted.
Sampat (2010) utilized the same database to examine whether applicants contribute more
prior art for important inventions. Utilizing three patent-specific measures of importance
for a subset of the sample patents—i.e., forward citations, renewal data, and patent family
size—he finds that applicants are more likely to contribute more prior art for important
inventions. Cotropia et al. (2013) built upon this early work by exploring how examiners
evaluate prior art. Utilizing a 1 percent random sample of all utility patents issued in 2007
(1564 patents), they find that examiners overwhelmingly relied upon prior art they found
themselves (rather than applicant provided) in considering whether an invention was new
3
We conducted a series of telephone interviews with former SPEs to confirm these details of
patent examination assignment. Our interviews further substantiated that SPEs do not make any
substantive evaluation of an application before assigning it to a particular examiner.

and nonobvious. To the extent that examiners are not reading and applying all the art
submitted to them, the Cotropia et al. analysis has several implications for patent policy.
First, it casts doubt on the need for the duty of candor in the United States, which, to
recap, requires patent applicants to submit prior art to the USPTO. Second, these find-
ings cast doubt on the premise underlying case law that states that prior art that was not
considered by the USPTO (or where there was at least no record that it was considered)
likely carries more weight than prior art that had previously been considered and rejected
by the USPTO. That is, if examiners are not systematically reading and evaluating
applicant-submitted prior art, there may be little reason to distinguish between prior art
that was and was not put forth before the examiner.
Another fundamental question that arises when assessing the functioning of the review
process is the degree to which patent examiners are capable of sorting meritorious from
non-meritorious applications. Zhen Lei and Brian Wright (2017) confront this question
by looking at a set of inventions that were patented in the United States and that likewise
sought patent protection at the European Patent Office (EPO). Given that the EPO is
an office with essentially similar patentability standards that invests considerably more
resources in application review than the USPTO, Lei and Wright follow an approach
applied by other scholars and use a rejection of an application at the EPO for an inven-
tion that received a patent at the USPTO as an indication for a weak patent—that is, as
an indication for what the USPTO would have done if it were to invest more heavily in
its examination system. With this benchmark in mind, Lei and Wright find that claimed
inventions in which U.S. examiners were able to identify a higher share of potentially
citable prior art—suggesting stronger search scrutiny—were more likely to be rejected by
the EPO. They interpret these findings as suggesting that U.S. examiners will exert greater
effort in an attempt to invalidate an application that they feel is weak. Despite the intuitive
argument that U.S. examiners may nonetheless issue invalid applications—that is, grant
applications on inventions that fail to meet the patentability requirements—due to certain
institutional features of the USPTO, their analysis uses variations across a set of U.S.
applications that nonetheless all issue to make an argument that U.S. examiners have the
capacity to sort meritorious applications (which should be granted) from non-meritorious
applications (which should be rejected).
In a recent paper—which we will discuss in greater detail below—Michael Frakes and
Melissa Wasserman (2017) find evidence that likewise supports the view of a rational
examination process that at least has the potential to target strong applications. Frakes
and Wasserman’s approach, however, generates this implication while arguably compar-
ing the underlying validity of a set of applications that were issued in the U.S. with a set
that were not, unlike Lei and Wright.
As will be explained further below, Frakes and Wasserman find evidence suggesting
that a tightening of time constraints facing examiners may contribute to the issuance
of additional patents on the margin—that is, the issuance of a set of patents that would
have been rejected if examiners were given more time to review. They then find evidence
suggesting that the patents being issued on the margin as a result of these time constraints
may be of weaker underlying validity (likewise assessing validity using EPO outcomes as
a benchmark) than the average issued patent—that is, than the patents that would have
likely issued even with unconstrained examination time. This suggests that U.S. examiners
may target their allowances first at meritorious applications. Should pressures mount that

cause them to grant additional patents on the margin, they will have to dip into the pool
of marginally weaker applications. While this analysis suggests that certain features of the
USPTO may be leading to undesirable outcomes—as discussed further below—it does
nonetheless imply a baseline degree of competency in the ability of examiners to exercise
their sorting responsibilities.
In a related vein, a series of papers have sought to assess the degree to which patent
examiners scrutinize applications and sort meritorious ones from non-meritorious
ones by simply calculating the USPTO’s grant rate—that is, the probability that a given
application will be allowed. This calculation is surprisingly complicated to pull off in light
of several features of the examination process, including the significant length of time it
takes to review applications (over which time, rates of filing may change considerably),
and the fact that a truly original application may ultimately lead to the filing of one or
more continuation applications (creating questions as to what values should be included
in the grant rate numerator and denominator). One oddity of the U.S. patent system is
that an aggrieved patent applicant can always choose to start the examination process
over by filing a continuation application or a request for continued examination. Early
investigations into this question estimated that as much as 97 percent of applications
are allowed by the USPTO, implicating concerns of a patent office that is no more than
a rubber stamp USPTO, potentially burdening society with the harms of over-granting
patents delineated above (Quillen and Webster, 2001). Subsequent studies employed a
number of different assumptions about the treatment of continuations and other matters
in forming the relevant numerator and denominator, while also employing a range of
methodological approaches—for example, some relying on aggregate counts of various
measures and some relying upon application-level data that allow one to follow individual
applications throughout their lifecycle. All told, the grant rates estimated in these studies
vary quite a bit, falling below 50 percent in some studies (Quillen and Webster, 2002;
Lemley and Sampat, 2008; Carley, Hedge, and Marco, 2015).
At the least, these grant-rate-calculation studies paint a picture in which the USPTO
is perhaps not a rubber stamp at all, but does indeed reject a non-trivial percentage of
applications. Given the findings mentioned previously of Lei and Wright and of Frakes
and Wasserman (2017), one might believe that those that are rejected by the USPTO are
indeed lacking in legal validity in the first place. Of course, these latter, largely descriptive
grant-rate studies cannot determine whether the observed rates are too high or too low;
and, if such rates do indeed deviate from the optimum, such studies have not causally
identified any features of the USPTO or of the examiners that may contribute to any such
distortion. In section IV below, we turn to a summary of studies that attempt to fill this
gap, including the aforementioned Frakes and Wasserman study.
It also bears emphasis that the grant rate itself provides an incomplete emphasis on the
degree to which patent examiners scrutinize applications and target rewards on meritori-
ous applications. The job of a patent examiner, after all, does not necessarily collapse to
a binary decision problem of accepting or rejecting an application. They are also tasked
with working with applicants to narrow the breadth of the requested claims. There may
be many initially weak applications that an examiner will ultimately allow. Focusing solely
on the fact of allowance may provide a limited sense of the degree to which examiners
scrutinized this questionable application and may miss any claim narrowing that the
examiners were able to force. One can view any such claim narrowing as a sorting of

meritorious claim breadth from non-meritorious claim breadth, even should the applica-
tion itself issue in some form.
The extent of claim narrowing occurring at the USPTO is an underdeveloped area of
scholarship, though it has been the subject of several studies to date. For instance, Mark
Lemley and Bhaven Sampat (2010) find that the vast majority of patents that issue do so
after undergoing some degree of amendment to the requested claims.4 In a 2004 article,
Doug Lichtman tracks textual changes in the claims between the time of the application
filing and the issuance of the relevant patent in order to quantify such amendments.
Lichtman’s analysis, however, aims to do more than that just empirically confirm that
the USPTO’s sorting function works along this claim-narrowing margin as well; rather,
it primarily aims to demonstrate the degree to which patent examiners vary in how they
perform this task. In this light, Lichtman joins another set of studies that seeks to explore
examiner heterogeneity in practices. We turn now to a survey of this literature.
IV. HETEROGENEITY IN PATENT OFFICE OUTCOMES
The discussion so far has focused on understanding the process by which examiners
make their decisions, as well as whether examiners are capable of sorting applications
based upon quality. Contemporaneous with these studies has also been the development
of another line of literature that has documented substantial examiner heterogeneity
in USPTO outcomes. That is, these studies provide evidence that the decision to grant
a patent application is driven not only by the merits of the invention, but also by hap-
penstance of the examiner to which the application is randomly assigned. The existence
of inter-examiner disparity raises concerns that the examination process is arbitrary, as
well as raising concerns as to whether examiners are systematically “missing the mark” in
making validity determinations.
The seminal work of Cockburn et al. (2003) analyzes 196 examiners who had worked on
182 patents involved in United States Court of Appeals for the Federal Circuit litigation
between 1997 and 2000. Evaluating all patents for which an examiner was associated,
as either primary or secondary examiner, Cockburn et al. demonstrate that a notable
portion of the overall variance in certain patent characteristics, including the citation
received per patent examination and the citations that appear in the patents examined,
can be explained simply by the identity of the examiner associated with the issued patent.
Stated differently, Cockburn et al. find evidence that an applicant’s outcome with the
examination process is largely a function of the patent examiner that her application is
randomly assigned.5
4
Frakes and Wasserman (2015) likewise investigate the claim-narrowing role of the USPTO.
However, rather than quantifying the extent of claim narrowing, they find evidence in support of
their prediction that the USPTO issues patents of excessive breadth (as captured by the number
of claims) in order to discourage the filing of continuation applications—which are costly to the
USPTO—during times in which the USPTO is resource constrained.
5
This empirical observation was not especially surprising to Cockburn et al. (2003). Their
extensive qualitative review of the USPTO examination system revealed a process in which
substantial discretion is given to patent examiners in how they deal with applications and how they

Lichtman (2004) also documents variation in the effect that examiners have on changes
in language of patent claims during the application review process. Starting with a dataset
of 300 000 patent applications, Lichtman, for various methodological reasons, narrowed
his database down to approximately 2200 applications which matriculated into patents.
For these 2200 patent applications Lichtman quantified the extent of textual changes
between the application and the issued patent. His results suggest that differences among
the assigned examiner account for about two-thirds of the variation in stringency of
editing patent claims.
Various studies have built upon the findings of Cockburn et al. and Lichtman by
beginning to explore certain features of the examiners that may contribute to the hetero-
geneity in patent examination process outcomes. Mann (2014) explored the way in which
examiner tenure and experience relate to characteristics of issued patents, such as number
of claims, prior art references cited in issued patent, and days in examination. Analyzing
over 500 000 issued patents associated with approximately 250 different examiners, he
finds that experience and tenure cut in opposite ways. That is, while the average number of
claims and average days of examination increase with examiner experience, they decrease
steadily with examiner tenure.6 Tu (2012) examines over 1.5 million issued patents and
argues that at least two distinct sets of examiners exist: a group of senior examiners who
issue a large number of patents, and a group of junior examiners who issue a small number
of patents. Although it is difficult to know exactly what to make out of Tu’s findings, as
he does not, among other things, correct for differences in the workload expectations of
examiners which increase with tenure at the USPTO, he contends that changes in scrutiny
of examiners’ work product explain his findings.
Frakes and Wasserman (2016) examine whether the year an examiner was hired by the
USPTO affects how she approaches the examination process, motivated by theories of
examiner learning in which initial training conditions shape persistent practice styles. In
contrast to the investigations above, which utilized a sampling frame of issued patents,
Frakes and Wasserman use application-level data, enabling them to shed light on argu-
ably the most important outcome of the examination process: whether the application
is granted. Analyzing over 1.5 million patent applications filed at the USPTO over a
ten-year period, Frakes and Wasserman find that the year a patent examiner was hired has
an enduring effect on her grant rate throughout her career. They also find that observed
differences in the mean grant rates of the various examiner hiring-year cohorts align with
search for prior art. Also, the fact that much of the learning within the USPTO, especially before
2006, comes through a process that is akin to an apprenticeship (as opposed to a purely centralized
training system) makes it more likely that examiners will begin to follow divergent pathways in their
examination styles.
6
One important limitation of Mann’s (2014) analysis, however, is his treatment of time
controls. The nature of the sampling frame that he employs is such that high-experience examiners
are more likely to show up in the examinations conducted later in the sample period, while low-
experience examiners are likely to show up more evenly throughout the sample. This is concerning
for any estimation of the link between examiner experience and examination duration to the extent
that mean duration is increasingly considerably over time. To address such concerns, Mann fits
linear time trends to his estimation. This control may be inadequate to the extent that prosecution
delays increase non-linearly over time. There may be little reason to think that duration increases
linearly over time over a several-decade period.

temporal changes in both the USPTO’s culture regarding the allowance of patents as well
as new hire training programs at the USPTO. Their results suggest that a substantial por-
tion of heterogeneity among examiners’ application outcomes is not simply idiosyncratic
to individual examiners, but instead is driven by the year in which they were hired by the
USPTO (combined with fluctuations over time in initial hiring conditions).
Lemley and Sampat (2012) and Frakes and Wasserman (2017)—articles that will be
discussed more below—examine how the experience of the examiner and the amount of
examination time allocated to the examiner respectively affect patent office outcomes.
Similar to Frakes and Wasserman (2016), Lemley and Sampat (2012) and Frakes and
Wasserman (2017) also use application level data and hence can examine how examiner
characteristics affect their decision to grant a patent. Though not the sole aim of their
projects, these studies further shape our understanding of those features of examiners
that contribute to the baseline heterogeneity in application outcomes across examiners.
Finally, Vishnubhakat and Rai (2015) compare patent office outcomes for Art Unit
1631, which reviews interdisciplinary inventions at the intersection of the biological and
information sciences with those in matched set of applications assigned to the traditional
software, Art Unit 2123. Utilizing a matched sample of approximately 60 applications
in each art unit, they find that applications in Art Unit 1631 experienced significantly
more rejections, particularly notice-related rejections, than the conventional software
applications. The authors suggest potential causal explanations including differences in
the training and educational attainment of examiners in Art Unit 1631, although they
acknowledge that their methodological design, along with limited data, does not enable
them to make causal inferences.
V. EVIDENCE THE USPTO IS DEVIATING FROM ITS MISSION
As noted above, there is general consensus that the USPTO is routinely issuing invalid
patents—that is, patents that are granted even though the invention fails to meet the
patentability requirements. It is indisputable that invalid patents impose a multitude
of costs on society. Low-quality patents can be utilized by non-practicing entities to
opportunistically extract licensing revenue from innovators; they can also stunt follow-on
innovation. More fundamentally, invalid patents can result in supracompetitive pricing
and diminished quantity without providing society with any innovative benefit.
The quality of issued patents has become such an important and visible issue that the
judiciary, Congress, and the USPTO have attempted to increase the quality of issued
patents. Perhaps most notably, in 2011 Congress enacted the first major patent reform bill
in over 60 years (Leahy-Smith America Invents Act, 2011). The USPTO was granted new
adjudicatory authorities and the ability to set its own fees in an effort to improve patent
quality. Despite the fact that major changes to the patent system are driven by concerns
that the USPTO allows too many invalid patents to issue, there is a lack of compelling
empirical evidence that particular features of the system bias the USPTO toward allowing
patents. As a result, policymakers have been making changes to the patent system absent
empirical evidence to help illuminate the actual problems at hand.
Most of the studies discussed in the previous two sections help to illuminate the
examination process as well as determinants of examiner behavior. They do not, however,

link their empirical analysis to particular features of the USPTO that may contribute to
over-granting or apply convincing strategies to causally link such features to observed
granting patterns. Effective policy guidance requires analytical steps of this nature. This
section now turns to discussing studies that document a pro-patent bias at the USPTO.
The important work of Lemley and Sampat (2012) links one feature of the patent
examination process to the outcome of whether the application is granted: the experience
level (in years) of the patent examiner randomly assigned to review the application.7
Looking at a sample of 10 000 applications filed in January 2001, they find a positive and
monotonically increasing relationship between grant rates and the temporal experience of
the examiner. They also find that as examiners gain more experience, they cite less prior
art—a possible indicator of lightened search scrutiny. Given that it is generally believed
that the EPO does a better job at screening patent applications, the authors’ finding that
the likelihood of the allowed US patent being rejecting by the EPO (for applications filed
in both the USPTO and the EPO) increases with the experience of the associated U.S.
examiner suggests that the additional patents allowed by experienced examiners may
be excessive in nature. Ultimately, however, Lemley and Sampat acknowledge that they
cannot determine whether the senior examiners are granting excessively or whether the
juniors are granting insufficiently.
While a big step forward, a limitation with Lemley and Sampat’s (2012) analysis is its
inability to attribute the positive correlation that it estimates to four possible explanations:
(1) that temporal experience genuinely causes patent examiners to grant at higher rates;
(2) a tenure effect: that patent examiners who happen to stay at the USPTO the longest,
and thus attain higher experience levels, differ in their granting proclivities from those who
stay the shortest amount of time; (3) that higher granting is due to the fact that examin-
ers at higher pay grades—something correlated with experience in years—are given less
time to review applications;8 and (4) a hiring-year cohort effect—the year examiners are
hired (and the conditions of the USPTO at that time) may have an enduring effect on an
examiner’s grant rate. The cross-sectional nature of Lemley and Sampat’s data leaves them
unable to fully distinguish between the first, second and fourth explanations. For instance,
by comparing senior examiners and junior examiners at a given point in time, they cannot
isolate a true experience effect as distinct from a selective retention effect. Moreover,
Lemley and Sampat cannot distinguish between the first and third story because they
do not have data sufficient to separate the experience in years of the examiner from the
pay-grade level of the examiner, which is determinative of the amount of time extended
to examiners.
7
Importantly, to the extent that patent application assignment within an Art Unit is not
truly random—for instance, interview results of formers SPEs suggest that at least for some Art
Units, a subset of examiners in an Art Unit may specialize in a subfield and hence applications
in this subfield are disproportionately assigned to this subset of examiners—does not necessarily
cast doubt on studies that rely upon randomization for their methodological design. For almost
all of the studies discussed in this subsection, the methodology utilized requires only that patent
applications be randomly assigned based upon patentworthiness. All interviews with former and
current patent examiners as well as high-ranking officials at the USPTO, highlight the impossibility
of SPEs to assign patent applications based upon the patentworthiness of an application.
8
We elaborate on this time allocations story below.

Frakes and Wasserman (2017) build upon Lemley and Sampat (2012) and develop
a methodological approach to separate these four stories: experience effects, tenure
effects, time-allocation effects, and cohort effects. Given that their goal is to help
illuminate the over-granting tendencies of the USPTO, Frakes and Wasserman target
their inquiry on the explanation that theorizes an excessive granting mechanism: the
time allocated to patent examiners. The operating theory here is quite simple. Patents
are assumed valid when filed. Examiners must form bases of rejection in order to reject.
As such, if examiners face binding time constraints, they may reject at insufficient rates.
Frakes and Wasserman test this by exploiting the fact that examination time allotments
shrink as examiners are promoted within the USPTO. By the time an examiner reaches
GS-14, she has nearly half of the time to review applications as do comparable GS-7
examiners. Using rich data on nearly 1.5 million individual applications over a ten-year
period, Frakes and Wasserman track individual examiners over the course of a series
of promotions that carry reductions in examination time allocations. After performing
a range of specification checks, in part designed to rule out other explanations for the
findings, the authors find that as examination time becomes more constrained, patent
examiners grant patents at substantially higher rates, conduct less prior art search-
ing, and perform fewer time-intensive prior-art-based rejections, mainly obviousness
rejections. Under an assumption that examiners, if given sufficient time, will grant
applications at an unbiased rate, the findings from this analysis suggest that current
time allocations may indeed represent binding time constraints and may be causing
examiners to grant too many patents. Their results imply that if all examiners were
given as much time to review applications as GS-level 7 examiners, the USPTO’s grant
rate would fall by 20 percent.
In the Online Appendix to their Review of Economics and Statistics article, Frakes
and Wasserman (2017) further demonstrate that, of the four possible explanations listed
above, the time allocation effects and hiring-year cohort effects may be most responsible
for the monotonically increasing relationship documented by Lemley and Sampat
(2012) between experience levels and grant rates. Higher levels of promotion within the
USPTO contribute to higher grant rates for the reasons just stated. Moreover, Frakes and
Wasserman estimate that, all else being equal (e.g., even after accounting for disposition
year fixed effects and other factors), examiners hired in the 1990s hold on to substantially
higher grant rates throughout their careers relative to examiners hired into the 2000s9—a
pattern that also explains at least some degree of the baseline observation of higher grant
rates associated with more years of examiner experience. After controlling for examina-
tion time allocations and hiring year effects, the relationship between examiner experience
in years and grant rates surprisingly takes on an inverse-U relationship, increasing at first
but then turning strongly negative.
9
In an analysis published in the Duke Law Journal (2016), Frakes and Wasserman argue that
this pattern of cohort effects is consistent with fluctuations over time in the USPTO’s granting
culture. USPTO heads promoted an atmosphere of permissive granting practices throughout the
1990s—at least to a relatively greater degree than they encouraged in the mid-2000s. The culture of
the USPTO at the time of hiring may imprint upon examiners as they develop enduring practice
styles that persist even in the face of future swings in the USPTO’s philosophy regarding the
application of patentability standards.

In a series of other papers, Frakes and Wasserman (2013, 2015) empirically explore
an additional set of explanations behind the USPTO’s perceived over-granting. In the
broadest terms, these studies suggest that the USPTO may respond to its highly publicized
financial instability and to its substantial backlog of applications awaiting review by
granting more permissively to certain types of applicants, drawing on various features
of the USPTO that provide financial relief in the event that patents are granted. Though
these papers do not purport to explain all determinants of the USPTO’s grant rate and
do not purport to fully explain observed levels or trends in grant rates, they do suggest, at
the least, that the USPTO’s observed grant rates may be higher than they otherwise would
have been absent financial pressures.
Frakes and Wasserman’s first contribution in this series targets the fee structure of the
USPTO. By way of quick background, the USPTO has been fully user fee funded since
1991, with roughly 85 percent of its fee collections coming from three types of fees: (1)
examination/application fees, (2) issuance fees which are collected at the time of issuance
of the patent, and (3) maintenance fees which are collected at four, eight, and 12 years
following the issuance of the patent in order for grantees to renew their patent protection.
Even though the cost to the USPTO of issuing and renewing patents is minimal, over
half of the USPTO’s fees come from the latter “post-allowance” fees. Moreover, while
the majority of the USPTO’s operational costs stem from examining patent applications,
the examination fees are set to cover only one-third of the costs incurred by the USPTO
to review an application. The fee structure of the USPTO raises a question: given that
the majority of the USPTO’s operating revenue comes only in the event that it grants
patents, might this cause the USPTO to grant patents at elevated rates—that is, to allow
applications that it would have otherwise rejected based on an application of proper
patentability standards? Frakes and Wasserman (2013) explore this question using a
range of natural experiment approaches. Mainly, they test for whether the USPTO began
to grant at higher rates: (1) when the USPTO adopted its user fee funded system in 1991
(relative to the pre-1991 period when it was largely funded by Congress) and thus when
this incentive was potentially created in the first place; (2) when the USPTO experienced
aggregate financial shocks that left it with less fee revenues available to satisfy the cost of
all applications awaiting review—that is, when the USPTO found itself in a position where
it needed additional revenue; and (3) when Congress, in 2004, weakened the mechanism by
which it had previously diverted USPTO fee collections away from USPTO use, thereby
making it easier for the USPTO to retain its fees and bolstering any existing incentives it
may have to elevate its fee revenues. While the authors predicted an elevation in grant rates
in each instance, they further predicted that any increase in granting would be targeted
on those types of applications that garner the greatest fee revenue for the USPTO. Such
types include large entity applicants (which pay double the fees) and applicants in tech-
nologies that consistently maintain their patents at higher rates (e.g., genetics and many
information and computer technologies).
Frakes and Wasserman (2013) find that the USPTO does appear to grant at higher
rates to these high fee-generating types when the USPTO experiences financial shocks
that leave it less able to satisfy its existing examination demand with its incoming fee
collections. Additionally, during times of financial strain, the USPTO’s backlog of
patent examinations is expanding. As such, the findings suggest that the USPTO may be
attempting to open up resources to deal with its growing backlog by biasing its granting

patterns upwards.10 It appears that the USPTO is balancing the harms associated
with elevated patenting with the harms associated with delays in its examinations that
naturally occur with large backlogs. Interestingly, Frakes and Wasserman (2013) do not
find that the USPTO attempts to increase revenues in all circumstances, even when its
financial state is healthy. They characterize the USPTO as being resource-constrained
but benevolent as opposed to budget maximizing. Finally, Frakes and Wasserman (2013)
document a notable spike in granting proclivities (among the high fee-generating patent
types) following 2004, when the USPTO attained a greater ability to retain its grant
rates.11
Frakes and Wasserman (2015) expand upon and refine this analysis in a follow-up
paper published in the Stanford Law Review. In this subsequent analysis, they likewise
theorize that in response to financial unsustainability and, concomitantly, in response to
the USPTO’s growing backlog, they may elevate their granting rates. But now, they do
not focus on features of the USPTO that allow it to raise more revenues—for example,
its fee structures. Instead, they focus on an additional feature of the USPTO that bears
on the financial health of the USPTO: its inability to finally reject a patent application
stemming from the ability of an aggrieved patent applicant to restart the examination
process by filing a continuation application or a request for continued examination
(RCE). In simple terms, if applicants can keep continuing their applications, the USPTO
may become overwhelmed and face a piling up of applications. The USPTO may attempt
to avoid this outcome and discourage the use of these repeat filing devices by simply
granting applications earlier in the process. Similar to the previous study (Frakes and
Wasserman, 2013), Frakes and Wasserman (2015) theorize that the USPTO may target
this elevated granting response on certain types of applicants over others, this time with
respect to applicants in those technologies that have consistently exhibited a greater
tendency to pursue continuation filings and RCEs upon rejections. They likewise find
evidence that upon negative resource shocks to the USPTO—or upon increases in the
USPTO’s aggregate backlog—the USPTO begins to grant at elevated rates to the targeted
technologies. Consistent with the hypothesized use of such elevated granting practices to
10
Again, Frakes and Wasserman (2013) do not contend that these financial motivations are
the only motivations driving the USPTO’s granting decisions; rather, they simply demonstrate that
financial considerations may matter on the margin.
11
Frakes and Wasserman (2014) theorize that the fee structure of the USPTO may, in part,
be responsible for the growth of the backlog in the first place. So not only may the fee structure
create an incentive to grant more to raise more money, which the USPTO may act upon when
it faces financial strain, but the fee structure may also contribute to that financial strain in the
first place. The idea here is also straightforward. By relying so extensively on post-allowance/
back-end fees to subsidize the cost of examination and by setting examination fees themselves
far below the cost of examination, the USPTO is exposing itself to a risk that the rate of
incoming applications will grow out of step with the rate of growth of the existing stock of
patents from which it garners post-allowance fees. Upon this imbalance, the USPTO will not
generate enough fee income to sustain that application growth and a backlog of applications
will naturally ensure.
Hegde (2012) similarly theorizes that certain features of the appropriations process with
Congress—for example, the delay between the time the USPTO submits its proposed budget and
when it is approved by Congress—expose the USPTO to risks that it will not be able to cover the
expense associated with an increase in examination demand over the intervening period.

stem off continuation filings, the authors likewise find a concurrent dip in year-to-year
continuation rates for the targeted technologies.12
Finally, Jaffe and Lerner (2004) provide some evidence that the examination standards
of the USPTO declined in the mid-1990s; but unlike the other studies discussed in this
section, they do not empirically identify a particular feature of the USPTO that is caus-
ing it to allow too many invalid patents. Jaffe and Lerner compared the growth of U.S.
originated patent families—patents that are directed to the same underlying inventions
that are filed and granted in multiple countries—with successful applications filed by U.S.
inventors in the United States that were not part of a patent family over the time period of
1987 through 1998. They reason that if examination standards were not changing in the
USPTO, then successful application in the United States by U.S. inventors should grow at
the same rate as U.S. originated patents that were also granted by the EPO and the Japan
Patent Office (JPO). Because they found the former grew at a rate that was twice as much
as the latter, they conclude the examination standards of the USPTO declined.
There may of course be reasons other than declining quality of examination review
that explain diverging trends between the rate of issuance of patents at the USPTO and
the rate by which U.S.-originated patents issue at the EPO and JPO. For instance, perhaps
there were general marketplace developments (or other developments in the innovation
space) that may have encouraged relatively more filings to be directed at the USPTO only
(a possibility that is difficult to disentangle when only working with counts of issued pat-
ents). Moreover, it may also be difficult to disentangle administrative developments at the
USPTO that bear on how it applies patentability standards from substantive developments
that reflect changes in policymakers’ or judges’ views over the fundamental nature of those
standards themselves. This distinction is important for this chapter insofar as our goal
has been to explore features of the USPTO that may bias its application of patentability
standards, whatever those standards may be—that is, our goal has been to identify reasons
for the issuance of invalid patents, where invalidity is identified by deviations from the
patentability requirements. EPO/JPO benchmarking as a means of assessing quality of the
prosecution process is useful as an exercise to the extent the legal patentability requirements
are comparable across these various institutions. There may be reason to question stability
in this comparability over the specific sample period employed by Jaffe and Lerner (2004).
We know that over their period of study (1987–98), U.S. patent eligibility standards were
becoming more relaxed relative to those of Japan and Europe, as the USPTO for the first
time enacted extensive guidelines regarding the patent eligibility of software, whereas the
EPO and JPO continued to disallow such applications. As such, patent applicants may
have chosen to file more applications in the USPTO than its foreign counterparts simply
because more inventions became patent eligible in the U.S. than elsewhere.
12
The authors also demonstrate that their new analysis (Frakes and Wasserman, 2015) is
empirically distinct from their prior analysis (Frakes and Wasserman, 2013). That is, they establish
that an increase in grant rates among the high continuation rate technologies cannot be explained
by the fact that such technologies overlap with high fee-generating technologies. The authors note,
however, that when estimating for both behavioral responses simultaneously, they find that the use
of elevated granting to raise additional fee revenues does not appear to materialize until the point
at which Congress changed its fee diversion practices in 2004 to make it easier for the USPTO to
actually retain any additional fees that it collects.

VI. CONCLUDING REMARKS
This chapter has provided a brief overview of the empirical literature that has explored
the patent examination process. More specifically, it has focused on studies that carry
a significant empirical component and that focus primarily on the U.S. patent system.
Although the studies discussed in this chapter have helped to shed light on the patent
application review process, including various features of the patent system that may
distort the USPTO’s decision making, the field is still in its infancy and there is ample
room for additional work in the area. This chapter will conclude by quickly delineating
two fruitful areas for additional research on the patent examination process.
The studies discussed in this chapter have largely utilized either a sampling period
of issued patents or application level data, the latter of which enables the exploration
of what influences an examiner’s decision to grant a patent. The examination process,
however, is far richer than a binary grant-versus-reject decision might suggest. The
economic significance of a patent, and hence the potentially negative consequences of
issuing an overly broad patent, depend on the patent’s scope. The broader the scope of
the patent, the larger the number of competing products and processes that the patent
can potentially block; whereas the narrower the scope of the patent, the smaller the
number of products and processes that the patent can potentially block. The job of
a patent examiner is not simply to decide whether to grant an application given the
claims set forth in the application, but also to force the applicant to narrow the scope
of the requested claims until the application satisfies the legal patentability standards.
Understanding how various features of the patent system affect not only the examiner’s
decision to grant a patent, but also the scope of patents issued, is essential to increase
the quality of patent issued by the USPTO. While a few studies discussed above have
examined the role examiners play in narrowing the scope of claims issued, much more
work is needed in this area.
More work is also needed in understanding the various levels at which decision making
operates within the USPTO. Many have viewed the USPTO as being at the whim of the
discretion of the individual examiners. The examiner heterogeneity literature discussed
above certainly supports this notion. However, even Cockburn et al. (2003) themselves
note that much of the heterogeneity they find could be a reflection of practice styles at
higher administrative units within the USPTO—for example, at the Art Unit level. Frakes
and Wasserman (2013, 2014, 2015) take a novel approach and model the decision making
of the USPTO itself rather than the decision making of individual examiners.13 Their
findings support the notion that the USPTO does have the ability to direct examination
outcomes to serve USPTO-level goals. Nevertheless, more work is needed in understand-
ing how other influences, such as mid-level managers or quality review, affect patent office
outcomes.
13
To be clear, Frakes and Wasserman (2013, 2014, and 2015) recognize the import that
examiner-level incentives play in USPTO outcomes, but examine how features of the patent system
may affect USPTO-level incentives.

REFERENCES
Alcacer, Juan, Michelle Gittelman, and Bhaven Sampat. 2009. ‘Applicant and Examiner Citations in U.S.
Patents: An Overview and Analysis’, 38 Research Policy 415–27.
Carley, Michael, Deepak Hedge, and Alan Marco. 2015. ‘What is the Probability of Receiving a U.S. Patent’,
17 Yale Journal of Law and Technology 203–23.
Cockburn, Iain, Samuel Kortum, and Scott Stern. 2003. ‘Are All Patent Examiners Equal? Examiners, Patent
Charactersitics, and Litigation Outcomes’, in Wesley Cohen and Stephen Merrill, eds., Patents in the
Knowledge Based Economy. National Academies Press.
Cotropia, Christopher, Mark Lemley, and Bhaven Sampat. 2013. ‘Do Applicant Citations Matter?’, 42 Research
Policy 844–54.
Frakes, Michael, and Melissa Wasserman. 2013. ‘Does Agency Funding Affect Decisionmaking?: An Empirical
Assessment of the PTO’s Granting Patterns’, 66 Vanderbilt Law Review 67–148.
Frakes, Michael, and Melissa Wasserman. 2014. ‘The Failed Promise of User-Fees: Empirical Evidence from
the U.S. Patent and Trademark Office’, 11 Journal of Empirical Legal Studies 602–36.
Frakes, Michael, and Melissa Wasserman. 2015. ‘Does the U.S. Patent and Trademark Office Grant Too Many
Bad Patents?: Evidence from a Quasi-Experiment’, 67 Stanford Law Review 613–76.
Frakes, Michael, and Melissa Wasserman. 2017. ‘Is the Time Allocated to Review Patent Applications Inducting
Examiners to Grant Invalid Patents?: Evidence from Micro-Level Application Data’, 99 Review of Economics
Frakes, Michael, and Melissa Wasserman. 2016. ‘Patent Office Cohorts’, 65 Duke Law Journal 1601–55.
Hegde, Deepak. 2012. ‘Funding and Performance the US Patent and Trademark Office’, 30 Nature
Biotechnology 148–50.
Jaffe, Adam, and Josh Lerner. 2004. Innovation and Its Discontents: How Our Patent System is Endangering
Innovation and Progress, and What to Do About it. Princeton University Press.
Lei, Zhen, and Brian Wright. 2017. ‘Why Weak Patents? Testing the Examiner Ignorance Hypothesis’, 148
Journal of Public Economics 43–56.
Lemley, Mark, and Bhaven Sampat. 2008. ‘Is the Patent Office a Rubber Stamp?’, 58 Emory Law Journal
181–203.
Lemley, Mark, and Bhaven Sampat. 2010. ‘Examining Patent Examination’, 2010 Stanford Technology Law
Review 2–14.
Lemley, Mark, and Bhaven Sampat. 2012. ‘Examiner Characteristics and Patent Office Outcomes’, 94 Review
Lichtman, Douglas. 2004. ‘Rethinking Prosecution History Estoppel’, 71 University of Chicago Law Review
151–82.
Mann, Ronald. 2014. ‘The Idiosyncrasy of Patent Examiners: Effects of Experience and Attrition’, 92 Texas
Quillen, Cecil, and Ogden Webster. 2001. ‘Continuing Patent Applications and Performance of the U.S. Patent
and Trademark Office’, 11 Federal Circuit Bar Journal 1–21.
Quillen, Cecil, and Ogden Webster. 2002. ‘Continuing Patent Applications and Performance of the U.S. Patent
and Trademark Office–Extended’, 12 Federal Circuit Bar Journal 35–55.
Sampat, Bhaven. 2010. ‘When Do Patent Applicants Search for Prior Art?’, 53 Journal of Law and Economics
399–416.
Tu, Shine. 2012. ‘Luck/Unluck of the Draw: An Empirical Study of Examiner Allowance Rates’, 2012 Stanford
Vishnubhakat, Saurabh, and Arti Rai. 2015. ‘When Biopharma Meets Software: Bioinformatics at the Patent
Office’, 29 Harvard Journal of Law and Technology 205–41.
Leahy-Smith America Invents Act, Pub. L. No. 112-29, 125 Stat. 284 (2011).

5. The USPTO’s Patent Trial and Appeal Board* 14
Arti K. Rai** and Saurabh Vishnubhakat***
Contents
I. Introduction
II. The Path to the PTAB
A. Why Administrative Review?
B. Opposition Mechanisms Prior to the PTAB
C. Administrative Opposition under the AIA
III. How has the PTAB Functioned?
A. Basic Descriptive Statistics
B. Intersection with District Court Litigation
C. Impact on Non-Practicing Entities
D. Relationship Between Patent and Patent-Examination Characteristics and
IPR
E. Directions for Future Research
References
I. INTRODUCTION
The post-grant administrative review of patent validity set up at the U.S. Patent and
Trademark Office’s (USPTO) Patent Trial and Appeal Board (PTAB) by the America
Invents Act of 2011 has transformed the relationship between Article III patent litigation
and the administrative state. Section II of this chapter reviews scholarly literature that
has made the case for such review and presents basic data on the functioning of several
predecessors to the PTAB. Section III reviews the nascent empirical literature on the
PTAB’s functioning and sets forth a research agenda.
*** © 2018 Arti K. Rai and Saurabh Vishnubhakat.

*** Elvin R. Latty Professor of Law at Duke Law School, Faculty Co-Director of the Duke
Law Center for Innovation Policy, and the former Administrator for Policy and External Affairs
of the USPTO.
*** Associate Professor of Law at the Texas A&M University School of Law, a Fellow at the
Duke Law Center for Innovation Policy, and a former Expert Advisor to the Chief Economist of
the United States Patent and Trademark Office (USPTO). The arguments in this chapter are the
authors’ own and should not be imputed to the USPTO or to any other organization.
92

The USPTO’s Patent Trial and Appeal Board 93
II. THE PATH TO THE PTAB
This section summarizes the literature making the case for post-grant administrative
review of patent validity. Arguments mooted in the literature, which played out over the
course of several decades, helped to propel the creation of various post-grant institutions.
The most significant of these is the PTAB.
A. Why Administrative Review?
Scholars have long noted that the initial patent examination process at the USPTO is
likely to produce at least some (and perhaps many) improper patent grants. To the extent
improperly granted patents impose unnecessary costs and call into question the credibility
of the patent system, improper grants ought to be corrected (Janis, 1997).
Additionally, scholars have long argued that USPTO errors should be corrected
outside Article III courts. Not only is district court litigation very expensive (American
Intellectual Property Law Association (AIPLA), 2013), but these great expenditures
don’t necessarily yield great accuracy in judicial outcomes. Because patent law often uses
science-based proxies such as “ordinary skill in the art” to tackle relevant legal and policy
goals, the subject matter of patent law can be highly complex as a scientific matter (Rai,
2003). With the possible exception of Federal Circuit judges, judges in the federal courts
tend to be generalists who may not be equipped to tackle complex questions at the inter-
section of law, science, and policy (Lee, 2010). District courts also have to contend with
juries, which may be even less equipped than federal judges to address complex questions
of law and science (Lemley, 2013).
For all of these reasons, theory would suggest that district court litigation is unlikely to
be the best mechanism for achieving the standard economic goal of minimizing the sum
of litigation costs and error costs (Kaplow and Shavell, 1994, 1996).
Empirical evidence bears out concerns about the capacity of district court judges to
resolve patent disputes. The generalist background of most district judges has led many
plaintiffs to seek out specific districts (Atkinson, Marco, and Turner, 2009; Moore, 2001),
with the natural result being a certain amount of de facto specialization. Among the subset
of judges who preside over patent cases regularly, increased experience may produce more
efficient and accurate case outcomes (Kesan and Ball, 2011). Yet this private ordering
toward certain districts only underscores the overall lack of expertise among district court
judges. Moreover, commentators have argued that aggressive attempts to specialize in
patent disputes by judges whose districts are found outside traditional technology centers
lead to overly plaintiff-friendly procedures rather than accurate adjudication (Anderson,
2015; Fromer, 2010; Klerman and Reilly, 2016).
In contrast, under conventional principles of administrative law, agencies are reposito-
ries of expertise. More specifically, administrative patent judges within the USPTO have
long been required to be “persons of competent legal knowledge and scientific ability”
(35 U.S.C. § 6, 2012).
Another reason to favor low-cost administrative review, rather than high-cost Article
III review, is that patent plaintiffs and defendants have asymmetric incentives. Supreme
Court case law builds into the patent doctrine asymmetric incentives to litigate. Under
the law’s estoppel provisions, a challenger that successfully invalidates a patent provides

a public good—the challenger benefits not only itself, but also all other potential
challengers. By contrast, a challenger that loses is uniquely estopped from challenging the
patent again (Blonder-Tongue Labs., Inc. v. Univ. of Ill. Foundation, 402 U.S. 313 (1971);
Farrell and Merges, 2004; Thomas, 2001).
Although this public-good-type problem may exist in the administrative context
as well (and, indeed, exists in both pre-America Invents Act (AIA) and post-AIA
inter partes proceedings as a consequence of the congressional decision to implement
relatively strong statutory estoppel provisions), the lower cost of the administrative
proceeding presumably reduces the scale of the problem. Moreover, because administra-
tive proceedings do not have a constitutionally based Article III standing requirement,
they create possibilities for additional entities—including entities that represent groups
of potential defendants in a given industry—to provide the public good of invalidating
an improperly granted patent.
For these reasons, a less costly, more expert, and more widely accessible institution
has long been thought desirable. On the other hand, even advocates of an administrative
mechanism have noted the potential for harassment of patent owners that might arise
in such administrative review (Kesan, 2002). Harassment potential exists as the obvious
flipside of the access created by low cost and the absence of an Article III standing
requirement. Moreover, to the extent that courts do not believe that administrative review
will in fact be accurate and efficient, and thus do not stay any related Article III litigation,
such review may create costly duplication rather than efficiency.
B. Opposition Mechanisms Prior to the PTAB
In 1980, Congress instituted a limited version of post-grant review by creating a

mechanism for ex parte administrative reexamination of patent validity (Bayh–Dole Act,
1980). A contemporaneous opinion from the Federal Circuit noted that the goal of the
procedure was to produce efficiency and expertise gains. Patlex Corp. v. Mossinghoff, 758
F.2d 594 (Fed. Cir. 1985).
Cost-wise, the USPTO’s ex parte reexamination fee has grown from $1500 in the early
1980s (50 Fed. Reg. 31818, 1986) to $12 000 at present (37 CFR § 1.20(c)(1), 2012) and
attorney costs have risen to approximately $20 000 at the mean and $15 000 at the median
(AIPLA, 2013). Thus, even today, the total expense of ex parte reexamination tends to
be below $35 000—some 20 times less costly than the lowest-stakes category of litigation.
Consistent with its mandate to correct examination errors, ex parte reexamination
requires a “substantial new question of patentability” as to one or more of the challenged
patent claims, and this standard may be met by reargument of information that was previ-
ously before the patent examiner (35 U.S.C. § 303(a), 2012). Access to reexamination is
unconstrained by traditional Article III standing requirements. Anyone at any time may
seek reexamination of a patent, including the patent owner and the USPTO itself (35
U.S.C. §§ 302, 303(a), 2012).
From the outset, however, commentators have criticized the reexamination procedure
for its ex parte nature, which excludes any third-party participation beyond filing the
initial request (Janis, 1997; Casey, 1995; Motsenbocker, 1994; Neff, 1986). According
to USPTO statistics (USPTO, 2014), 29 percent of the ex parte reexaminations filed
between July 1981 and September 2014 were filed by the patent owner itself, presumably

as a potential mechanism for strengthening the patent. Indeed, certain Federal Circuit
cases have indicated that patents that survive reexamination should be viewed even more
deferentially by the courts than ordinary patents. Gould v. Control Laser Corp., 705 F.2d
1340 (Fed. Cir. 1983).
To improve public participation in the administrative review of patent validity, in 1999,
Congress created a new procedure: inter partes reexamination. Designed to coexist with
the old ex parte procedure, inter partes reexamination conferred significant rights upon
third-party requestors to participate in the USPTO’s review of patent validity. A requestor
could comment on every substantive response by the patent owner to an examiner action
and could appeal the examiner’s decision to the USPTO’s administrative review board.
However, inter partes reexamination also posed significant barriers. One was a strong
estoppel provision, barring the challenger from raising in Article III litigation any issues it
raised or could have raised during the inter partes reexamination (Thayer, 2004). Perhaps
even more significant was the prolonged duration of reexamination. The reexaminations
themselves took an average of 39.5 months, and then had to be appealed to the Board
of Patent Appeals and Interferences (USPTO, 2016a). From its creation in 1999 through
2012, when it was abolished by the new administrative review system established by
the AIA, inter partes reexamination was never widely used as a means for challenging
the validity of patents. From November 29, 1999, through the abolition of inter partes
reexamination, effective September 16, 2012, fewer than 2000 requests were filed, and in
most years the usage of inter partes reexamination represented only a fraction of ex parte
reexamination (USPTO, 2016a).
C. Administrative Opposition under the AIA
The legislative path to a more robust system was advanced significantly by two landmark
reports, one from the Federal Trade Commission (2003) and the other from the National
Academies of Science (2004). Both of these reports strongly supported robust post-grant
review.
The AIA abolishes inter partes reexamination and creates four new procedures for
reevaluating the validity of patents. One procedure, post-grant review, is just beginning its
operation, as it only applies to patents that issue from applications filed under the AIA’s
new first-inventor-to-file regime (i.e. patent applications with a priority date on or after
March 16, 2013). Another procedure, supplemental examination, allows patent owners
themselves to provide new information that helps fortify the validity of their patents
against potential charges of inequitable conduct.
Thus far, the opportunity for robust public participation in challenging existing patents
has arisen in the two remaining procedures: inter partes review (IPR) and the transitional
program for covered business method (CBM) review. Of the two proceedings, IPRs have
been more frequently used and discussed from an empirical standpoint. The empirical
discussion of this chapter will likewise focus on IPRs.
In keeping with the principle that questions of patent validity should be decided
outside the courts, the AIA contemplates challengers bringing PTAB petitions on
validity after they have been sued in district court. Specifically, the AIA requires that a
challenger bring a petition for IPR review on a given patent within one year of the filing
of a district court lawsuit asserting that patent against the challenger (35 U.S.C § 315

(b), 2012). Meanwhile, CBM review explicitly requires the petitioner to be “charged with
infringement,” language the USPTO has interpreted as requiring the petitioner to prove
standing necessary to bring a declaratory judgment action in district court (37 C.F.R.
§ 42.302(a), 2012).
That said, in the case of IPRs, the statutory language certainly does not preclude
petitioners which have previously not been sued, or even threatened with suit. IPR
proceedings have no constitutional standing requirement (35 U.S.C. § 311(a), 2012) and
the USPTO can continue to conduct IPRs even if the challenger drops out of the case (35
U.S.C. § 317(a), 2012). Administrative review also differs from district court litigation in
that challengers in IPR and CBM review must prove invalidity only by a preponderance
of the evidence (35 U.S.C. § 316(e), 2012). By contrast, challengers in district court litiga-
tion must prove invalidity by clear and convincing evidence. Microsoft Corp. v. i4i Ltd.
P’ship, 564 U.S. 91 (2011). The Supreme Court decision in Cuozzo Speed Technologies,
LLC v. Lee, 136 S. Ct. 2131 (2016) noted these differences between administrative review
and district court litigation in upholding as reasonable under Chevron v. Nat. Res. Def.
Council, 467 U.S. 837 (1984) the USPTO regulation directing the PTAB to use a different
claim construction standard than district courts.
An IPR challenger may attack patent claims only on the grounds that they fail to satisfy
the novelty or nonobviousness requirements. To decide that an IPR petition warrants
institution of an IPR proceeding, the USPTO must find a “reasonable likelihood that
the petitioner would prevail with respect to at least [one] of the claims challenged in the
petition” (35 U.S.C. § 314(a), 2012).
Both IPR and CBM review proceedings generate estoppel effects, though not in the
same way. The estoppel generated by IPR is quite strong. An IPR resulting in a final
written decision precludes the petitioner from asserting any claim in either the USPTO,
the federal courts, or the International Trade Commission that the petitioner raised, or
could have raised, in the IPR proceeding (35 U.S.C. § 315(e), 2012).
By contrast, a CBM review that results in a final written decision creates full estoppel
within the USPTO only—it precludes the petitioner from asserting any claim in the
USPTO that the petitioner raised, or could have raised, in the CBM review proceeding (35
U.S.C. § 325(e)(1), 2012). With respect to the courts, CBM petitioners are not estopped
from invoking the invalidity grounds that they raised at the USPTO.
Both IPR and CBM review proceedings trigger automatic stays of co-pending declara-
tory judgment litigation. Just as a would-be petitioner cannot challenge a patent in an
IPR if it has previously challenged that patent in a civil action (35 U.S.C. §§ 315(a)(1), (3),
2012), if a petitioner files such a civil action after the IPR petition, then that civil action
is automatically stayed (35 U.S.C. § 315(a)(2), 2012). Likewise, if a petitioner files a civil
action challenging the patent after filing a CBM petition on the same patent, then that
civil action must automatically be stayed (35 U.S.C. § 325(a)(2), 2012).
Meanwhile, courts still have the discretion to stay existing infringement litigation
brought by a patent owner pending the outcome of an IPR or CBM review proceeding.
For IPRs, where the AIA does not specify the standard for such stays, prior standards
pertaining to ex parte and inter partes reexamination remain valuable (though not neces-
sarily conclusive) precedent. Drawing on the reexamination case law, courts continue to
consider the familiar three factors in deciding whether to issue stays: the potential for
prejudice or tactical disadvantage; the timing of the desired stay relative to that of the

administrative proceeding itself; and the likelihood that resolution of the administrative
proceeding may simplify the pending litigation (Frontz, 2015).
In contrast, the AIA specifies a four-factor test for CBM-related stays. This four-
factor test encompasses the three factors courts previously used in determining contested
motions for stay under the old reexamination system and adds a fourth factor—“whether
a stay, or the denial thereof, would reduce the burden of litigation on the parties and on
the court” (Leahy-Smith America Invents Act, Pub. L. No. 112-29, § 18(b)(1), 2011).
Moreover, for purposes of “ensur[ing] consistent application of established precedent,”
the AIA provides for immediate interlocutory appeal of the district court’s decisions
regarding stays. It also states the Federal Circuit’s standard of review on appeal from a
district court decision “may be de novo.” Using this standard, the Federal Circuit has
held that district courts have limited discretion to deny CBM-related stays when all claims
asserted in litigation are also under CBM review. VirtualAgility Inc. v. Salesforce.com, Inc.,
759 F.3d 1307 (Fed. Cir. 2014).
III. HOW HAS THE PTAB FUNCTIONED?
Given the recency of the PTAB’s creation, the empirical literature on its functioning is
necessarily limited. Thus far, this literature divides into four major categories: (1) basic
descriptive statistics; (2) the intersection (or lack thereof) with district court litigation,
including across different technology areas and different courts; (3) the impact of the
PTAB on non-practicing entities (NPEs); and (4) the relationship between various patent
and patent-examination characteristics and IPRs.
A. Basic Descriptive Statistics
Consistent with early commentary noting that the strong estoppel provisions associated
with IPRs might chill such use, use of both IPRs and CBMs began slowly. The popularity
of these proceedings began to rise starting in early 2014, to an average of about 100–150
IPR petitions per month and approximately 10–15 CBMs per month (USPTO, 2016;
Vishnubhakat, Rai, and Kesan, 2016).
Thus far, IPRs do appear substantially cheaper than district court litigation. According
to the 2015 AIPLA Economic Survey, the median cost of an IPR through a PTAB hearing
was $275 000 and the cost through appeal was $350 000. (AIPLA, 2015). Although the
AIPLA survey does not differentiate between IPRs based on amount of money at risk,
these figures are more than 50 percent lower than the median cost of district court litiga-
tion, even for the lowest-stakes cases.
Breakdown by technology depends to some extent on time period sample, technology
classification system used (e.g. USPTO Technology Centers or the National Bureau of
Economic Research (NBER) six-part technological classification developed by Hall,
Jaffe, and Trajtenberg (2001)), and whether IPRs and CBMs are pooled together or
treated separately.
Vishnubhakat, Rai, and Kesan (2016) use NBER technological classifications and
study IPRs and CBMs filed between September 16, 2012 (when the PTAB began opera-
tions) and June 30, 2015. They find that at least half of IPR petitions involve Computer

or Communications Technology. Electrical technologies are challenged in another 15.4

percent of IPR petitions; chemical technologies (excluding drugs) in about 6 percent; and
drugs and medicine in about 13 percent.
As for CBM petitions, because CBM proceedings are oriented by definition toward
business-method-related technologies such as information and communications, it is
unsurprising that an overwhelming majority (82.2 percent) of CBM petitions challenge
Computers or Communications method-related patents. Mechanical-related patents
make up another 15.9 percent of CBM petitions, and only a negligible share of CBM
petitions fall in any other category (Vishnubhakat, Rai, and Kesan, 2016).
Although IPR petitions may challenge patent claims as to either novelty or non
obviousness, nonobviousness challenges predominate across all major technology areas
(Vishnubhakat, Rai, and Kesan 2016). Unlike IPR petitions, CBM petitions may chal-
lenge patent claims on a fuller range of patentability requirements: in addition to novelty
and nonobviousness, subject-matter eligibility, enablement, written description, and
indefiniteness are available grounds. Across this range of options, petitioners have focused
their attention primarily on subject-matter eligibility and nonobviousness (Vishnubhakat,
Rai, and Kesan, 2016).
Possibly because of selection effects associated with the earlier IPR petitions, the rate
of IPR institution on at least one claim has diminished over time, from over 90 percent
to about 70 percent (Vishnubhakat, Rai, and Kesan, 2016). Across technologies, rates of
institution appear higher for patents emerging from USPTO Technology Centers that
evaluate electrical/computer applications and chemical applications (excluding pharma-
ceuticals) than for biotechnology and pharmaceutical applications (USPTO 2016b).
As of May 2016, the PTAB had processed 3114 IPR petitions and 355 CBM petitions,
with 1580 IPR trials instituted and 192 CBM trials instituted (USPTO, 2016b). For both
IPRs and CBMs, a notable percentage settle before or after the institution decision—over
38 percent of IPR petitions and 33 percent of CBM decisions (USPTO, 2016b). Of the
988 IPRs for which trials were completed, 71 percent held all instituted claims unpatent-
able. Of the 122 CBMs for which trials were instituted, 79 percent held all instituted claims
unpatentable (USPTO, 2016b).
At the claim level, 20 991 of 47 364 claims challenged in IPR were instituted, and 10 597
were found unpatentable in a final written decision. As for CBM review, 3205 of 6893
claims challenged in this review were instituted, and 1871 found unpatentable. In general,
biotechnology and pharmaceutical claims on which proceedings had been instituted
were less likely to be found unpatentable in a final written decision than claims in other
technological areas (USPTO 2016b).
B. Intersection with District Court Litigation
Scholars have explored how USPTO petitions intersect with district court litigation. One
issue is the extent to which administrative review is serving as a substitute for district
court litigation.
At a minimum, patents challenged at the PTAB and at the USPTO overlap quite a bit.
During the period between September 16, 2012 and June 30, 2015, 87 percent of patents
challenged at the PTAB were also in some type of district court litigation (Vishnubhakat,
Rai, and Kesan, 2016). According to another study, which addressed only IPRs and

covered the period from September 16, 2012 to March 31, 2014, district court litigation
involving not only the same patents, but also the same parties occurred in about 80 percent
of cases (Love and Ambwani, 2014).
Moreover, if there is no Article III litigation, Computer and Communications patents
are particularly unlikely (relative to other areas of technology) to be challenged at the
PTAB (Vishnubhakat, Rai, and Kesan, 2016). In this area of technology, district court
assertion may be necessary to force parties to overcome an absence of clarity regarding
the merits of a validity challenge created by lack of boundary notice, as well as informa-
tional hurdles created by the sheer volume of Computer and Communications patents.
The average patent challenged at the PTAB has been asserted at least three times in
court (Vishnubhakat, Rai, and Kesan, 2016). However, this average reflects considerable
variation, and the relationship between the number of IPR or CBM petitions that were
filed on a patent and the number of times that the patent was asserted in district court is
not monotonic. It is clear that, in addition to number of district court assertions, numbers
of PTAB petitions on a given patent are affected by other factors.
One factor that may have an influence is the district court in which the patent is
asserted. Of the eight leading district courts—which together account for nearly 70
percent of litigated patents during the observed time period—the top three courts were
overrepresented in sending patents into PTAB validity challenges (District of Delaware,
Eastern District of Texas, and Northern District of Delaware), and the remaining five
(including Central District of California, District of New Jersey, and Northern District
of Illinois) were underrepresented (Vishnubhakat, Rai, and Kesan, 2016).
The significant overlap between patents challenged in the PTAB and in district court
does not mean, however, that defendants which have previously been sued are the only
entities bringing AIA challenges. To the contrary, in the case of IPRs, 30 percent of
petitioners have not previously been defendants in district court litigation involving the
patents they now challenge (Vishnubhakat, Rai, and Kesan, 2016). This subset of cases is
particularly interesting if one looks across technologies. Perhaps most notably, about 52
percent of petitioners in the Drugs and Medical area have not previously been sued. In
some cases, generic firms may be filing even prior to being sued in order to clear the path
toward eventual entry into the market. In other cases, third parties, including hedge funds
that are shorting the stock of the patent owner, have filed petitions.
Of course, any efficiency gains relative to district court litigation can only be realized
if the parallel district court litigation is stayed. Grant rates on stay motions vary signifi-
cantly by district and by stage of district court proceeding. District court judges also grant
motions in part or deny motion without prejudice until there has been a PTAB decision on
institution. Full denial rates are relatively low, in the order of about 25 percent for IPRs
and 20 percent for CBMs (Vishnubhakat, Rai, and Kesan, 2016). Love and Ambwani
(2014) found that, for instituted IPRs that had a parallel district court suit involving the
same parties, stays were granted, at least in part, 82 percent of the time.
C. Impact on Non-Practicing Entities
Love and Ambwani (2014) investigate the impact of the PTAB on NPEs, which they
define as a group comprising patent-assertion entities, universities, pre-product start-
ups, and IP-holding subsidiaries of product-producing companies. According to this

efinition of NPEs, and during roughly the period from September 16, 2012 to March
d
31, 2014, 48 percent of IPR petitions were brought against NPEs. NPEs were somewhat
more likely than product-producing firms to have IPRs instituted on at least one claim.
However, of claims that reached the merits stage, NPE claims were no more likely to be
invalidated than other claims.
D. Relationship Between Patent and Patent-Examination Characteristics and IPR
A nascent literature initiated by the USPTO (2015) is applying matched case-control

approaches commonly used in studies analyzing which patents enter district court
litigation to systematically investigate relationships between certain patent and patent-
examination characteristics and IPR proceedings. This literature compares patents
that become the subject of IPRs with a matched set that are not subject to IPR review.
(Matching is done based on characteristics that are not the subject of study, specifically
year of issue, art unit to which the patent was assigned, and forward citations.) This litera-
ture also evaluates relationships between patent and patent-examination characteristics
and likelihood of IPR institution. Finally, the literature manually examines final written
decisions in IPRs to determine what they reveal about the adequacy of the examiner’s
prior art search, the appropriateness of the examiner’s evaluation of prior art references,
and the drafting of proper rejections.
As compared with matched samples of patents for which IPRs are not filed, patents
on which IPRs are filed are more likely to be owned by small entities, less likely to be of
foreign origin, and more likely to have a higher number of parent applications. Patents
on which IPRs are filed also tend to have more independent claims and fewer words per
independent claim, though the effect of these variables is smaller in magnitude. The grade
of the primary examiner is also relevant (USPTO, 2015).
A comparison of patents that are instituted relative to those that are not instituted
indicates few differences, with the exception of a cohort effect for patents issued in the
2000–04 time period. This 2000–04 cohort was more likely to have IPRs instituted.
Manual examination of final written decisions revealed that in those cases where at
least one claim was held patentable, the prior art used by the PTAB had been before the
examiner during prosecution 59 percent of the time (USPTO, 2015).
E. Directions for Future Research
As more data on the PTAB accumulates, descriptive statistics of the type discussed in
section III.A will need to be updated and chronological patterns (if any) discerned.
Another important area of description, and possible analysis, will include Federal Circuit
review of the PTAB. Thus far, Federal Circuit review has been quite favorable to the
PTAB (Davis, 2016). As numbers of PTAB appeals increase, systematic comparison to
the Federal Circuit’s treatment of district courts (and trends in numbers of cases appealed
from district courts) may be useful.
The Court of Appeals for the Federal Circuit’s frequent use of summary affirmances
in disposing of appeals from the PTAB may also be worth studying. To the extent that
summary affirmances are being used to affirm results in narrow, fact-specific cases,
their use may be appropriate. But determining this will require detailed review of

cases for purposes of determining the extent (if any) to which the PTAB is developing
law.
As noted, USPTO research on which patents will be the subject of IPRs focuses on
certain patent and patent-examination characteristics. More research should be done on
such characteristics. They are important not only for purposes of prediction, but also as
potential indicators of poor quality (on the theory that in a comparison between matched
samples of patents that have equal private value, IPR petitions are more likely to be filed
on those of lower quality).
However, given that many of the characteristics that determine entry into district
court litigation and IPRs are the same (USPTO, 2015), and that only a subset of patents
asserted in district court litigation are the subject of IPR petitions, researchers should
also develop systematic models assessing the factors that affect IPR filing conditional on
assertion in litigation. As suggested by the work of Vishnubhakat, Rai, and Kesan (2016),
these factors might include area of technology and district court in which the patent was
asserted. Factors that determine the frequency of IPR filing on a given patent may also be
worth examining. The work of Vishnubhakat, Rai, and Kesan (2016) demonstrates that
some patents are the subject of significant serial petitioning at the PTAB.
Further research is also needed to determine whether serial petitioning is being man-
aged in a manner that deters harassment. Relatedly, research is needed on whether serial
petitioning is by the same or different petitioners and, if the latter, whether procedures
such as joinder that have the potential to foster useful collective action are being utilized
properly.
Finally, as more data accumulates, scholars should examine in detail the use of IPRs
to challenge so-called “Orange Book” patents that are subject to the special litigation
provisions of the Hatch-Waxman regime. Moreover, given longstanding concerns about
antitrust problems with settlements in Orange Book cases litigated in Article III courts,
scholars should analyze settlement behavior involving Orange Book patents at the PTAB.
REFERENCES
American Intellectual Property Law Association. 2013. Report of the Economic Survey.
American Intellectual Property Law Association. 2015. Report of the Economic Survey.
Anderson, J. Jonas. 2015. “Court Competition for Patent Cases”, 163 University of Pennsylvania Law Review
631–98.
Atkinson, Scott, Alan C. Marco, and John H. Turner. 2009. “The Economics of Centralized Judiciary:
Uniformity, Forum Shopping, and the Federal Circuit”, 52 Journal of Law & Economics 411–43.
Casey, Shannon M. 1995. “The Patent Reexamination Reform Act of 1994: A New Era of Third Party
Participation”, 2 Journal of Intellectual Property Law 559–73.
Davis, Ryan. 2017. “Fed. Circ. Taking Harder Look At PTAB Appeals”, Law 360.
Farrell, Joseph, and Robert P. Merges. 2004. “Incentives to Challenge and Defend Patents: Why Litigation
Won’t Reliably Fix Patent Office Errors and Why Administrative Patent Review Might Help”, 19 Berkeley
Federal Trade Commission. 2003. To Promote Innovation: The Proper Balance of Competition and Patent
Law and Policy. Available at www.ftc.gov/sites/default/files/documents/ reports/promote-innovation-proper-
balance-competition-and-patent-law-and-policy/innovationrpt.pdf.
Fromer, Jeanne C. 2010. “Patentography”, 85 New York University Law Review 1444–520.
Frontz, Matthew R. 2015. “Staying Litigation Pending Inter Partes Review and Effects on Patent Litigation”,
24 Federal Circuit Bar Journal 469–520.
Hall, Bronwyn H., Adam B. Jaffe, and Manuel Trajtenberg. 2001. “The NBER Patent Citations Data File:

Lessons, Insights and Methodological Tools”, 13 National Bureau of Economics Research Working Paper
No. 8498.
Janis, Mark D. 1997. “Rethinking Reexamination: Toward a Viable Administrative Revocation System for U.S.
Patent Law”, 11 Harvard Journal of Law & Technology 1–122.
Kaplow, Louis, and Steven Shavell. 1994. “Accuracy in the Determination of Liability”, 37 Journal of Law &
Economics 1–15.
Kaplow, Louis, and Steven Shavell. 1996. “Accuracy in the Assessment of Damages”, 39 Journal of Law &
Kesan, Jay P. 2002. “Carrots and Sticks to Create a Better Patent System”, 17 Berkeley Technology Law Journal
763–97.
Kesan, Jay P., and Gwendolyn G. Ball. 2011. “Judicial Experience and the Efficiency and Accuracy of Patent
Adjudication: An Empirical Analysis of the Case for a Specialized Patent Trial Court”, 24 Harvard Journal
Klerman, Daniel, and Greg Reilly. 2016. “Forum Selling”, 88 Southern California Law Review 241–316.
Lee, Peter. 2010. “Patent Law and the Two Cultures”, 120 Yale Law Journal 1–82.
Lemley, Mark. 2013. “Why Do Juries Decide If Patents Are Valid?”, 99 Virginia Law Review 1673–736.
Love, Brian J., and Shawn Ambwani. 2014. “Inter Partes Review: An Early Look at the Numbers”, 81 Chicago
Law Review Dialogue 93–107.
Moore, Kimberly A. 2001. “Forum Shopping in Patent Cases: Does Geographic Choice Affect Innovation?”,
79 North Carolina Law Review 889–938.
Motsenbocker, Marvin. 1994. “Proposal to Change the Patent Reexamination Statute to Eliminate Unnecessary
Litigation”, 27 John Marshall Law Review 898–907.
National Academies of Sciences. 2004. Stephen A. Merrill, Richard C. Levin, and Mark B. Myers, eds., A Patent
System for the 21st Century. Washington, D.C.: The National Academies Press.
Neff, Gregor N. 1986. “Patent Reexamination—Valuable, But Flawed: Recommendations for Change”, 68
Journal of Patent & Trademark Office Society 575–95.
Rai, Arti K. 2003. “Engaging Facts and Policy: A Multi-Institutional Approach to Patent System Reform”, 103
Thayer, Patricia. 2004. “Examining Reexamination: Not Yet an Antidote to Litigation”, 5 Sedona Conference
Journal 1–122.
Thomas, John R. 2001. “Collusion and Collective Action in the Patent System: A Proposal for Patent Bounties”,
2001 University of Illinois Law Review 305–53.
Vishnubhakat, Saurabh, Arti K. Rai, and Jay P. Kesan. 2016. “Strategic Decision Making in Dual PTAB and
District Court Proceedings”, 31 Berkeley Technology Law Journal 45–116.
USPTO. 2014. Ex Parte Reexamination Filing Data. Available at www.uspto.gov/learning-and-resources/
statistics/reexa mination-information.
USPTO. 2015. Patent Review Processing System. Available at www.uspto.gov/sites/default/files/documents/2015-
10-31%20PTAB.pdf.
USPTO. 2016a. Inter Partes Reexamination Filing Data. Available at www.uspto.gov/sites/default/files/documen
ts/inter_parte_historical_st ats_roll_up.pdf.
USPTO. 2016b. Patent Trial and Appeal Board Statistics. Available at www.uspto.gov/sites/default/files/
documen ts/2016-5-31%20PTAB.pdf.
Case List
Blonder-Tongue Labs., Inc. v. Univ. of Ill. Found. (1971), 402 U.S. 313.
Chevron v. Nat. Res.’ Def. Council (1984), 467 U.S. 837.
Cuozzo Speed Tech.’s, LLC v. Lee (2016), 136 S. Ct. 2131.
Gould v. Control Laser Corp. (1983), 705 F.2d 1340.
Microsoft Corp. v. i4i Ltd. P’ship. (2011), 564 U.S. 91.
Patlex Corp. v. Mossinghoff (1985), 758 F.2d 594.
Virtual Agility Inc. v. Salesforce.com, Inc. (2014), 759 F.3d 1307.
35 U.S.C. § 302 (2012).

35 U.S.C. § 303(a) (2012).
35 U.S.C. § 311(a) (2012).
35 U.S.C. §§ 315(a)(1), (3) (2012).

35 U.S.C. § 315(a)(2) (2012).

35 U.S.C. § 315(b) (2012).
35 U.S.C. § 315(e) (2012).
35 U.S.C. § 316(e) (2012).
35 U.S.C. § 317(a) (2012).
35 U.S.C. § 325(a)(2) (2012).
35 U.S.C. § 325(e)(1) (2012).
35 U.S.C. § 6 (2012).
37 C.F.R. § 1.20(c)(1) (2012).
37 C.F.R. § 43.302(a) (2012).
Bayh–Dole Act, Pub. L. No. 96-517, 94 Stat. 3015 (1980).
Leahy-Smith America Invents Act, Pub. L. No. 112-29, § 18(b)(1), 125 Stat. 284 (2011).
Revision of Patent Fees, 50 Fed. Reg. 31818 (Aug. 6, 1968), codified in various parts of 37 C.F.R. Part 1.

6. The Federal Circuit as an institution
Ryan Vacca* 15
Contents
I. Introduction
II. Creation of the Federal Circuit
A. Concerns and Inadequacies
B. A Court with a Mission
III. Uniformity
A. Forum Shopping
B. Diversity and Percolation
IV. Certainty and Predictability
A. Pressure for and Undermining Certainty
B. Reversal Rates, Panel Dependency, and Indeterminacy
1. Validity
2. Claim construction
3. Infringement
V. Quality
A. Formalism
B. Scholarship
VI. Structure and Staffing
A. Individual Judges
1. Judicial backgrounds and characteristics
2. Turnover
3. Case distribution
B. Internal Dynamics
C. En Banc Review
VII. Relationships with Other Institutions
A. Supreme Court
B. USPTO
C. District Courts
D. Solicitor General
E. Congress
VIII. Conclusion
References
* Professor of Law, University of New Hampshire School of Law.
104

The Federal Circuit as an institution 105
I. INTRODUCTION
The Court of Appeals for the Federal Circuit is a unique institution. Unlike other circuit
courts, the Federal Circuit’s jurisdiction is bound by subject area rather than geography,
and it was created to address a unique set of problems specific to patent law. These
characteristics have affected its institutional development and made the court one of the
most frequently studied appellate courts. This chapter examines this development and
describes the evolving qualities that have helped the Federal Circuit distinguish itself, for
better or worse, as an institution.
This chapter begins with an overview of the concerns existing before the creation of the
Federal Circuit and then discusses the Federal Court Improvements Act of 1982, includ-
ing what Congress was trying to achieve with the Federal Circuit and how the court was
structured in terms of subject-matter jurisdiction and the courts it replaced.
Next, this chapter describes the problem of not having a uniform patent law and
how the Federal Circuit has impacted uniformity. It describes how forum shopping was
influenced and whether uniformity has affected the diversity and percolation of ideas
related to patent law.
This chapter then discusses how the Federal Circuit has impacted certainty and predict-
ability of patent law. It begins by looking at where the pressure for certainty comes from. It
then focuses on the Federal Circuit’s treatment of a few representative issues to determine
whether the court has succeeded in providing it.
Next, this chapter explores the quality of the Federal Circuit’s patent jurisprudence,
including whether the Federal Circuit is too formalistic in its decision-making and whether
it pays enough attention to policy and technological considerations and scholarship.
Given the focus on the Federal Circuit as an institution, this chapter describes how
the Federal Circuit is structured, the make-up of the individual judges on the court, and
how these qualities affect the court as such. The role of individual judges and the types of
judges on the Federal Circuit are examined, as well as how internal dynamics and en banc
hearings play a role in the decision-making process.
This chapter then explores the Federal Circuit’s interactions with other institutions and
examines which institutions are and should be performing what role in the development
of patent law. This section notes the interesting and dynamic interactions the Federal
Circuit has had with the Supreme Court, the United States Patent and Trademark Office
(USPTO), district courts, the Solicitor General, and Congress. Finally, this chapter closes
with concluding thoughts about empirical research on the Federal Circuit and suggests
further areas of exploration.
II. CREATION OF THE FEDERAL CIRCUIT
Before delving into the studies analyzing whether the Federal Circuit has been successful,
we must know what problems existed prior to the court’s creation and what Congress
hoped to achieve by establishing the Federal Circuit. This section begins by laying out the
concerns and inadequacies that existed and the failed attempts to resolve those problems.
It then turns to Congress’s creation of the Federal Circuit and describes what goals it
sought to achieve.

A. Concerns and Inadequacies
The antecedent to the creation of the Federal Circuit was the Senate’s Commission
on Revision of the Federal Court Appellate System in the early 1970s (Dunner, 2010,
p. 776). This commission, also known as the Hruska Commission, was formed to
evaluate the systemic strains posed on the federal appellate system and how to efficiently
and effectively resolve them (Commission on Revision of the Federal Court Appellate
System, 1975, p. 208). As part of this evaluation, several problems with respect to patent
litigation were brought to the attention of the commission, including great disparities
between the regional circuits’ treatments of patents; some were hostile to patents, while
others were friendlier (Commission on Revision of the Federal Court Appellate System,
1975, p. 220; Dunner, 2010, p. 777). As a result, patent litigators strategically raced to the
courthouses to make sure their clients’ interests would be protected on appeal (Dunner,
2010, p. 777). As some believed, forum shopping led to an unfair administration of justice
and created an unpredictable patent jurisprudence, which led to increasing difficulty in
advising technology developers and users (Commission on Revision of the Federal Court
Appellate System, 1975, p. 220; Dreyfuss, 1989, p. 7; Dunner, 2010, p. 777). The Hruska
Commission’s consultants conducted a survey of patent litigators to determine whether
it would be worthwhile to recommend the creation of a specialized appellate court for
patents, but the respondents were evenly divided (Dunner, 2010, p. 777).
The Hruska Commission concluded that it would not be desirable to have a special-
ized court of patent appeals to solve the problems of a disharmonious national law
(Commission on Revision of the Federal Court Appellate System, 1975, p. 234). Part of
the reason was a concern that specialized judges would suffer from tunnel vision and see
the cases without the influence of exposure to legal problems in other fields (Commission
on Revision of the Federal Court Appellate System, 1975, pp. 234–35). The other concerns
included: (1) judges imposing their own views of policy, (2) judges not writing thorough
and persuasive opinions, (3) depriving the other circuit courts of a broad exposure to
cases, and (4) finding capable judges who were not captured by particular interest groups
(Commission on Revision of the Federal Court Appellate System, 1975, p. 235; Dreyfuss,
1989, p. 25).
Instead, the Hruska Commission ultimately recommended creating a National Court
of Appeals, which would adjudicate issues of national law so as to provide consistency
and uniformity by avoiding and resolving circuit splits (Commission on Revision of the
Federal Court Appellate System, 1975, pp. 208–09). Although the National Court of
Appeals would hear a variety of types of cases, it was expected that this court would
appropriately monitor the patent decisions in the regional circuits and consequently
reduce the forum-shopping problem (Commission on Revision of the Federal Court
Appellate System, 1975, p. 236). Congress never adopted this recommendation.
B. A Court with a Mission
Although Congress and the Hruska Commission rejected the National Court of Appeals
and a court of patent appeals respectively, the Federal Circuit originated from these
proposals (Dreyfuss, 1989, p. 6). Over the next several years, Congress drafted bills and
held hearings to solve the problems of lack of uniformity in, and poor administration

of, the patent law (S. Rep. 97-275, 97th Cong., 1st Sess. 1981, 1982 U.S.C.C.A.N. 11;
H.R. Rep. 97-312, 97th Cong., 1st Sess. 1981). Around the same time, President Carter
convened a Domestic Policy Review, which was aimed at remedying the recession by
spurring technology-based innovation (Newman, 2002, pp. 541–42). A committee was
charged with examining the role of the patent system in supporting technological innova-
tion and concluded, in relevant part, that the judiciary was compromising innovation
and economic development (Newman, 2002, p. 542). The Congressional hearings and the
political need for increased innovation led to the creation of the Federal Circuit in 1982
(Federal Court Improvements Act of 1982, P.L. 97-164, April 2, 1982, 96 Stat. 25; Dunner,
2010, pp. 777–78; Gugliuzza, 2012, p. 1454).
The Federal Court Improvements Act of 1982 merged the Court of Customers and
Patent Appeals (CCPA) and the Court of Claims into the Federal Circuit, which would
hear, inter alia, patent appeals from the district courts and from the USPTO (Dreyfuss,
1989, pp. 3–4). Congress thought that creating the Federal Circuit and eliminating the
number of decision-makers in the patent system would largely eradicate the pervasive
inconsistency plaguing patent law and provide predictability to stakeholders through
uniformity (S. Rep. 97-275, 97th Cong., 1st Sess. 1981, 1982 U.S.C.C.A.N. 11; H.R. Rep.
97-312, 97th Cong., 1st Sess. 1981). Moreover, by forcing appellate jurisdiction into one
court, Congress believed this would reduce forum shopping (S. Rep. 97-275, 97th Cong.,
1st Sess. 1981, 1982 U.S.C.C.A.N. 11; H.R. Rep. 97-312, 97th Cong., 1st Sess. 1981). As
described below, much ink has been spilled analyzing whether and how Congressional
goals for the Federal Circuit have influenced the Federal Circuit’s decision-making.
But Congress was not just focused on defining the Federal Circuit’s mission. It was also
well aware of the potential problems of creating a “specialized” court (S. Rep. 97-275,
97th Cong., 1st Sess. 1981, 1982 U.S.C.C.A.N. 11). To avoid or minimize these problems,
Congress did not limit the Federal Circuit’s jurisdiction to patent appeals (S. Rep. 97-275,
97th Cong., 1st Sess. 1981, 1982 U.S.C.C.A.N. 11). Instead, Congress made sure the
Federal Circuit would hear a variety of types of cases (S. Rep. 97-275, 97th Cong., 1st
Sess. 1981, 1982 U.S.C.C.A.N. 11), including veterans’ benefits, government-employment
disputes, government contracts, customs and tariff issues, tax refunds, and trademark
registration appeals from the USPTO (Gugliuzza, 2012, p. 1453). Congress believed
this exposure to non-patent cases would keep judges from developing tunnel vision and
avoid special interest groups from dominating the court’s development (S. Rep. 97-275,
97th Cong., 1st Sess. 1981, 1982 U.S.C.C.A.N. 11). Like Congress’s goals for the Federal
Circuit, the accompanying concerns and the effectiveness of a diverse caseload have also
been studied in depth to determine whether the Federal Circuit experiment has succeeded.
III. UNIFORMITY
Before creation of the Federal Circuit, patent law differed throughout the country and
this lack of uniformity was of paramount concern. One cause of this lack of uniformity
flowed from the USPTO and CCPA interpreting and developing patent law with respect
to patentability, but not having the ability to bind the federal courts which regularly
resolved these issues in the context of infringement litigation (Dreyfuss, 1989, p. 6). The
other cause of the lack of uniformity was that there were pronounced differences amongst

the regional circuits’ understanding and application of patent law (Dreyfuss, 1989, p. 7;
Dreyfuss, 2010, p. 828). Statistics from the 1940s and 1950s showed that “a patent was
twice as likely to be held valid and infringed in the Fifth Circuit than in the Seventh
Circuit, and almost four times more likely to be enforced in the Seventh Circuit than in the
Second Circuit” (Dreyfuss, 1989, p. 7). As a result, forum shopping became widespread
(Dreyfuss, 1989, p. 7; Dreyfuss, 2010, p. 828).
In theory, the Supreme Court can serve as the tribunal that unifies patent law through-
out the nation (Dreyfuss, 1989, p. 2). But as a practical matter, this is impossible given
how few cases the Supreme Court hears each term (Dreyfuss, 1989, p. 2; Cotropia, 2010,
p. 804). Although the Supreme Court has taken much more interest in patent cases over
the last ten or 15 years, the total number of patent cases and issues it resolves is relatively
insignificant in the grand scheme of patent litigation (Dreyfuss, 2010, pp. 829–30, 839).
The Federal Circuit was created, in large part, to solve the lack of uniformity problems
plaguing patent law (Atkinson, Marco, and Turner, 2009, p. 411). In short, the different
interpretations of patent law by the regional circuits and the CCPA would cease to
exist once a single court was tasked with interpreting patent law for the entire country
(Dreyfuss, 2010, p. 828). This, in turn, would eliminate the rampant forum shopping
(Dreyfuss, 1989, p. 7).
Despite the perceived benefits provided by having the Federal Circuit serve as the
unifier of patent law, there were, and continue to be, risks associated with a single circuit
hearing all patent appeals. The major risk commentators have identified is a lack of
diversity in the development of patent jurisprudence (Nard and Duffy, 2007, p. 1623). The
diversity concern comes in two flavors—inter-circuit percolation and entrenchment. The
remainder of this section examines the Federal Circuit’s impact on forum shopping and
whether centralization has resulted in a dearth of diversity.
A. Forum Shopping
The Federal Circuit has been keenly aware of its mission to achieve uniformity. As
Howard Markey (1992), former Chief Judge of the Federal Circuit, noted, “[f]rom its very
first case, the Federal Circuit set out to meet Congress’ express intent that it contribute
to increased uniformity and reliability in the fields of national law assigned” (Markey,
1992, p. 577). Not only did Judge Markey believe the Federal Circuit had purposefully
undertaken this task, he believed the Federal Circuit had achieved it. In his words, “the
Federal Circuit met the desire of its congressional creators for increased uniformity and
elimination of forum shopping in its assigned areas of national law” (Markey, 1992,
p. 577). And to the extent forum shopping was a major problem in and of itself, Judge
Markey is correct that forum shopping has been completely resolved if the concern was
forum shopping based on appellate court jurisdiction (Dreyfuss, 2008, p. 788–89).1 By
1
Until passage of the America Invents Act (AIA), forum shopping at the appellate court level
was possible in limited circumstances because of Christianson v. Colt Industries Operating Corp.,
486 U.S. 800 (1988) and Holmes Group, Inc. v. Vornado Air Circulation Systems, Inc., 535 U.S. 826
(2002), which limited the Federal Circuit’s jurisdiction to those cases “arising under any Act of
Congress relating to patents” under the well-pleaded complaint rule. The AIA amended the Federal
Circuit’s jurisdiction so that it has exclusive jurisdiction “in any civil action arising under, or in

eliminating every other circuit’s appellate jurisdiction over patent cases, litigants have no
other appellate fora to shop for.
Despite this seemingly banal conclusion about uniformity and elimination of forum
shopping, there was little to no empirical evidence supporting the underlying assumption
that forum shopping was widespread and problematic. Congress’s belief that forum shop-
ping was a problem was based on the 240 survey responses conducted during the Hruska
Commission’s study (Atkinson, Marco, and Turner, 2009, pp. 414–15). To determine
whether forum shopping was prevalent, and to what extent, Atkinson, Marco, and Turner
(2009) performed an econometric study of non-uniformity and forum shopping before
and after the creation of the Federal Circuit.
In this study, they concluded that when the patentee was the plaintiff, significant
forum shopping existed before creation of the Federal Circuit (Atkinson, Marco, and
Turner, 2009, p. 441). This corresponded to a showing of significant non-uniformity in
validity outcomes across the circuits during that same time period (Atkinson, Marco, and
Turner, 2009, p. 441). In contrast, when the patentee was the defendant (i.e., a declaratory
judgment action), the authors found no evidence of systematic nonuniformity during the
pre-Federal Circuit period and no strong evidence of forum shopping (Atkinson, Marco,
and Turner, 2009, p. 441).
The authors also concluded that forum shopping during the period after creation of
the Federal Circuit had been mitigated (Atkinson, Marco, and Turner, 2009, p. 441).
This, unsurprisingly, corresponded to a finding that non-uniformity across district courts
within the regional circuits remains, but has been greatly reduced by the presence of the
Federal Circuit (Atkinson, Marco, and Turner, 2009, p. 441). Interestingly, the authors
estimate that systemic forum shopping ended by 1978, four years prior to creation of
the Federal Circuit (Atkinson, Marco, and Turner, 2009, pp. 412, 441). They suspect that
patentees saw the writing on the wall and anticipated the impact of the Federal Circuit
several years before its creation (Atkinson, Marco, and Turner, 2009, p. 413).
Even though the extent of forum shopping may have been exaggerated and the Federal
Circuit seems to have eliminated forum shopping based on appellate court jurisdiction,
this does not necessarily mean the forum shopping problem has been solved. Instead,
studies suggest that forum shopping has remained prevalent and the focus has simply
shifted to the district court level (Moore, 2001, p. 892; Sichelman, 2010, p. 1165). For
example, then-Professor Moore (2001) conducted an empirical study of patent litigation
in district courts after creation of the Federal Circuit. Examining the population of patent
cases that went to trial from 1983–99 and all patent cases that were terminated by any
means from 1995–99, she demonstrates that patent litigation at the district court level
was not evenly dispersed throughout the country; nor was it dispersed according to the
district courts’ civil docket size (Moore, 2001, pp. 892, 903–04). Instead, patent litigation
was concentrated in a few select jurisdictions. (Moore, 2001, pp. 892, 903–04).2 Jeanne
Fromer’s (2010) study of forum selection refines Moore’s analysis by analyzing forum
any civil action in which a party has asserted a compulsory counterclaim arising under, any Act of
Congress relating to patents . . .” (28 U.S.C. § 1295(a)(1), 2012).
2
These jurisdictions have varied somewhat over time. In Moore’s (2001) study, the Eastern
District of Texas was not a favorite district; but as Sichelman (2010) notes, “[o]ver the last decade
plaintiffs have increasingly chosen the remote Eastern District of Texas with 860 percent more

selection vis-à-vis the underlying technology (Fromer, 2010, pp. 1499, 1502, 1512, 1514,
1516–20). Moore’s results also illustrate that district courts have procedural and substan-
tive variation in adjudicating patent cases and, although it is a multi-dimensional inquiry,
this creates an environment hospitable to forum shopping (Moore, 2001, pp. 892, 907–19).
At the end of the day, Moore concludes that the differences in win rates in infringement
and declaratory judgment suits are likely attributable to the choice of forum (Moore,
2001, p. 923).
And though one may posit that these differences between districts will be resolved at
the Federal Circuit on appeal and uniformity will ensue, most patent cases settle before
judgment and before appeal to the Federal Circuit (Sichelman, 2010, p. 1165; Gruner, 2010,
pp. 1024, 1029–33). Until the Federal Circuit makes a broad pronouncement to resolve
district court variances, it is thought that parties will settle based on forum considerations
rather than just how a uniform law applies to the facts of the case (Sichelman, 2010, p. 1169).
Although the Federal Circuit may have been created, in part, to eliminate forum
shopping, its success is certainly questionable. Scholars have declared forum shopping
to be “alive and well” (Moore, 2001, p. 937), and that it “remains a pernicious feature of
the patent litigation landscape” (Sichelman, 2010, p. 1171). Although there is hope that
the Federal Circuit will exercise more oversight regarding motions to transfer venue and
that this will reduce forum shopping, it is a pessimistic case for hope—significant forum
shopping in patent litigation is likely unavoidable (Sichelman, 2010, p. 1171).
B. Diversity and Percolation
As noted above, a major risk of centralization is a dearth of diversity in patent jurispru-

dence due to a lack of inter-circuit percolation and because of entrenchment or path
dependency—that is, the risk is that the Federal Circuit has become too successful in
producing uniformity (Cotropia, 2010, p. 806).
Percolation “is based on the notion that law evolves through interchange among
the regional circuits and experimentation within their territories [and that e]ventually,
experience demonstrates which rules work best” (Dreyfuss, 2010, p. 829). It was thought
that centralizing all patent appeals in the Federal Circuit would make percolation impos-
sible and the best rules would not necessarily come to fruition (Nard and Duffy, 2007,
pp. 1627–37; Dreyfuss, 2010, p. 829).
Entrenchment, or path dependency, is related to percolation. But entrenchment is
concerned with later panels of the Federal Circuit being required to follow the precedent
of prior panel decisions (Nard and Duffy, 2007, pp. 1645–46; Petherbridge, 2009, p. 424).
Because of court rules binding subsequent panels, the concern is that attorneys are less
likely to challenge precedent and subsequent panels are less likely to revisit and reconsider
existing rules (Nard and Duffy, 2007, pp.1645–46; Petherbridge, 2009, p. 424). As a
result, the Federal Circuit’s patent jurisprudence becomes homogenous and irretrievably
entrenched, and “compromises the efficient evolution of the law through common law
mechanisms” (Petherbridge, 2009, p. 425).
cases filed there in 2009 than in 2000, while the background growth of total cases filed nationwide
was only 15 percent.”

Professors Nard and Duffy (2007), in a thought-provoking article, suggested that the
Federal Circuit be stripped of its exclusive appellate jurisdiction over patent appeals, that
at least one additional regional circuit be given jurisdiction to hear patent appeals from
district courts, and that the DC Circuit also be given jurisdiction over appeals from the
USPTO (Nard and Duffy, 2007, pp. 1623–25).3 They argued that adding a limited number
of circuit courts to hear patent appeals would inject “an ongoing, lively jurisprudential
debate at the circuit court level” (Nard and Duffy, 2007, pp. 1623–24). This diversity of
views, they argue, enhances competition between circuits, but also sparks more creative
lawyering, both of which lead to better incremental innovation in patent law—that is, the
law is allowed to percolate (Nard and Duffy, 2007, pp. 1623–24, 1633). As to entrench-
ment, they argue that the antidote is having more than one circuit hearing appeals so that
when circuit splits arise, lawyers and the Supreme Court can step up to develop the best
arguments and resolve the circuit split (Nard and Duffy, 2007, pp. 1645–46). Judge Diane
Wood (2014) of the Seventh Circuit has recently echoed these sentiments and suggested a
variation of Nard and Duffy’s (2007) proposal (Wood, 2014, pp. 9–10).
The claim that the Federal Circuit yields a lack of diversity has been disputed and
the proposed solution has been critiqued (Plager and Pettigrew, 2007; Golden, 2009;
Petherbridge, 2009; Cotropia, 2010). For example, Cotropia (2010) asserts that creation of
the Federal Circuit has not resulted in stagnant decision-making or impeded the common
law development of patent law (Cotropia, 2010, p. 803). In support of this claim, Cotropia
(2010) measured diversity by comparing rates of dissents and en banc review in the
Federal Circuit with those of selected regional circuits (Cotropia, 2010, p. 803).
Looking at the proportion of dissents to written decisions is useful because, it is argued,
the dissents are usually disagreements about the law or application of the law, not about
facts (Cotropia, 2010, p. 810). As a result, dissents show that a court is not of one mind
(Cotropia, 2010, p. 810). Dissents also signal and communicate new ideas or approaches
to those outside of the specific panel—future litigants, Congress, future panels, and the
court sitting en banc (Cotropia, 2010, p. 810).
The results of this study show that the Federal Circuit has a dissent in 3.51 percent of
its opinions (Cotropia, 2010, p. 815). The only selected circuit with a higher dissent rate
was the Ninth Circuit at 4.56 percent (Cotropia, 2010, p. 815). When limiting the Federal
Circuit opinions to patent cases, the dissent rate climbed to 9.28 percent (Cotropia, 2010,
p. 816). Even when taking a circuit’s total caseload, caseload per judge, and outlier judges
within the Federal Circuit into account, the Federal Circuit’s comparatively high dissent
rate persisted (Cotropia, 2010, pp. 818–20). A similar analysis undertaken by Petherbridge
(2009) focuses on dissenting and concurring opinions in the areas of the doctrine of
equivalents, claim construction, and obviousness (Petherbridge, 2009, p. 455). This study
shows a substantial rate of alternative writings (Petherbridge, 2009, p. 456). These results
challenge the belief that there is a lack of diversity within the Federal Circuit; there
appears to be a fair amount of discord within the court (Petherbridge, 2009, pp. 456–57;
Cotropia, 2010, p. 818). From these analyses, it has been reasoned that the high rate of
3
Nard and Duffy (2007) recognize that such a solution has the potential to exacerbate the
forum shopping problem Congress tried to address. To minimize forum shopping, they suggest
randomizing assignment of the appellate court (Nard and Duffy, 2007, pp. 1668–69).

disagreement at the Federal Circuit flows from repeated exposure to a small set of subject
matters, and that over time, the judges develop deeper and more nuanced perspectives on
those subject matters (Cotropia, 2010, p. 820).
With respect to examining the proportion of en banc cases to written decisions, it was
reasoned that dissenting opinions were a significant trigger for en banc review, and that if
there was a high degree of uniformity on an issue, then there would be little intra-circuit
conflict and dissent and hence, no need for en banc review (Cotropia, 2010, pp. 813–14).
Likewise, en banc review may suggest “how willing a court is to innovate new legal rules,
percolate these concepts, and then eventually adopt them in future decisions” (Cotropia,
2010, p. 822). Cotropia’s results show that the Federal Circuit decided cases en banc 0.18
percent of the time (Cotropia, 2010, p. 817). This was the lowest of the selected circuits,
but not much lower than the next two lowest regional circuits (Cotropia, 2010, p. 817).
Unlike Cotropia’s (2010) study, which examined all en banc decisions from 1998–2009 and
included all subject matters as a function of the total number of opinions, Vacca (2011)
measured only en banc patent cases as a function of the court’s total patent cases. Vacca
(2011) shows that from 2001–09, the Federal Circuit decided patent cases en banc 0.23
percent of the time (Vacca, 2011, p. 738). When compared to the general en banc rates
of all other circuits, these results show that the Federal Circuit had the highest rate of en
banc decision-making (Vacca, 2011, p. 738).
In explaining how these results explain or impact percolation and entrenchment,
Cotropia (2010) offers a few explanations. One is simply that although there is not
uniformity within the court, the law remains entrenched (Cotropia, 2010, p. 822). Another
explanation could be that the disputes provoking dissenting opinions are minor and
not sufficiently important for en banc review (Cotropia, 2010, pp. 822–23). The final
explanation is that the Federal Circuit could be harmonizing the law without using the
transparent en banc procedure, but instead using the eight-day comment period on prec-
edential opinions to resolve potential conflicts (Cotropia, 2010, p. 823). As a result, the
law does not become entrenched; instead it evolves, but does so in a less transparent way
than en banc review provides (Cotropia, 2010, p. 823). In contrast, Vacca (2011) shows
that the Federal Circuit’s high en banc rate and practices are illustrative of an exceedingly
transparent form of jurisprudential development (Vacca, 2011, pp. 757–58).
Assuming Cotropia’s dissent and en banc rates are accurate and his comparisons to
other selected circuits are correct, there has been disagreement with his conclusions that
the Federal Circuit’s jurisprudential diversity is no worse than the other circuits, and
that it may have more diversity given the frequent dissents (Dreyfuss, 2010, pp. 832–34).
First, a simple comparison of dissent and en banc rates does not take into account the
traditions and heritage of each circuit; nor does it consider that different areas of the
law develop differently (Dreyfuss, 2010, p. 832). Second, merely comparing the rates of
dissent and en banc decision-making does not tell us whether this compensates for a lack
of inter-circuit percolation (Dreyfuss, 2010, p. 832). To be meaningful and answer these
questions, the substance of those dissents and en banc issues must be examined (Dreyfuss,
2010, p. 832). If the topics of dissent are insubstantial and not related to key issues suf-
fering from ossification, then the high degree of diversity based on dissent rates does
not adequately address the entrenchment concern (Dreyfuss, 2010, p. 832). Finally, the
concern about a lack of diversity is not just that the law becomes entrenched, but that the
law is suboptimal—“that the ‘group’ is not engaged in serious thinking at all” (Dreyfuss,

2010, p. 833). It could be that the large high and low en banc rates show that the judges
regularly disagree, but “fail to frame their disagreements in ways that make en banc review
fruitful enough to identify optimal rules on which to converge” (Dreyfuss, 2010, p. 834).
Of course, Vacca’s (2011) study showing the Federal Circuit has a high en banc rate for
patent cases suggests this may not be the case (Vacca, 2011, p. 738).
In another challenge to the claim that the Federal Circuit has yielded too much
uniformity, Petherbridge (2009) demonstrates that, at least with respect to the doctrine
of equivalents, the Federal Circuit does not lack jurisprudential diversity (Petherbridge,
2009, p. 427). Using a dataset consisting of all Federal Circuit doctrine of equivalents
analyses from 1992–2007, Petherbridge (2009) shows that the major response variables—
(1) patentee success at the Federal Circuit on a doctrine of equivalents issue, (2) whether
that ruling was dispositive of the case in favor of the patentee, and (3) whether the Federal
Circuit affirmed the lower tribunal’s decision—varied greatly between each other at any
particular moment and greatly shifted over time (Petherbridge, 2009, p. 442). Looking at
the average rates of success for two of the response variables in this study shows that there
is great volatility between the first and second half of the dataset (Petherbridge, 2009,
pp. 442–43). Looking closer at the specific issues within the doctrine of equivalents yields
similar results—different components of the doctrine have been on the rise, while others
have declined (Petherbridge, 2009, p. 459).
When examined as a whole, the results do not suggest a stagnating jurisprudence
(Petherbridge, 2009, p. 459). Based on these results, the study suggests that the law con-
cerning the doctrine of equivalents is fairly dynamic and “is becoming more uncertain and
unpredictable than uniform” (Petherbridge, 2009, p. 444). Having a wide degree of uncer-
tainty and hence little uniformity suggests a lack of entrenchment, but does not directly
address whether ideas percolate (Dreyfuss, 2010, p. 833). Nonetheless, the study concludes
that the results provide evidence contradicting the notion that the Federal Circuit’s patent
jurisprudence is uniform and entrenched (Petherbridge, 2009, p. 457). Nor, it concludes,
has the Federal Circuit’s decisions become path-dependent so that debate is foreclosed,
a variety of views are suppressed, and refinements through traditional common law pro-
cesses are retarded (Petherbridge, 2009, p. 457). These conclusions are modest; the study
makes clear that it cannot address the question of whether the Federal Circuit develops
its patent jurisprudence at an optimal rate (Petherbridge, 2009, p. 464).
Recall that Nard and Duffy’s (2007) proposed solution to the lack of diversity in
patent jurisprudence was to give at least two additional circuits jurisdiction over patent
appeals (Nard and Duffy, 2007, pp. 1623–25). Like their predicate belief that there is a
lack of diversity, the proposed solution too has been challenged (Plager and Pettigrew,
2007; Golden, 2009, p. 661). One objection questions whether Nard and Duffy’s proposed
solution begets better results (Plager and Pettigrew, 2007, p. 1756). That is, has it been
adequately proven that multiple circuits disagreeing about what the law is will provide
better decisions than the Federal Circuit currently provides (Plager and Pettigrew, 2007,
p. 1756)?
The other objection concerns not whether another tribunal should play a role in the
percolation process, but which tribunal should play that role (Golden, 2009). Rather than
sacrificing the benefits of a single appellate court for an unknown benefit of percolation,
an alternative solution is to urge the Supreme Court and Federal Circuit to work together
to ensure the percolation of patent law (Golden, 2009, p. 662). The suggestion is to have

the Supreme Court view itself as the “prime percolator” of patent law rather than the
“final law sayer” (Golden, 2009, p. 662). Specifically, the Supreme Court would hear
patent cases “to combat undesirable ossification of legal doctrine” and do so in an effort
to spur subsequent development of the patent law (Golden, 2009, p. 662). This would
occur by the Supreme Court only providing partial direction in its opinions and letting the
lower courts and USPTO flesh out the details and allow for continued growth (Golden,
2009, p. 663).
In conjunction with the Supreme Court taking this role, the Federal Circuit itself could
undertake responsibility to foster percolation (Golden, 2009, p. 717). First, the Federal
Circuit should hesitate to write or read opinions unnecessarily broadly (Golden, 2009,
p. 717). This would limit entrenchment and allow subsequent panels to develop more
nuanced and considered interpretations of the law (Golden, 2009, p. 717). Second, the
Federal Circuit could promote percolation by hearing cases en banc when there is a
risk of entrenchment on a substantive question and there is reason to believe the settled
approach is substantially inferior to the alternative approach (Golden, 2009, p. 717).
Given the Supreme Court’s revived interest in patent law (Dreyfuss, 2010, pp. 829–30) and
the Federal Circuit’s high en banc rate in patent cases (Vacca, 2011, p. 738), it might be that
this solution is being implemented. Another suggestion to assist in the percolation process
is to use advisory panels (Pedraza-Fariña, 2015, pp. 154–59). These advisory panels, with
a mixture of technological, economic, and sociological expertise, could advise the Federal
Circuit on areas of the law in need of clarification and provide input on how a decision
may impact innovation in particular fields (Pedraza-Fariña, 2015, pp. 154–55). Further
research is warranted.
As noted in the beginning of this section, the Federal Circuit’s creation spawned concern
that there would be a lack of diversity in the development of patent jurisprudence (Nard
and Duffy, 2007, p. 1623). Much ink has been spilled over whether this dearth of diversity
exists and what, if anything, should be done to compensate for it. Novel approaches to
evaluating the issues have been suggested and implemented, and although we have more
data to analyze based on over 30 years of experimentation at the Federal Circuit, the role
and impact of uniformity are still ripe for exploration.
IV. CERTAINTY AND PREDICTABILITY
Closely related to, and sometimes intertwined with, uniformity is a concern over certainty.
Certainty comes in two varieties. The first is certainty with respect to having clear rules
as opposed to vague standards. The second variety of certainty, sometimes referred to
as panel-dependency, is consistently applying the same legal rule or standard to a given
scenario. Despite the differences in the types of certainty, both are aimed at the same
goal—predictability—being able to accurately forecast the outcome of application of the
law to a new set of facts (Michel, 1994, p. 1233). Predictability, of course, was one of the
major problems Congress sought to overcome by creation of the Federal Circuit, although
the proposed route to solving this problem was uniformity (S. Rep. 97-275, 97th Cong., 1st
Sess. 1981, 1982 U.S.C.C.A.N. 11; H.R. Rep. 97-312, 97th Cong., 1st Sess. 1981; Wagner
and Petherbridge, 2004, p. 1108). As it turns out, uncertainty can exist despite uniformity.
This section begins by exploring the pressures the Federal Circuit faces to foster certainty

and uncertainty. Then, this section examines reversal rates and indeterminacy involving
validity, claim construction, and infringement to see whether patent law and litigation
have become more or less certain under the Federal Circuit’s stewardship.
A. Pressure for and Undermining Certainty
The benefits of certainty and the risks of uncertainty in patent law abound. Uncertainty
not only harms individual litigants coming before the court, but also negatively affects
competitors, potential licensees, follow-on inventors, and the consuming public (Michel,
1994, pp. 1234, 1241; Mullally, 2010, pp. 1112–14). For litigants, uncertainty increases the
costs of litigation (Michel, 1994, p. 1234; Mullally, 2010, pp. 1112–13). Certainty gives the
parties the ability to decide their futures, and as argued by Paul Michel (1994), former
Chief Judge of the Federal Circuit, parties resolving their disputes is superior to judges
doing so and would result in fewer lawsuits altogether (Michel, 1994, p. 1234). More
generally, uncertainty as to patentability may cause inventors and businesses to elect for
trade secrecy rather than patent protection or may even result in failure to engage in the
relevant research and development (Dreyfuss, 1989, p. 7; Michel, 1994, p. 1241; Mullally,
2010, pp. 1112–13). As a result, the benefit of innovation and disclosure, along with the
foundation for further innovation, may be lost (Michel, 1994, p. 1241; Mullally, 2010,
pp. 1112–13). Uncertainty regarding the patent’s validity can also result in the patentee
being unable to obtain financing to bring its innovation to market (Mullally, 2010,
pp. 1112–13). For competitors, and consequently the consuming public, uncertainty
with respect to claim construction and the doctrine of equivalents plagues their ability
to effectively compete by improving the patented technology (Michel, 1994, p. 1241;
Mullally, 2010, pp. 1113–14).
The pressure for certainty and uncertainty in patent law stems from a variety of
sources. One such source is the patent bar (Thomas, 2003, p. 794). The patent bar was
outspoken before creation of the Federal Circuit about the need for certainty and has
continued down the path by seeking more rules and fewer standards from the Federal
Circuit (Thomas, 2003, p. 794). Despite this claim, it has been suggested that the patent
bar could be the source of pressure for uncertainty in an attempt to enhance its value
(Mullally, 2010, pp. 1142–45).
Another source of pressure comes from district courts (Thomas, 2003, p. 795). District
court judges’ general inexperience with cutting-edge technology and patent law, coupled
with complaints about the Federal Circuit’s high rates of reversal, have caused them to
seek more certain rules to apply in their cases to help compensate for this lack of expertise
(Thomas, 2003, p. 795; Mullally, 2010, pp. 1128–30; O’Malley, Saris, and Whyte, 2004,
p. 682; Lee, 2010, pp. 9–20). The complexity of patent law itself also contributes to this
problem (Thomas, 2003, pp. 794–95). It has been suggested that the Federal Circuit has
tried to provide clear rules in an attempt to “provid[e] a well-meaning judiciary with a
thread through the labyrinth” (Thomas, 2003, pp. 794–95).
Another source of pressure for increased certainty flows from the USPTO (Thomas,
2003, pp. 795–96; Mullally, 2010, pp. 1126–28). The USPTO, because of its structure and
employment of technical agents, many of whom are not lawyers, must synthesize Federal
Circuit case law in a way that is suitable for examiners to implement at the prosecution
stage (Mullally, 2010, pp. 1126–28). As a result, there is pressure from the USPTO for

the Federal Circuit to provide clear rules that can easily be translated into the Manual of
Patent Examining Procedure and guidelines used by patent examiners (Mullally, 2010,
pp. 1126–28). But this is not always the case; sometimes the USPTO has openly criticized
the Federal Circuit for not allowing the USPTO to exercise its discretion (Nard, 1995,
pp. 1421–22).
Despite these pressures for increased certainty, the Supreme Court, as of late, has been a
consistent counter-pressure, encouraging or mandating that the Federal Circuit take a less
rule-like approach and adopt more flexible standards (Mullally, 2010, p. 1130; Pedraza-
Fariña, 2015, pp. 105–07). Recent examples include eBay Inc. v. MercExchange, L.L.C.,
547 U.S. 388 (2006), Festo Corp. v. Shoketsu Kinzoku Kogyu Kabushiki Co., Ltd., 535 U.S.
722 (2002), KSR Int’l Co. v. Teleflex Inc., 550 U.S. 398 (2007), and Bilski v. Kappos, 561
U.S. 593 (2010). In sum, the Federal Circuit has strived to provide predictability, but is
caught between competing stakeholders and actors within the patent ecosystem that exert
upward and downward pressure for more or less certainty (Mullally, 2010, p. 1126). To
what extent the Federal Circuit has provided certainty is a separate question to which we
now turn.
B. Reversal Rates, Panel Dependency, and Indeterminacy
An oft-examined method of measuring uncertainty caused by the Federal Circuit is to

observe and compare the Federal Circuit’s reversal rates over various aspects of district
court litigation and how the make-up of the panel affects outcomes. Patent litigation
before the Federal Circuit has been analogized to deciding the case by coin flip or dice roll
(Hill and Cote, 2002, p. 2). The argument is that as the reversal rate increases, this makes
it increasingly difficult for the parties and lower courts to have much confidence in the
accuracy of their decision-making process (Mullally, 2010, pp. 1113–14; O’Malley, Saris,
and Whyte, 2004, p. 682). To the extent outcomes are based on the make-up of the panel,
this also undermines certainty and predictability as the panels are not known until just
before the case is argued (Wagner and Petherbridge, 2004, p. 1174). Beyond the quantita-
tive analyses of reversal rates and panel dependence, another source of uncertainty is the
indeterminacy of the law because of the use of vague rules or standards as opposed to
clearly defined rules. The following subsections explore how reversal rates, panel depend-
ence, and unclear rules affect the most frequently examined areas of patent law: validity,
claim construction, and infringement.
1. Validity
In a study claiming to be the first to examine data from decisions from both before and
after creation of the Federal Circuit, Henry and Turner (2006) constructed a dataset of
approximately 5000 patent cases from 1953–2002. In this study, the authors look at the
affirmation rate of invalidity decisions in the pre-Federal Circuit era and Federal Circuit
era (Henry and Turner, 2006, pp. 100–01). During the Federal Circuit era, the affirma-
tion rate for invalidity was .573, nearly .28 points less than the pre-Federal Circuit era
affirmation rate of .850 (Henry and Turner, 2006, p. 100). As a result, they found the
Federal Circuit to be nearly three times more likely to reverse an invalidity decision than
the regional circuits were (Henry and Turner, 2006, p. 100). Yet when a district court held
the patent to be valid and infringed, the Federal Circuit’s affirmation rate was .723, .12

points higher than the pre-Federal Circuit era (Henry and Turner, 2006, pp. 100–03). The
authors noted that several factors complicate their interpretation of these descriptive
statistics and advised using a time-series analysis (Henry and Turner, 2006, p. 103). These
results illustrate that the Federal Circuit has been three times more likely to reverse an
invalidity decision (Henry and Turner, 2006, p. 108). Unsurprisingly, patentees have
appealed invalidity findings more frequently since the Federal Circuit’s creation (Henry
and Turner, 2006, p. 112). With respect to district court decisions, where the patent was
held to be valid and infringed, the authors concluded that there was a short-lived initial
surge in the affirmation rate, but this rate returned to the pre-Federal Circuit era rate in
the 1990s (Henry and Turner, 2006, p. 112).
Although the Henry and Turner (2006) study provides interesting general results, it is
not terribly specific with respect to the grounds of invalidity. There could, in fact, be vari-
ation within the Federal Circuit on the various requirements for a valid patent (Barnes,
2013, p. 973). Barnes (2013) conducted a study that examined, inter alia, reversal rates
for invalidity based on Section 112’s enablement and written description requirements
(Barnes, 2013, p. 983). The results show that the Federal Circuit reversed on enablement
grounds 28.1 percent of the time, reversed on written description grounds 32.1 percent of
the time, and had an overall reversal rate of 32.9 percent when combined (Barnes, 2013,
pp. 991–992). When viewed as a three-year moving average from 1997–2011, there was
substantial variation in reversal rates; from 2006–08, the reversal rate for written descrip-
tion and enablement was at its lowest point of 18.9 percent, but from 2000–02, the rate was
at a high of 46 percent (Barnes, 2013, pp. 1004–1006). One major limitation of Barnes’s
(2013) study was that it excluded Rule 36 cases (Barnes, 2013, p. 989). If these cases were
included, the reversal rates would necessarily drop (Barnes, 2013, p. 989).
In comparison, Sichelman (2010) analyzed reversal rates from 2000–07 on several
validity issues and included Rule 36 affirmances (Sichelman, 2010, pp. 1174–75). These
results show reversal rates of approximately 10–15 percent for written description and
enablement (Sichelman, 2010, p. 1175). In comparison to other invalidity issues, such as
anticipation, indefiniteness, obviousness, and statutory bars, the reversal rates for written
description and enablement are relatively low (Sichelman, 2010, p. 1175). These data
are further broken down to see whether the reversal stems from the patentee or accused
infringer prevailing at the district court (Sichelman, 2010, pp. 1177–78). For some validity
issues, the disparity in reversal rates is substantial, such as for on-sale bar issues, where the
reversal rate is 30 percent for prevailing patentees and only 9 percent for prevailing accused
infringers (Sichelman, 2010, p. 1178). But for some areas, the disparity is negligible, such
as the public use bar, which is 19 percent and 20 percent, respectively (Sichelman, 2010,
p. 1178). Although there may be low reversal rates for one type of litigant versus another,
the contrasting high reversal rates still inject uncertainty and unpredictability into the
process.
Of course, appeals to the Federal Circuit come from district courts and the USPTO.
There could be some difference in reversal rates based on where the appeal stems from.
Barnes’s (2013) study breaks the data down into appeals from district courts and from
the Board of Patent Appeals and Interferences (BPAI), and shows that for Section 112
reversals, the reversal rate for district courts was 38 percent, but only 17.5 percent for
the BPAI (Barnes, 2013, pp. 993–94). Based on these results, one could conclude that the
Federal Circuit is uncertain when it comes to appeals from district courts, but relatively

certain when it comes to appeals from the BPAI. But as Barnes (2013) notes, selection
bias could play a larger role in explaining what BPAI decisions are appealed as opposed
to district court decisions (Barnes, 2013, p. 994).
Despite the results of these studies involving reversal rates, we should be hesitant to
blindly follow them to the conclusion that this proves a high level of uncertainty and
unpredictability. High reversal rates could be a result of extremely certain doctrine;
specifically, parties may be more likely to appeal a particular issue when the correct result
is certain and the lower tribunal clearly erred (Rantanen and Petherbridge, 2015, p. 2013).
In such situations, we would expect the reversal rates to be high, but the cause of the high
reversal rate is inversely proportional to uncertainty (Rantanen and Petherbridge, 2015,
p. 2013).
Panel dependence is another source of uncertainty and unpredictability. As Judge
Paul Michel (1994) has noted, although there will always be some circumstances where
outcomes are unpredictable, the worst scenario is when the legal borderline shifts based
on the composition of the appellate panel (Michel, 1994, p. 1243).
In an early study, John Allison and Mark Lemley (2000) studied patent validity deci-
sions from 1989–96 (Allison and Lemley, 2000, pp. 745–46). Despite the firm belief of
patent litigators that validity outcomes are panel-dependent, the results of this study did
not provide strong empirical evidence for such a belief (Allison and Lemley, 2000, pp. 746,
753). After noting several limitations of the study, including subject-matter bias by the
exclusion of appeals from the USPTO, the exclusion of remands, an inability to account
for multiple causative case-specific facts, and the exclusion of Rule 36 affirmances, the
authors described their results (Allison and Lemley, 2000, pp. 748–49). Their results show
that “validity votes were remarkably consistent” (Allison and Lemley, 2000, p. 754).
There did not seem to be any substantial differences between judges appointed before
and after creation of the Federal Circuit (52.5 percent versus 55 percent), or between those
with and without prior patent experience (57.9 percent versus 52.1 percent) (Allison and
Lemley, 2000, p. 754). Using the Chi-squared test to apply their population data to the
superpopulation, the authors came to similar conclusions: they could not confidently
predict that the time of appointment or prior patent experience influenced validity voting
(Allison and Lemley, 2000, pp. 754–55). When examining individual judges, the authors
concluded that “the overwhelming majority of [Federal Circuit] judges show no leaning
whatsoever for or against the validity of the patent” (Allison and Lemley, 2000, p. 757).
When looking at affirmance rates (rather than validity and invalidity rates), Allison
and Lemley (2000) similarly concluded that the time of appointment and prior patent
experience could not confidently be used to predict reversal rates for invalidity (Allison
and Lemley, 2000, pp. 758–59). To be sure, there were some differences between judges
with and without patent experience, but these differences were rather modest (Allison and
Lemley, 2000, p. 759). When examining individual judges, most were close to the overall
affirmation rate and even the outliers were close to the norm (Allison and Lemley, 2000,
p. 759). Using the Chi-squared test to apply their population data to the superpopulation,
the authors could only weakly reject the hypothesis that “there is no difference in the
likelihood that [the various judges of the Federal Circuit] will vote to affirm the district
court” (with a p-value of 0.069) (Allison and Lemley, 2000, pp. 759–60).
Finally, Allison and Lemley (2000) briefly explored specific validity issues and whether
the outcomes were dependent on the judge (Allison and Lemley, 2000, p. 764). For best

mode, judges with prior patent experience rejected best mode invalidity challenges 74.1
percent of the time, whereas judges without prior patent experience rejected such chal-
lenges 61.5 percent of the time (Allison and Lemley, 2000, p. 764). They concluded that
this difference was modest (Allison and Lemley, 2000, p. 764). A similar conclusion was
drawn with respect to obviousness challenges (Allison and Lemley, 2000, pp. 764–65). For
Section 102 prior art challenges, judges appointed prior to the creation of the Federal
Circuit rejected these challenges 87 percent of the time, whereas judges appointed after
the creation of the Federal Circuit rejected these challenges only 50 percent of the time
(Allison and Lemley, 2000, p. 764). This result was surprising and it is unclear why such
a difference exists (Allison and Lemley, 2000, p. 764). In sum, Allison and Lemley (2000)
concluded that, with respect to validity, there is little variation between the judges on the
Federal Circuit, and that validity cases “depend[] on the facts of the case, and not on the
composition of the panel” (Allison and Lemley, 2000, p. 766). Because this study is now
19 years old and many limitations existed, the time may be ripe for an updated study on
panel dependency and patent validity.
Turning now to how vague rules or the use of standards affect certainty and predictabil-
ity with respect to validity issues, Paul Michel (1994), before becoming Chief Judge, argued
that the Federal Circuit had failed to increase predictability (Michel, 1994, p. 1233). Judge
Michel (1994) argued that the Supreme Court’s four-part test from Graham v. John Deere
Co., 383 U.S. 1 (1966) bewilders judges, and that determining whether the person having
ordinary skill in the art could have invented without undue experimentation becomes
subjective and elusive (Michel, 1994, pp. 1240–41). After initially using the CCPA’s test
from In re Winslow, 365 F.2d 1017 (Fed. Cir. 1966), the Federal Circuit acknowledged the
limitations of the Winslow (1966) tableau and, in terms of certainty, seemed to “[leave]
too much discretion to the decision maker, [and] rendered the results of the nonobvious-
ness uncertain” (Thomas, 2003, pp. 790–91). As a result, the Federal Circuit began to
stress the importance of the prior art references teaching, motivating, or suggesting their
combination (Thomas, 2003, p. 791; Petherbridge, 2010, pp. 910–11). Judge Michel’s
comments concerning the uncertainty of obviousness were during the Federal Circuit’s
teaching, suggestion or motivation (TSM) period and before the Supreme Court’s KSR
(2007) decision. KSR (2007) relaxed the standard for proving obviousness by rejecting the
rigid TSM test as the sole test for obviousness (KSR, 2007). If certainty and predictability
were difficult under the Federal Circuit’s TSM test, KSR’s (2007) more fluid approach to
obviousness introduced additional uncertainty to the mix.
Similarly, John Thomas (2003) reviewed the developments of other patent validity
doctrines and concluded that the Federal Circuit has been moving from standards to
rules (Thomas, 2003, pp. 773–74). As Thomas (2003) argued, the origin of this shift from
standards to rules is Pfaff v. Wells Electronics, 124 F.3d 1429 (Fed. Cir. 1997). Prior to
Pfaff (1997), the Federal Circuit used a totality of the circumstances test for the on-sale
bar trigger (Thomas, 2003, pp. 778–79). The Supreme Court rejected this test in Pfaff
v. Wells Electronics, 525 U.S. 55 (1998) and, Thomas (2003) argues, caused the Federal
Circuit to take a more rule-oriented approach (Thomas, 2003, p. 779).
We have seen similar movements toward clear and definite rules in determining
patentable subject matter. The judge-made exceptions to patentable subject matter
muddied the waters of the broad categories capable of being patentable (Thomas, 2003,
pp. 786–87). The Federal Circuit’s decision in State Street Bank & Trust Co. v. Signature

Fin. Group, Inc., 149 F.3d 1368 (Fed Cir. 1998) was a move toward simplicity (Thomas,
2003, pp. 787–88). The existing exceptions had formed a “Gordian knot” and the Federal
Circuit’s decision in State Street severed, rather than untied, this knot and adopted a test
that simply required that the process produce a “useful, concrete, and tangible result”
(Thomas, 2003, pp. 788–89). Later, the Federal Circuit, sitting en banc in In re Bilski,
545 F.3d 943 (Fed. Cir. 2008), addressed the patentable subject-matter test for processes.
In Bilski (2008), the court rejected State Street’s (1998) test, but replaced it with the
“machine or transformation” test (In re Bilski, 2008). This seemed to be a limited retreat
from certainty, but the Federal Circuit did limit this test as the only one to be used (In re
Bilski, 2008). But again, as with obviousness, the Supreme Court rejected this approach
in favor of a more flexible standard (Bilski, 2010). This lack of clear rules has continued
in the rest of patentable subject-matter jurisprudence, and has “created more legal conflict
than [it] resolve[d]” (Rantanen and Petherbridge, 2015, pp. 2027–28).
Regardless of whether it is because of or despite the Supreme Court’s interventions in
cases such as Pfaff (1998) and KSR (2007), there is a strong basis for believing that the
Federal Circuit has been trying to increase certainty and predictability by providing clear
and definite rules. Patent law’s current lack of certainty and predictability stemming from
the use of standards and vague rules is not likely solely attributable to the Federal Circuit.
Instead, the Supreme Court may be the major proponent of this type of uncertainty.
Of course, as discussed with respect to uniformity, perhaps these are examples of the
Supreme Court serving as the prime percolator.
2. Claim construction4

Of all the areas of patent law, claim construction has received the most scrutiny from
scholars, and reversal rates have been the central focus of these studies (Schwartz, 2010,
p. 1075; Gruner, 2010, pp. 994–96). The Federal Circuit’s decisions in Markman v. Westview
Instruments, 52 F.3d 967 (Fed. Cir. 1995) and Cybor Corp. v. FAS Technologies, Inc., 138
F.3d 1448 (Fed. Cir. 1998), and the Supreme Court’s decision in Markman v. Westview
Instruments, 517 U.S. 370 (1996), were partially motivated to enhance predictability of the
claim construction process (Schwartz, 2010, p. 1075). The question scholars have wrestled
with is whether this goal has been achieved (Schwartz, 2010, p. 1075).
Early studies examining claim construction report reversal rates ranging from 27
percent to 53.5 percent (Gruner, 2010, p. 995). Some included Rule 36 affirmances and
some did not, and the time periods varied as well (Gruner, 2010, pp. 995–1001). Many
of the authors of these studies concluded that the Federal Circuit’s claim construction
reversal rates “are excessive and indicative of major flaws in claim construction standards
and processes” (Gruner, 2010, p. 1001).
In a study of claim construction appeals from 1996–2007, Schwartz (2008) found that
29.7 percent of appeals from district courts were reversed, vacated, or remanded in a way
to vacate the judgment (Schwartz, 2008, p. 249). That study further showed that more
experience with patent cases did not affect the reversal rate (Schwartz, 2008, pp. 255–56).
Other studies show that judges with relatively recent patent experience are less likely to
4
Because another chapter of this book is dedicated to claim construction, this subsection
provides only a cursory overview of claim construction studies.

be reversed (Kesan and Ball, 2011, pp. 442–43; Stiernberg, 2013, pp. 294–95). Schwartz
(2008) hypothesized that possible explanations for this have included trial judges being
unable to master claim construction unless they have a technical background, the Federal
Circuit’s case law on claim construction being unclear, and claim construction being
inherently indeterminate because of the limitations of language (Schwartz, 2008, p. 267).
Yet the Kesan and Ball (2011) and Stiernberg (2013) studies show no support for the
theory that claim construction is inherently indeterminate so that district judges cannot
learn from prior experience (Kesan and Ball, 2011, p. 442; Stiernberg, 2013, pp. 294–95).
In a comprehensive study of Federal Circuit claim construction opinions from
1991–2008, Schwartz (2010) reports the reversal rates of district courts before and after
Markman (1995, 1996) and Cybor (1998) to see whether these cases improved the certainty
of claim construction and consequently reduced the reversal rates (Schwartz, 2010,
p. 1076). Importantly, Schwartz (2010) included Rule 36 affirmances in his dataset, which
gives a more accurate look at reversal rates (Schwartz, 2010, pp. 1090–91).
Schwartz’s (2010) results show that the reversal rate from January 1, 1991 through
the Federal Circuit’s April 1995 Markman (1995) decision was 20.8 percent (Schwartz,
2010, p. 1093). Between the Federal Circuit’s Markman (1995) decision and the Supreme
Court’s Markman (1996) decision in April 1996, the reversal rate dropped to 17.7 percent
(Schwartz, 2010, p. 1093). Between the Supreme Court’s Markman (1996) decision and
the Federal Circuit’s March 1998 decision in Cybor (1998), which held that the claim
construction was reviewed de novo, the reversal rate rose to 26.2 percent (Schwartz, 2010,
p. 1093). The seven-year period between Cybor (1998) and Phillips v. AWH Corp., 415 F.3d
1303 (Fed. Cir. 2005) (March 1998–July 2005) saw the reversal rate grow to 32.1 percent
(Schwartz, 2010, p. 1093). And from Phillips (2005) through the end of 2008, the reversal
rate shrank to 28 percent (Schwartz, 2010, p. 1093). In sum, the reversal rates were higher
after the Markman (1995, 1996) cases than before (Schwartz, 2010, pp. 1094–95).
Interestingly, from 1991–2008, the reversal rate was 28.5 percent and an additional 6.6
percent of appeals involved an affirmance of the district court, but still noted error in the
claim construction (Schwartz, 2010, p. 1095). This is consistent with prior studies, where
the reversal rates from 1996–2007 averaged 29.7 percent (Moore, 2005, p. 239; Schwartz,
2008, pp. 248–49). For 2008, the reversal rate was 37.5 percent (Schwartz, 2010, p. 1095).
Schwartz’s (2010) post-Markman (1996) reversal rates are abnormally high when consid-
ering that the average reversal rate of the Federal Circuit, excluding claim construction,
is 18 percent (Sichelman, 2010, pp. 1171–72).
Despite the results indicating that the claim construction reversal rate rose dramatically
after Markman (1996), Schwartz (2010) cautions that further empirical analysis is neces-
sary, including regressions for control variables to eliminate possible explanations for the
difference (Schwartz, 2010, p. 1098). Schwartz and others also warn of the limitations of
this study from selection effects (Schwartz, 2010, p. 1098; Gruner, 2010, pp. 985–88).
The importance of certain selection effects has been debated (Sichelman, 2010, p. 1172;
Gruner, 2010, pp. 1007–52). It has been argued that the high reversal rate for claim
construction is not problematic because the selection effects result in inherently uncertain
claims being appealed (Gruner, 2010, pp. 1007–09). This provoked a response that such
an explanation does not adequately explain the high reversal rates, because when com-
pared to reversal rates of other complex legal issues outside of the patent context, claim
construction remains the leader (Sichelman, 2010, pp. 1172–74). And even comparing

claim construction to other patent doctrines, the reversal rate far exceeds most other areas
(Sichelman, 2010, pp. 1174–76).
As far as implications of Schwartz’s (2010) study to certainty and predictability, it could
be that removing claim construction from the province of the jury has made patent law less
predictable (Schwartz, 2010, p. 1101). But, as Schwartz (2010) notes, de novo review for
claim construction is a confounding factor (Schwartz, 2010, p. 1101). Additional empirical
study is necessary to determine the precise effects that changes to claim construction have
had on predictability (Schwartz, 2010, p. 1107).
One step toward additional empirical research was Ted Sichelman’s (2010) analysis of
reversal rates from 2000–2007 on a variety of patent issues (Sichelman, 2010, pp. 1174–75).
Building on Schwartz (2010), Sichelman’s (2010) results show that claim construction has
a reversal rate of 33 percent (Sichelman, 2010, p. 1175). Compared to the reversal rates
of other patent issues, claim construction was near the top, only behind novelty and
indefiniteness, each with 38 percent reversal rates (Sichelman, 2010, pp. 1175–76). When
broken down by who prevailed at the district court level, there was no difference in the
claim construction reversal rates (Sichelman, 2010, pp. 1177–78). Gruner asserts that
these comparisons are not particularly enlightening because of the selection effects he
previously articulated, but acknowledges that future studies of reversal rates that control
for variables such as the stake asymmetry of the parties, the parties’ uncertainties as to
the quality of their cases, and the ability of one party to better estimate outcomes would
be worthwhile endeavors (Gruner, 2010, pp. 1062–65).
Rather than accepting reversal rates as the correct measure of uncertainty in claim
construction, Gruner (2010) argues that settlement rates would be better indicators of the
clarity of the claim construction standards (Gruner, 2010, p. 1064). When doing so, the
settlement rate for patent cases was 88 percent in 2008, up from 76 percent from 1995–99
(Gruner, 2010, p. 1065). Gruner (2010) concludes that these figures illustrate that predict-
ability is high among patent litigants and is increasing (Gruner, 2010, pp. 1065–66). These
results are interesting, but not specifically tied to claim construction; further research is
necessary to explore this relationship (Gruner, 2010, pp. 1069–70).
Lefstin (2007) adds a new twist to the empirical study of claim construction by analyz-
ing dissenting opinions at the Federal Circuit from 1998 through mid-2005 (Lefstin,
2007, pp. 1031, 1069). The theory is that the frequency of dissents can be used to measure
indeterminacy within the Federal Circuit (Lefstin, 2007, p. 1031). Lefstin (2007) argues
that dissent rates are better measures of uncertainty than reversal rates because the
observed group is small and relatively fixed as opposed to the more varied population of
district court judges (Lefstin, 2007, p. 1032). Additionally, because the two sets of courts
do not necessarily review the same set of facts and from the same vantage point, this
could skew the results (Lefstin, 2007, p. 1032). And although there are a host of potentially
confounding variables, comparing dissents within the same court, but in different fields of
law helps provide a benchmark to determine whether the uncertainty is driven by the law
itself or characteristics of the court and the judges (Lefstin, 2007, pp. 1032–34).
The results of this study show that the overall dissent rate from 1998–2005 for claim
construction was 8.3 percent (ranging from 1.4 percent to 19.5 percent) (Lefstin, 2007,
p. 1072). This was nearly identical to the dissent rate for infringement, invalidity, and ineq-
uitable conduct (Lefstin, 2007, p. 1072). From this, Lefstin (2007) concludes that “there is
no support for the hypothesis that over the entire period of this study, claim construction

was any more or less indeterminate than any other issue of patent law” (Lefstin, 2007,
p. 1072). However, because of the variation over time, a logistic regression was undertaken
and suggests that there was an increasing disparity in claim construction methodology
over time, and that from 2002–04 there was likely an increase in legal indeterminacy
(Lefstin, 2007, p. 1073). This corresponds, perhaps not coincidentally, to when the Federal
Circuit ordered Phillips v. AWH Corp., 363 F.3d 1207 (Fed. Cir. 2004) to be heard en banc
(Lefstin, 2007, p. 1074).
Lefstin (2007) also compared claim construction dissent rates with dissent rates of
contract interpretation in the regional circuits (Lefstin, 2007, p. 1075). It was thought that
contract interpretation is analogous to claim construction (Lefstin, 2007, p. 1075). The
dissent rates for contract interpretation ranged from 0.00 percent to 12.03 percent, with
an average of 5.74 percent (excluding the Federal Circuit) (Lefstin, 2007, p. 1077). When
compared to the overall dissent rates for each regional circuit, Lefstin (2007) observes that
the variation likely has more to do with structural differences between the circuits than
variation in the indeterminacy of the cases (Lefstin, 2007, pp. 1080–81). When comparing
claim construction to contract interpretation dissent rates, Lefstin (2007) finds that “over
the period from Cybor (1998) to Phillips (2005), the average indeterminacy of patent
claim construction was virtually indistinguishable from the indeterminacy associated with
contract interpretation at the regional [circuits]” (Lefstin, 2007, p. 1087). When limited
to reported cases, claim construction was more indeterminate, but the difference was
unremarkable (Lefstin, 2007, p. 1087). In sum, Lefstin’s (2007) study supports the view
that claim construction did not suffer from uncertainty; at least, not any more than other
areas within and outside patent law.
Despite the tremendous number of studies empirically analyzing claim construction
reversal rates and other objective outcomes, the Supreme Court’s recent decision in Teva
Pharmaceuticals USA, Inc. v. Sandoz, Inc., 135 S. Ct. 831 (2015) adds another layer of
complexity. In Teva (2015), the Court rejected the Federal Circuit’s de novo standard of
review for claim construction. Instead, the Federal Circuit must now review district court
claim construction under the clear error standard for subsidiary fact-finding, but review
the ultimate interpretation de novo (Teva, 2015). What change, if any, Teva (2015) will
have on the Federal Circuit’s development of claim construction law and how this will
affect certainty and predictability are worthy of future study.
Turning now to panel dependence, as Wagner and Petherbridge (2004) illustrate, from
April 23, 1996 through November 1, 2002, there was a sharp division at the Federal
Circuit between the “procedural” and “holistic” approaches to claim construction, and
claim construction was thought to be panel dependent (Wagner and Petherbridge, 2004,
pp. 1111–12, 1163). Others have concurred in this conclusion (Sichelman, 2010, pp. 1170,
1189).
Individual judges varied widely in their approaches to claim construction (Wagner
and Petherbridge, 2004, p. 1112). Some judges used a “procedural” approach, whereby
the judge starts with the ordinary understanding of the claim language and then departs
from that meaning only upon significant proof that such a departure is required under
the circumstances and according to an established hierarchy of sources (Wagner and
Petherbridge, 2004, pp. 1133–34). Other judges used a “holistic” approach, whereby the
judge seeks meaning of the claimed term by looking to the particular circumstances
presented rather than following formal steps and an established hierarchy (Wagner and

Petherbridge, 2004, p. 1134). The results of their study show that the procedural approach
to claim construction was used in 63.1 percent cases and the holistic approach was used
in 36.9 percent of cases, but the strength of the methods was polarized (Wagner and
Petherbridge, 2004, pp. 1148–49). When observed over time, the move toward the extremes
of these approaches increased (Wagner and Petherbridge, 2004, pp. 1149–51).
When broken down into individual judges on the court, this explains why there was
a shift over time to a more procedural approach: two additional procedural judges were
added to the court in 2000, Judges Linn and Dyk (Wagner and Petherbridge, 2004,
p. 1153). These same additions may explain why there was a shift in the strength of the
methods (Wagner and Petherbridge, 2004, pp. 1153–55). Further analysis shows that there
were factions on the Federal Circuit when it came to claim construction methods (Wagner
and Petherbridge, 2004, p. 1159). Judges Dyk, Clevenger, and Linn (the proceduralist) led
the shift toward a procedural approach and have varied little from this approach in com-
parison to the other members of the court (Wagner and Petherbridge, 2004, pp. 1161–63).
Unsurprisingly, the authors found that individual membership and composition of the
panel have a statistically significant effect on the claim construction approach used
(Wagner and Petherbridge, 2004, pp. 1163–69).
Despite the argument that panel dependence undermines certainty and predictability,
Wagner and Petherbridge (2004) were not necessarily troubled (Wagner and Petherbridge,
2004, p. 1169). They argue that those judges driving panel dependence provide some
certainty to the court, and that the rest of the judges who are not clearly classified
contribute to unpredictability (Wagner and Petherbridge, 2004, pp. 1169–70). This may be
true at the institutional level, but for individual litigants who do not learn of the make-up
of their panel until the day of oral argument, this level of predictability is likely of little
value (Wagner and Petherbridge, 2004, p. 1174). This study is now 15 years old and the
composition of the court has changed quite dramatically since then. An updated version
of this study may be appropriate.
In a more recent study of panel dependence for claim construction, Stiernberg (2013)
found that when a Federal Circuit judge has expertise in the technological field concerning
the patent in suit, the judge is more likely to find claim construction errors (Stiernberg,
2013, p. 295). Specifically, the probability of claim construction reversal increased by
53.42 percent when the judge had a technical background relevant to the patent at issue
(Stiernberg, 2013, p. 295). Unfortunately, there was insufficient data to assess whether the
Federal Circuit’s technical judges are equally likely to reverse district judges that share
technological expertise and whether these Federal Circuit judges are “unduly predisposed
to substitute their own claim construction for those of the district courts” (Stiernberg,
2013, p. 296). Further research, although admittedly difficult to conduct, would help bear
out whether the composition of the panel has an effect on claim construction (Stiernberg,
2013, p. 296).
Finally, turning to how vague rules or the use of standards affects certainty and pre-
dictability, the Federal Circuit has adopted various canons of construction for use in the
claim construction process and has done so, at least in part, in a quest for predictability
(Sichelman, 2010, p. 1191; Schwartz, 2010, pp. 1079–80; Petherbridge, 2010, p. 934). The
thought is that these canons produce a system of transparent rules, which may increase
certainty and predictability (Sichelman, 2010, pp. 1191–92; Schwartz, 2010, pp. 1079–80;
Wagner and Petherbridge, 2004, p. 1133; Petherbridge, 2010, p. 934). But because of

contradictions and tensions between them, application of these canons may lead to
several constructions and judges are then forced to exercise their best judgment in prop-
erly construing the claim (Sichelman, 2010, pp. 1191–92; Schwartz, 2010, pp. 1079–80;
Wagner and Petherbridge, 2004, p. 1133; Petherbridge, 2010, p. 934). As one scholar notes,
“commentators, practitioners, trial judges, and even some judges of the Federal Circuit
themselves seem united in their view that uncertainty and unpredictability are the order
of the day” (Lefstin, 2007, p. 1026).
In the early 2000s, judges on the Federal Circuit split over the proper process for claim
construction (Wagner and Petherbridge, 2004, p. 1111; Schwartz, 2010, p. 1080). Some
believed that initially resorting to the dictionary was proper, while others felt that the first
source for interpretation was intrinsic evidence—the claims, specifications, and prosecu-
tion history (Schwartz, 2010, p. 1080; Lefstin, 2007, p. 1074). These approaches are similar
to the procedural and holistic approaches described above (Wagner and Petherbridge,
2004, pp. 1133–36). The disagreement in approaches persisted for several years and
became more pronounced over time; then the Federal Circuit issued its en banc opinion
in Phillips v. AWH Corp. (2005) (Schwartz, 2010, p. 1080; Wagner and Petherbridge,
2004, p. 1155). Phillips (2005) rejected the dictionary-first approach and emphasized that
there is no “right” way to analyze claims other than ensuring that particular sources do
not contradict the unambiguous meaning of the claim in light of the intrinsic evidence
(Schwartz, 2010, p. 1080; Petherbridge, 2010, p. 937; Phillips, 2005). But disagreement
over how to construe claims remains (Schwartz, 2010, p. 1080; Petherbridge, 2010, p. 938).
Indeed, Phillips (2005) suggests that claim construction is not guided by clear and certain
rules, but is determined on a case-by-case basis (Petherbridge, 2010, p. 938). The Federal
Circuit’s loosely guided approach in Phillips (2005) rests on the premise that courts and
parties should focus on the understanding of the person having ordinary skill in the art
(Phillips, 2005; Sichelman, 2010, p. 1192). But, as Sichelman (2010) observes, the problem
is that the person having ordinary skill in the art “generally cannot understand a disputed
claim term without knowing the applicable law” (Sichelman, 2010, p. 1192). Gruner
(2010) takes a different view and concludes that the Federal Circuit’s claim construction
standards are succeeding because parties are settling in 88 percent of filed cases (and
numerous disputes that are never filed) (Gruner, 2010, pp. 986–88).
Before Phillips (2005) was decided, commentators noted that it was still an open ques-
tion as to whether the Federal Circuit was succeeding in bringing certainty and predict-
ability to claim construction, but the results were encouraging that the court was headed
in that direction (Wagner and Petherbridge, 2004, p. 1179). Despite additional studies in
the decade since Phillips (2005), it may still be too early to say for sure whether this goal
has been attained. An author who held out hope before Phillips (2005) was decided has
subsequently stated that Phillips (2005) is derationalizing and that it seems to encourage
unpredictability (Petherbridge, 2010, pp. 941–42). As it was a decade ago, the question of
the Federal Circuit’s successful contribution of certainty to the claim construction process
is still open.
3. Infringement
In addition to validity and claim construction, certainty regarding infringement is prob-
ably of the utmost concern to patentees and potential infringers. Henry and Turner’s
(2006) study of approximately 5000 patent cases from 1953–2002 focused on reversal

rates for infringement (Henry and Turner, 2006, pp. 95–99). They found that the Federal
Circuit affirmed district court findings that the patent was valid and infringed at a higher
rate than during the pre-Federal Circuit era (Henry and Turner, 2006, pp. 100–03). The
Federal Circuit’s affirmation rate was .723 for valid and infringed patents, whereas during
the pre-Federal Circuit era, the rate was .603 (Henry and Turner, 2006, pp. 102–03). For
district court findings of non-infringement, the Federal Circuit’s affirmation rate was
.685, which is nearly the same as the pre-Federal Circuit era affirmation rate of .694
(Henry and Turner, 2006, pp. 102–03).
As with their analysis of invalidity, the authors note that several factors complicate
their interpretation of these descriptive statistics and advised using a time-series analysis
(Henry and Turner, 2006, p. 103). The results of this analysis show that the Federal
Circuit’s affirmation rate of non-infringement findings was similar to that of the regional
circuits during the pre-Federal Circuit era (Henry and Turner, 2006, p. 109). For valid and
infringed decisions, there was an upward jump in affirmation until the late 1980s, but then
the affirmation rate dropped down to the pre-Federal Circuit rates for the rest of the study
period (Henry and Turner, 2006, pp. 110–12).
Interestingly, the focus of reversals of valid and infringed findings has shifted (Henry
and Turner, 2006, p. 112). During the pre-Federal Circuit era, 71 percent of reversals
of valid and infringed decisions were based on validity and 16 percent were based on
infringement (Henry and Turner, 2006, p. 112).5 During the Federal Circuit era, reversals
of valid and infringed decisions were based on validity 29 percent of the time compared to
40 percent for infringement (Henry and Turner, 2006, p. 112).6 Drilling down a bit more,
from 1983–1992, reversals were 41 percent and 32 percent for validity and infringement,
respectively (Henry and Turner, 2006, p. 112). From 1992–2002, the rates changed to 22
percent and 44 percent, respectively (Henry and Turner, 2006, p. 113). Like with their
study of invalidity, Henry and Turner’s (2006) study provides interesting general results,
but is not specific with respect to the particular infringement issues facing district courts
and the Federal Circuit.
Sichelman (2010), in his study of different reversal rates, shows that the doctrine of
equivalents and literal direct infringement both had overall reversals rate of 15 percent
(Sichelman, 2010, p. 1175). Contributory infringement had an overall reversal rate of 16
percent (Sichelman, 2010, p. 1175). To put these rates in perspective, the average reversal
rate across all issues during the relevant time was 21 percent (18 percent excluding claim
construction) (Sichelman, 2010, p. 1175). From this perspective, reversal rates related to
infringement were lower than average and perhaps not so uncertain.
But when broken down by whether the patentee or accused infringer prevailed at the
district court level, the reversal rates for the doctrine of equivalents were 33 percent and
12 percent, respectively (Sichelman, 2010, p. 1177). For literal direct infringement, the
reversal rates were 27 percent and 11 percent, respectively (Sichelman, 2010, p. 1177). And
for contributory infringement, the reversal rates were 0 percent and 21 percent, respec-
tively (Sichelman, 2010, p. 1177). Again, to give a point of reference, the average reversal
5
Vacated and remanded decisions make up the remaining cases.
6
This also shows that the Federal Circuit vacated or remanded more frequently (Henry and
Turner, 2006, p. 112).

rates across all issues (except claim construction) during the relevant time periods were 20
percent and 16 percent, respectively (Sichelman, 2010, p. 1177). From these perspectives,
some areas related to infringement are quite certain (contributory infringement), while
others are less so (doctrine of equivalents for prevailing patentees).
Petherbridge’s (2009) study of the doctrine of equivalents shows different results
(Petherbridge, 2009, p. 442). Using moving (lagged) averages, this research shows that
from 1992–2007, the Federal Circuit’s affirmation rate of doctrine of equivalents cases
ranged from 20 percent to 90 percent (Petherbridge, 2009, p. 442). When breaking
down the time periods, the results show an average affirmation rate of 71 percent from
1992–2000, but a drop to 49 percent from 2001–07 (Petherbridge, 2009, p. 443). These data
suggest that the doctrine of equivalents became less certain and predictable for district
court judges during this later period (Petherbridge, 2009, pp. 442–44).
Turning now to panel dependence, Petherbridge (2009) also reviews the role of
individual judges on the Federal Circuit to determine whether judge selection affects the
reversal rate (Petherbridge, 2009, pp. 445–46). His study rejects the hypothesis that “[t]here
are no differences in the likelihood of [affirmations of the district court’s decisions] when
different judges hear appeals” with a p-value of .000 (Petherbridge, 2009, p. 446). The
odds ratios for the various Federal Circuit judges indicate that several judges impacted
the affirmation outcome in a statistically significant way (Petherbridge, 2009, p. 447).
For example, Judges Bryson, Rich, Skelton, and Smith had strong positive effects on
affirmation, but Judges Linn, Prost, and Schall had strong negative effects (Petherbridge,
2009, p. 447).
Focusing on contemporaneously appointed judges and more active judges did not
change the story much. The hypothesis that “[t]here are no differences in the likelihood
of [affirmations of the district court’s decisions] when Judges Radar, Lourie, Clevenger,
Schall, and Bryson hear appeals” was rejected with a p-value of .008 (Petherbridge, 2009,
p. 450). The hypothesis that “[t]here are no differences in the likelihood of [affirmations
of the district court’s decisions] when Judges Bryson, Clevenger, Gajarsa, Linn, Lourie,
Michel, Newman, Radar, Schall, and Dyk author analyses between January 1, 2001 and
December 31, 2005” was also rejected with a p-value of .032 (Petherbridge, 2009, p. 450).
Grouping together particular judges as “proceduralists,” “swings,” and “holistics”
showed a statistically significant effect on affirmation outcomes (Petherbridge, 2009,
p. 455). In contrast, grouping together particular judges by the appointing President or
whether the judge had a patent background before appointment to the Federal Circuit
had no statistically significant effect on affirmation outcomes (Petherbridge, 2009, p. 455).
These results suggest that reversal rates related to the doctrine of equivalents may very
well be panel dependent. Of course, the composition of the court has changed quite
dramatically over the last several years, so an updated study would be worthwhile.
Finally, turning to how vague rules or the use of standards affect certainty and predict-
ability, after just over a decade of the Federal Circuit, Judge Michel (1994) described
the problem of uncertainty as it pertained to the doctrine of equivalents (Michel, 1994,
p. 1236). First, he noted that limitations on the doctrine, such as prosecution history estop-
pel and the rule against extending protection beyond the prior art, were rarely applied
(Michel, 1994, p. 1236). Moreover, the limitation that restricts extension to be within the
available range of equivalents that is proportionate to the scope of the claimed invention
viewed against the prior art has been useless to judges and the attorneys because of its

vagueness (Michel, 1994, pp. 1236–37). Notwithstanding the Federal Circuit’s element-by-

element test for equivalence, Judge Michel (1994) opined that with good jury instructions,
lay jurors struggled to understand the meaning of the tests and limitations to it (Michel,
1994, p. 1237). And when poor jury instructions were issued, Michel concluded that “the
jury’s task seems impossible” (Michel, 1994, p. 1237). Even as a judge on the Federal
Circuit, Judge Michel admitted to occasionally being “bafffled by the more basic question
under the doctrine of whether an accused device with structurally different components
nevertheless works in ‘substantially the same way’” under Graver Tank & Mfg. Co., Inc.
v. Linde Air Prod. Co., 339 U.S. 605 (1950) (Michel, 1994, p. 1237). Continuing on, Judge
Michel (1994) noted that the Federal Circuit has said the doctrine of equivalents should
not erase a plethora of meaningful claim limitations, but at the same time suggests that all
limitations are material and therefore meaningful (Michel, 1994, pp. 1237–38). As a result,
juries and district court judges find infringement analyses difficult and their findings
become unpredictable (Michel, 1994, p. 1238).
The results of Lunney’s (2004) study support the belief that infringement is uncertain
and unpredictable at the Federal Circuit. Instead of emphasizing reversal rates, Lunney
focused on success rate variability for infringement (Lunney, 2004, pp. 75–76). This study
shows that after creation of the Federal Circuit, patentees were more likely to prevail
on appeal on validity grounds, but less likely to succeed when it came to infringement
(Lunney, 2004, p. 14). This corresponds to a narrowing of the doctrine of equivalents by
the Federal Circuit (Lunney, 2004, pp. 27–28). If this narrowing was supposed to increase
certainty and predictability, then the success rate variability at the Federal Circuit should
decrease (Lunney, 2004, p. 75). However, the data indicate the opposite—success rate
variability increased (Lunney, 2004, p. 76). Prior to creation of the Federal Circuit, the
standard deviation for infringement success rates was 3.25 percent (Lunney, 2004, p. 76).
When analyzing the period of time after the Federal Circuit began to narrow the doctrine
of equivalents, the standard deviation increased to 10.2 percent (Lunney, 2004, p. 76).
From this, Lunney (2004) concludes that the increased success rate variability “suggests
that parties are less able to predict appellate litigation outcomes under the Federal Circuit
[and that] . . . the Federal Circuit and its doctrinal changes have brought less certainty and
predictability to patent enforcement” (Lunney, 2004, p. 76).
In contrast, nearly a decade after Judge Michel’s comments, Thomas (2003) argues
that the Federal Circuit has shifted toward bright-line rules with respect to the doctrine
of equivalents (Thomas, 2003, p. 781). In 1996, the court announced the rule that subject
matter disclosed, but not claimed, was dedicated to the public and could not serve as a
basis for infringement under the doctrine of equivalents (Thomas, 2003, pp. 781–82). But
two years later, the court opined that there was no per se rule that subject matter disclosed,
but not claimed, is not an equivalent (Thomas, 2003, p. 782). In 2002, the court resolved
this conflict by issuing an en banc opinion in Johnson & Johnson Associates, Inc. v. R.E.
Service Co., Inc., 285 F.3d 1046 (Fed. Cir. 2002), which adopted the public dedication doc-
trine, whereby subject matter disclosed, but not claimed, in the patent cannot be reached
under the doctrine of equivalents (Thomas, 2003, p. 782). This per se rule is an example
of a move from a flexible standard toward a bright-line rule (Thomas, 2003, pp. 782–83).
Likewise, with prosecution history estoppel, there was disagreement over the extent to
which this rule limited application of the doctrine of equivalents (Thomas, 2003, p. 783).
Some courts used a strict bar approach, which precluded any range of equivalents for the

amended limitation (Thomas, 2003, p. 784). Other courts adopted a flexible approach,
which only limited application of the doctrine of equivalents to situations where the
person having ordinary skill in the art would reasonably believe the patentee surrendered
the subject matter (Thomas, 2003, p. 784). The Federal Circuit had favored the flexible
approach, but changed course in its en banc opinion in Festo Corp. v. Shoketsu Kinzoku
Kogyu Kabushiki Co., Ltd, 234 F.3d 558 (Fed. Cir. 2000) and held that the strict bar
approach was appropriate (Thomas, 2003, pp. 784–85; Lunney, 2004, p. 35; Petherbridge,
2010, p. 924). To help justify this rule, the court noted that the need for certainty of the
patent’s scope was being emphasized (Thomas, 2003, p. 785; Festo, 2000).
Although the Federal Circuit moved toward a bright-line rule and an effort to
enhance certainty, similar to what we saw for validity, the Supreme Court rejected this
approach and replaced this rule with a rebuttable presumption of abandonment of
the subject matter (Thomas, 2003, pp. 785–86; Lunney, 2004, pp. 35–36; Petherbridge,
2010, pp. 924–25). Despite the move away from the more certain strict bar approach, the
Supreme Court’s new test was a middle ground between the strict bar approach and the
more flexible bar standard, so the blow to certainty was expected to be minor (Thomas,
2003, p. 786; Lunney, 2004, p. 36; Petherbridge, 2010, p. 925).
In sum, the Federal Circuit may have successfully moved toward more certainty and
predictability when it comes to adopting infringement rules (Thomas, 2003, pp. 781–86),
but as with validity, the source of any remaining uncertainty may stem more from the
Supreme Court than the Federal Circuit. Assuming these moves toward clear rules pro-
vide more certainty and predictability (which is questionable given the empirical studies
on reversal rates and panel dependence), the question arises whether such bright-line
rules are the right tools for the job; that is, has the Federal Circuit sacrificed quality for
certainty? We turn to that issue now.
V. QUALITY
A common critique of the Federal Circuit is that its rules and decisions are of poor,
or at least suboptimal, quality (Dreyfuss, 2008, pp. 791, 809; Nard and Duffy, 2007,
pp. 1620–21, 1644, 1645; Dreyfuss, 2010, p. 830). The first lens through which quality is
measured has been whether the court has been too formalistic in its decision-making;
that is, does the court pay enough attention to policy and technological considerations
(Dreyfuss, 2008, pp. 791, 803, 809)? Closely associated with this is the question of what
extent being a “specialized” court has had on the quality of the Federal Circuit’s rules
and opinions. The second lens through which quality is measured has been the Federal
Circuit’s use of scholarship (legal and otherwise) in its written opinions (Nard, 2002). This
criticism of quality and the measures of it has provoked rejoinders, most notably from
judges on the Federal Circuit (Plager and Pettigrew, 2007, pp. 1740–45). The remainder
of this section explores these claims, methods, and responses.
A. Formalism
The major criticism of the quality of the Federal Circuit’s decision-making is that it is
too formalistic (Dreyfuss, 2008, p. 791; Rai, 2003, pp. 1103–04; Lefstin, 2010, p. 846).

That is, there is a concern that the court is not engaged in serious thinking, at least with
respect to policy, when it adopts and applies the legal rules; instead, it produces isolated
and sterile opinions that are disconnected from those affected by patent law (Dreyfuss,
2010, p. 833; Nard and Duffy, 2007, pp. 1620–21). As Dreyfuss (2008) explains, “the
Federal Circuit tends to favor a kind of formalism that is more characteristic of legal
thinking in the nineteenth century than in the twenty-first” (Dreyfuss, 2008, p. 809). The
court “rarely provide[s] insight into the goals the court sees the law as achieving [and]
‘policy’ discussions take the form of incantations of standard justifications for statutory
terms” (Dreyfuss, 2008, p. 809). And as discussed earlier in regards to certainty, there has
been a move at the Federal Circuit away from flexible standards toward bright-line rules
(Thomas, 2003, pp. 773–74). Although bright-line patent rules may enhance certainty,
there is a risk that this distances patent law from innovation policy (Thomas, 2003, p. 774).
Examples of the Federal Circuit’s formalist approach abound. In Phillips v. AWH
Corp. (2005), despite “the torrent of ink that has been spilled on theories of statutory
construction and contract interpretation,” none of the judges engaged with those policy
discussions in their decisions (Dreyfuss, 2010, p. 835). Similarly, Dreyfuss (2004) notes
that in Schering Corp. v. Geneva Pharms. Inc., 339 F.3d 1373 (Fed. Cir. 2003), the court
cited a tremendous amount of precedent, but never focused on underlying policies
(Dreyfuss, 2004, pp. 777–78). Importantly, the two approaches at issue in Schering (2003)
have dramatically different effects on innovation systems—one “preserves incentives
to study known inventions and to find valuable features, while [the other] protects first
comers, the public, and the temporal limits of patents” (Dreyfuss, 2004, pp. 777–78).
Arti Rai (2003) discusses several examples of the Federal Circuit adopting a formalistic
bright-line approach to patent law, where a more flexible, policy-oriented approach may
make sense (Rai, 2003, pp. 1103–10). Rai points to the court’s patentable subject-matter
decisions in State Street (1998) and AT&T Corp. v. Excel Communications, Inc., 172 F.3d
1352 (Fed. Cir. 1999) as examples (Rai, 2003, pp. 1103–07). She also points to the court’s
rigid and policy-void opinions in In re Indep. Serv. Org. Antitrust Litigation CSU, L.L.C.
v. Xerox Corp., 203 F.3d 1322 (Fed. Cir. 2000) and Madey v. Duke University, 307 F.3d
1351 (Fed. Cir. 2002) concerning antitrust limitations and experimental use, respectively
(Rai, 2003, pp. 1107–10).
Burk and Lemley (2003) note that the Federal Circuit’s actual or functional elimination
of policy levers, such as the prohibition against business method patents, the printed
matter doctrine, the reverse doctrine of equivalents, and experimental use, was not based
on policy changes rendering the rules obsolete, but on a lack of specific authorization in
the Patent Act (Burk and Lemley, 2003, pp. 1672–73). These are additional examples of
formalism—strictly sticking to the text of the statute in spite of the need and invitation to
use policy to develop patent law and ensure its responsiveness to changed circumstances
(Rai, 2003, p. 1040; Burk and Lemley, 2003, pp. 1672–73).
More recently, Osborn (2012) describes the Federal Circuit’s formalism in terms of
failing to read precedent in context. For example, in Bilski (2008), the Federal Circuit
relied on the Supreme Court’s 1972 decision in Gottschalk v. Benson, 409 U.S. 63 (1972),
when it announced that the machine-or-transformation test was the sole test for patent-
able subject matter for process claims (Osborn, 2012, p. 429). The cornerstone of the
Federal Circuit’s explanation of this rule was that Benson (1972) described the machine-
or-transformation test as “the” clue, not “a” clue to patentability (Osborn, 2012, p. 431).

And although this is a reasonable interpretation “if the ‘clue’ language appeared in a
vacuum, . . ., it did not” (Osborn, 2012, p. 431). In short, the Federal Circuit clung to
specific precedential language, but failed to read it in context (or at least give due weight
to that context) (Osborn, 2012, pp. 431–32).
Assuming for the moment that the Federal Circuit is too formalistic in its decision-
making, what could be the cause? Lefstin (2010) argues that the Federal Circuit’s formal-
ism largely stems from its history (Lefstin, 2010, p. 847). In particular, he argues that
because the Federal Circuit adopted the CCPA’s body of case law as its own, and because
of the CCPA’s particular structure and needs, this has strongly shaped the development
of the Federal Circuit’s patent jurisprudence and decision-making, and yielded a court
steeped in formalism (Lefstin, 2010, p. 847).
The USPTO, Lefstin (2010) argues, only dealt with patentability (and occasional
interferences) and not infringement actions, and therefore had little need to appreciate
the consequences of granting patents of broad or narrow scope (Lefstin, 2010, p. 853).
The CCPA was limited to reviewing USPTO decisions and likewise had no need to
consider these consequences or those affecting innovation and competition (Lefstin,
2010, pp. 853–55). The CCPA’s case law reflects this lack of interest in downstream effects
from its and the USPTO’s decisions (Lefstin, 2010, pp. 853–56). Lefstin (2010) also notes
that the CCPA strongly emphasized that USPTO rejections must have been grounded
in a provision of the Patent Act, and that the CCPA used strict adherence to the text of
the statute to maintain control over USPTO practices (Lefstin, 2010, pp. 858–59). CCPA
references outside the statute were rare (Lefstin, 2010, p. 858).
By adopting the CCPA’s patent jurisprudence, the Federal Circuit implicitly rejected
the patent jurisprudence developed by the regional circuits (Lefstin, 2010, pp. 868–70).
And even when the CCPA had not spoken on a particular issue, such as infringement, the
Federal Circuit tended to ignore the law developed by the regional circuits (Lefstin, 2010,
p. 870). In fact, the Federal Circuit’s first case, South Corp. v. United States, 690 F.2d 1368
(Fed. Cir. 1982), which adopted the CCPA’s case law as precedent, set the tone for the
formalism for which the Federal Circuit is now critiqued (Lefstin, 2010, pp. 870–71). In
South Corp. (1982), Chief Judge Markey, rejecting a suggestion to adopt the laws of other
courts, noted that resolving existing conflicts between the circuits would require “a care-
ful, considered, cautious, and contemplative approach” (South Corp., 1982; Lefstin, 2010,
pp. 870–71). As described earlier, the Federal Circuit not engaging in careful, considered,
cautious, and contemplative decision-making is a similar refrain of legal scholars.
Another source of the Federal Circuit’s formalism has been attributed to it being a
specialized court. Although some take issue with the term “specialized court” because the
Federal Circuit hears a variety of types of cases (Plager and Pettigrew, 2007, p. 1740), it
cannot be disputed that the Federal Circuit has more exposure to patent cases than any
other court. As a result, Federal Circuit judges develop some level of expertise in emerg-
ing sciences and encounter new problems that arise when patent law is applied to new
technology (Dreyfuss, 2010, p. 839; Kesan and Ball, 2011, pp. 401–02, 408). And, as the
theory goes, specialization can improve the quality of the adjudication (Dreyfuss, 2004,
p. 770; Kesan and Ball, 2011, pp. 401–02, 408).
However, a potential problem with specialization is that channeling a particular type of
case to a single court may result in a lack of contact with other legal developments, which
may result in tunnel vision and yield stagnated law (Kesan and Ball, 2011, p. 404; Dreyfuss,

2004, p. 778). As Laura Pedraza-Fariña (2015) illustrates using sociological literature, the
Federal Circuit engages in typecasting and has an inability to self-coordinate, and these
contribute to what we understand as tunnel vision (Pedraza-Fariña, 2015, pp. 136–37,
140–42). Typecasting happens when an expert community treats a particular problem as
similar to other problems already solved within its domain of expertise, and is thus less
likely to look for other—potentially better—solutions outside of this area of expertise
(Pedraza-Fariña, 2015, pp. 136–37). The inability to self-coordinate arises when a problem
requires cooperation and coordination between two or more expert areas and one expert
institution places inadequate weight on the competing considerations and interests of the
other expert institution (Pedraza-Fariña, 2015, pp. 140–42). Pedraza-Fariña (2015) argues
that the Federal Circuit exhibits these qualities and explains that this leads to undesirable
consequences in the context of the Federal Circuit (Pedraza-Fariña, 2015, p. 142).
A few suggestions have been made on how to ameliorate the problems associated with
specialization. One is to expand the Federal Circuit’s jurisdiction to cover more issues,
especially those overlapping with the jurisdiction of the other circuits (Dreyfuss, 2004,
p. 786). Others are skeptical that such an approach would have any effect (Pedraza-Fariña,
2015, p. 153). A second is to have judges on the Federal Circuit sit by designation on
other courts (Dreyfuss, 2004, p. 796). Doing so would expose Federal Circuit judges
to jurisprudential trends across the country and other approaches to decision-making
(Dreyfuss, 2004, p. 796). The data suggest that in comparison to those in other circuits,
Federal Circuit judges have been infrequent visitors (Dreyfuss, 2004, pp. 794–95).
A third proposal is to have the Federal Circuit use visiting judges more often (Dreyfuss,
2004, p. 796). As with the second proposal, the interaction with other judges would
acquaint the Federal Circuit with different practices (Dreyfuss, 2004, p. 796). The data
here also show that the Federal Circuit has been an infrequent host to judges from
other courts (Dreyfuss, 2004, pp. 794–95). There are limitations to this third proposal,
however. Having visiting judges might be disruptive because other judges might not have
the expertise to deal with the technical materials frequently involved in patent litigation,
and as a result, the case might be viewed as compromised (Dreyfuss, 2004, p. 796). This
argument seems to assume, though, that a lack of exposure to a certain subject matter
results in a lower-quality opinion. To the extent this is true, having a visiting judge from a
district court only makes the problem worse, as studies show that visiting district judges
are overly deferential to the appellate judges (Dreyfuss, 2004, p. 796).
Alternatively, Laura Pedraza-Fariña (2015) posits a sociological theory for the Federal
Circuit’s formalism. This theory describes the Federal Circuit as a “weak expert com-
munity” and suggests that it uses formal rules to teach and control subordinate communi-
ties, such as the USPTO and district courts, to legitimize itself before relevant external
audiences, and to manage internal dissent (Pedraza-Fariña, 2015, pp. 128–36). Calls for
empirical research to test these functions have been made (Pedraza-Fariña, 2015, p. 136).
Such studies could include measuring “whether a court is more likely to prefer bright-line
rules over flexible, indeterminate tests in periods of high-judge turnover, in periods with
great epistemic diversity among judges, or in periods of crises of negative public opinion”
(Pedraza-Fariña, 2015, p. 136). Comparisons between patent and non-patent dockets and
one or more specialist courts could also prove informative (Pedraza-Fariña, 2015, p. 136).
Regardless of the reason for the Federal Circuit’s formalism, why does it matter? What
benefits flow from an anti-formalist approach to decision-making or what costs are asso-

ciated with a formalist jurisprudence? Ideally, enhanced quality would result in opinions
where the court sets out the policies it is trying to achieve, discusses alternative approaches
and explains why the chosen position best achieves the objectives (Dreyfuss, 2008, p. 803).
Articulating rules and application in this matter could promote coherence; legal doctrines
could fit together in a purposeful way rather than haphazardly (Dreyfuss, 2008, p. 803).
From the other perspective, formalism undermines confidence in the court’s decisions
because it fails to “test[] the accuracy of its positions by trying to explain them” (Dreyfuss,
2008, p. 809). Finally, the Federal Circuit’s formalistic approach creates a vacuum in which
there is no law, incoherent law, or bad law to apply to changed circumstances (Dreyfuss,
2010, pp. 838–39). Reinterpreting patent law is necessary as the innovation economy
has changed dramatically—the explosion of biotechnology and computer science, the
development of nanotechnology, the movement to upstream patenting of fundamental
advances in new fields, and shifts in the business of patenting (Dreyfuss, 2010, pp. 838–39).
But not everyone is in accord with the view that the Federal Circuit is formalistic or that
if it is, that this is problematic. Several Federal Circuit judges have been outspoken about
the resistance to using policy in their decision-making (Taylor, 2013, p. 641). For example,
Judge Newman (1993) has stated, “I caution against . . . policy-driven activism whereby
the application of the law will not be known until the Federal Circuit hears the case . . .
It is policy choices that lead to departure from precedent, into the judicial activism that
weighs against legal stability” (Newman, 1993, p. 688; Taylor, 2013, p. 641). She concludes
with “policy choices are not the province of judges” (Newman, 1993, p. 688). Judge Lourie
(2006) notes that the court is “not a policy-making legislature” or a “debating society
having debates with outside groups on what the law should be” (Lourie, 2006, p. 41;
Taylor, 2013, p. 641). Instead, he describes the decision-making process as just applying
the precedent as best as the judges can to the cases coming before it (Lourie, 2007, p. 22).
Judge O’Malley (2012) has also eschewed policy in the court’s decision-making (O’Malley,
2012, p. 92; Taylor, 2013, p. 644). Judge O’Malley’s rationale is that the court needs to
determine what the law is and judges consult a crowded field of authority that limits the
court’s ability to rely upon policy (O’Malley, 2012, pp. 93–94; Taylor, 2013, p. 644).
Judge Plager echoes the same sentiment, but deconstructs the argument a bit more
(Plager and Pettigrew, 2007, pp. 1737–38; Taylor, 2013, p. 642). In examining whether the
Federal Circuit’s decisions are of sufficient quality, Nard and Duffy (2007) ask whether
the decisions “adequately reflect[] current knowledge regarding the beneficial functions of
the patent system in generating technological innovation, the potential problems of patent
rights in foreclosing legitimate competition, and the need for predictable rules capable of
curtailing litigation costs” (Nard and Duffy, 2007, p. 1620). Plager and Pettigrew (2007)
challenge the first two of those questions as not necessarily being indicative of high-
quality decision-making (Plager and Pettigrew, 2007, pp. 1740–45).
Moreover, Plager and Pettigrew (2007) reject the view that the Federal Circuit should
engage in setting patent policy and correcting the perceived errors of Congress and the
Supreme Court (Plager and Pettigrew, 2007, p. 1743). They note that the court is merely
interpreting a federal statutory scheme and is limited accordingly (Plager and Pettigrew,
2007, p. 1753). As a result, the court cannot engage in common law-esque development of
patent law (Plager and Pettigrew, 2007, p. 1753). Although he acknowledges that courts
must sometimes exercise judgment in interpreting a statute, Judge Plager rejects the notion
that the court should “assess the extent to which the congressional policy is responsive

to current problems or to determine how well-tuned the statute is to subtle changes in

people’s behavior or market conditions” (Plager and Pettigrew, 2007, pp. 1737–39).
Although there is some minor variation in the judges’ responses to the critique of
formalism, a common thread seems to be a fear that the court will be seen as an “activist”
court (Osborn, 2012, p. 458; Burk and Lemley, 2003, p. 1638; Rai, 2003, p. 1116). But as
some commentators have noted, this should be less of a concern for the Federal Circuit
because the Patent Act was designed for judicial discretion to fill the gaps in the statutory
scheme (Osborn, 2012, p. 458; Burk and Lemley, 2003, p. 1638; Rai, 2003, p. 1116). The
court should not shy away from this duty, but embrace it (Osborn, 2012, p. 458; Burk and
Lemley, 2003, pp. 1638, 1674; Rai, 2003, p. 1116).
Federal Circuit judges responding to arguments that the Federal Circuit produces
low-quality opinions may, of course, be biased. Taylor (2013) provides academic support
to the judges’ responses and argues against the view that the Federal Circuit is overly
formalistic. Taylor (2013) asserts that the court does express its policy views when
identifying the underlying, basic purpose behind a doctrine and when the court hears a
case en banc (Taylor, 2013, p. 638). That being said, Taylor (2013) does not go so far as
to argue that the Federal Circuit is a shining example of anti-formalism. He recognizes
that there are numerous instances where the court’s opinions, even in important cases, are
devoid of policy discussion when such a discussion might be appropriate (Taylor, 2013,
pp. 652–57). Nonetheless, Taylor (2013) supports his argument with examples of cases
where the Federal Circuit does address policy in its decision-making process, including
Pfaff v. Wells Electronics (1997) (on-sale bar), Inc., Integra Lifesciences I Ltd. v. Merck
KGaA, 331 F.3d 860 (Fed. Cir. 2003) (statutory safe harbor), Board of Trustees of Leland
Stanford Junior University v. Roche Molecular Systems, Inc., 583 F.3d 832 (Fed. Cir. 2009)
(application of Bayh–Dole), and LG Electronics, Inc. v. Bizcom Electronics, Inc., 453 F.3d
1364 (Fed. Cir. 2006) (exhaustion doctrine) (Taylor, 2013, pp. 657–58).
Taylor (2013) also recognizes that the use of policy arguments has been used in dis-
senting opinions from denials of en banc rehearing and in en banc opinions (including
concurring and dissenting opinions) (Taylor, 2013, pp. 659–61). A prime example of the
Federal Circuit using policy in its decision-making is its en banc decision in Festo (2000)
(Taylor, 2013, p. 661). The majority opinion considered the notice function of claims and
the need for certainty of patents’ scope and how adopting the complete bar to prosecution
history estoppel will stimulate innovation and competition (Taylor, 2013, p. 661). Judge
Lourie’s concurrence and the dissenting opinions in Festo (2000) also heavily relied upon
policy considerations (Taylor, 2013, pp. 661–63). Based on cases such as this, Taylor (2013)
illustrates that the Federal Circuit does use policy rationales to justify its decisions, and
that these policy considerations range from basic ones underlying any law to policies
generally relevant to patent law to policies specific to particular patent law doctrines
(Taylor, 2013, p. 663).
Despite his claims that the Federal Circuit does, in fact, engage in policy analyses,
Taylor (2013) emphasizes that in the ongoing dialogue between the Supreme Court and
the Federal Circuit, there should be a discussion of the role of policy (Taylor, 2013,
pp. 675–76). To the extent this discussion over policy continues, Taylor (2013) believes
patent law, as a whole, should improve (Taylor, 2013, p. 676). To encourage this, Taylor
(2013) suggests (1) identifying doctrines susceptible to policy analysis (i.e., they have not
been crowded out by other sources of interpretation), (2) reading statutory provisions

with an eye toward their underlying policy or policies, (3) considering all policies, not
just the one of most interest to the Federal Circuit—certainty; (4) requesting additional
briefing from parties and amici on policy issues in cases heard by panels and the en banc
court; and (5) making an effort to familiarize themselves with secondary materials related
to innovation policy, economics of the patent system, and specific patent doctrines
(Taylor, 2013, pp. 677–82).
B. Scholarship
The Federal Circuit’s use of scholarship has also been a way to evaluate the quality of
the court’s decisions. In 1994, Judge Michel noted that unlike the Supreme Court, the
Federal Circuit seldom cites scholarship, and speculates that this may have been because
the patent literature was not empirical (Michel, 1994, p. 1245). He called upon scholars
to produce empirical scholarship and host conferences to share their work so these
studies could be cited by the parties and amici and help ensure the court’s decisions
reflect true relationships rather than assumed ones (Michel, 1994, p. 1255). If there was
a lack of empirical patent scholarship, it appears as if that era is long gone (Nard, 2002,
pp. 669–71). The existence of this handbook strongly evidences that this is no longer the
case! Perhaps we should see an increased reliance on scholarship.
Despite the production of an enormous amount of empirical scholarship about patent
law’s relationships with firms’ innovations and research and development, the role of
juries, litigation trends, voting patterns, claim scope, as well as social science research
on economics and innovation policy (Nard, 2002, pp. 669–71), scholars argue that this
literature has largely been absent from the Federal Circuit’s patent opinions (Nard, 2002,
pp. 673–74; Dreyfuss, 2004, p. 772). Dreyfuss gives the example of the court’s struggles
with the doctrine of equivalents (Dreyfuss, 2004, p. 783). As she describes:
The Federal Circuit launched its attack with the charge that patent attorneys were making “the
doctrine of equivalents . . . simply the second prong of every infringement charge.” The suspi-
cion that lawyers were abusing the system led the court to adopt a rather ungenerous view of the
doctrine. That interpretation severely limited the breadth of patent claims, and ultimately led to
two rounds of remand by the Supreme Court. Had the Federal Circuit considered the accelerat-
ing pace of technological innovation—an issue richly studied in the economic literature—then
it might have viewed reliance on the doctrine more benignly, as an attempt to deal with the
abundance of after-arising technology and the ease with which this technology can now be used
to replace elements of the invention as literally claimed in the patent. Moreover, the court failed
to draw on empirical evidence demonstrating that thickets of narrow patents raise transaction
costs and give rise to other innovation-deadening problems (Dreyfuss, 2004, p. 783).
To test whether the Federal Circuit is less likely to use scholarship, Nard (2002) reviewed
published Federal Circuit opinions from 1983–2000 to determine how frequently the
court cited scholarship in patent and non-patent cases (Nard, 2002, p. 678). Nard’s (2002)
results show that the Federal Circuit cited scholarship more frequently in patent cases
than non-patent cases and the court had steadily declined its use over the last ten years of
the study (Nard, 2002, pp. 678–80).
Nard (2002) then compared the Federal Circuit’s citation rates in patent cases with
those of the Second and Ninth Circuits in copyright and trademark cases (Nard,
2002, p. 681). This study showed that the Second and Ninth Circuits cited scholarship

considerably more than the Federal Circuit (Nard, 2002, pp. 681–82). In particular, these
two circuits cited scholarship nearly four times more frequently than the Federal Circuit
(Nard, 2002, p. 683).
These data, of course, do not explain the cause of this discrepancy. Nard (2002)
speculates that it could be that the Federal Circuit is more familiar and comfortable with
its subject matter than the Ninth and Second Circuits are with theirs, and therefore has
less need to consult scholarship (Nard, 2002, p. 683). Alternatively, it could be that the
judges on the Second and Ninth Circuits are drawn more from the academy than judges
on the Federal Circuit (Nard, 2002, p. 683). Another possibility is that more copyright and
trademark scholarship exists, so there is more to draw from (Nard, 2002, p. 683).
Rather than compare the Federal Circuit’s use of scholarship in patent cases to the Ninth
and Second Circuits’ uses in copyright and trademark cases, Schwartz and Petherbridge
(2012) compared the citation rates across all circuits (Schwartz and Petherbridge, 2012,
p. 1563). In this study, Schwartz and Petherbridge (2012) study citations to legal scholar-
ship by all regional circuits from 1990–2008 (Schwartz and Petherbridge, 2012, pp. 1569,
1576–77). This study concludes that the Federal Circuit’s use of legal scholarship is similar
to the other regional circuits, thus undercutting the argument that the Federal Circuit’s
use of scholarship is inadequate because it uses it less than the other circuits (Schwartz
and Petherbridge, 2012, p. 1569).
In particular, the Federal Circuit used legal scholarship in 5.07 percent of its reported
opinions (Schwartz and Petherbridge, 2012, p. 1578). In comparison, the Federal Circuit
cited legal scholarship more frequently than the Eleventh and Eighth Circuits (4.79
percent and 2.41 percent, respectively) (Schwartz and Petherbridge, 2012, p. 1578). After
eliminating outlier judges, the Federal Circuit fell near the middle of the pack (Schwartz
and Petherbridge, 2012, pp. 1579–81). When limiting the study to the Federal Circuit’s
patent cases in comparison to the regional circuits’ uses in all cases and in all civil cases,
the results show the Federal Circuit cited legal scholarship in 6.01 percent of its patent
opinions, and that this number is consistent with many of the other circuits (Schwartz and
Petherbridge, 2012, pp. 1586–88). Schwartz and Petherbridge (2012) also confirm Nard’s
(2002) conclusion that the Federal Circuit uses scholarship in patent cases slightly more
than it does in non-patent cases (Schwartz and Petherbridge, 2012, pp. 1570, 1588).
Regardless of the Federal Circuit’s comparative citation rates, there are numerous
criticisms of these measures and limitations to these studies. One objection is whether
engaging with scholarship in opinions is proper. In dismissing the criticism that the
Federal Circuit rarely cites legal literature about patent law, Judge Plager (2009) questions
whether engaging with and relying on such scholarship is appropriate because it may
not be presented by the parties to the litigation and they may not have an opportunity
to review and comment upon it (Plager, 2009, p. 1338; Plager and Pettigrew, 2007,
pp. 1750–52). To the extent the parties themselves rely upon this literature, this may be a
more jurisprudentially sound way to encourage the court to engage with the legal scholar-
ship (Plager, 2009, pp. 1338–39).
A limitation of these studies is that citations might not properly measure the court’s
engagement with the literature. Just because an opinion does not cite a piece of scholar-
ship does not mean the court was unaware of it (Nard, 2002, p. 685; Plager, 2009, p. 1339).
It could be that the court was unpersuaded by it, or that other forms of analysis were
superior or sufficient to arrive at the given outcome (Nard, 2002, p. 685). Nonetheless,

Nard argues that judges should “should err on the side of citation or, more dramatically,
offer a discussion of the cited scholarship,” because citing and discussing scholarship
signals to the relevant community that the court is aware of it and that its decisions have
consequences (Nard, 2002, p. 685).
Another criticism of the suggestion to rely upon scholarship is that there are substantial
costs in doing so. Given the judges’ and clerks’ other obligations, the judges cannot
keep up with scholarship (Plager, 2009, p. 1339). Moreover, scouring the literature is
resource-intensive and separating the wheat from the chaff is even more so (Nard, 2002,
pp. 686–88). Although the court may prefer to use the literature more, it may be impracti-
cal for it to do so.
And finally, as noted before, the make-up of the Federal Circuit has changed sub-
stantially over the last several years. Schwartz and Petherbridge’s (2012) study only ran
through 2008; it may be a worthwhile endeavor to see what changes, if any, have taken
place with respect to the Federal Circuit’s use of scholarship over the past few years.
Furthermore, as Schwartz and Petherbridge (2012) suggest, scholars might also look to
see what influences the Federal Circuit’s decision to use legal scholarship and whether
these influences differ from those used in the other circuits, whether the Federal Circuit
normatively uses legal scholarship enough, and whether legal scholarship is helping with
the development of patent law (Schwartz and Petherbridge, 2012, pp. 1570, 1593–97).
VI. STRUCTURE AND STAFFING
The previous sections of this chapter have reviewed the extent to which the Federal Circuit
has achieved the goals it was established to achieve. This section takes a different approach
and investigates the Federal Circuit’s structure, the make-up of the individual judges
on the court, and how these affect the court as an institution. This section starts with
individual judges and eventually expands to the entire court sitting en banc. In particular,
this section explores how individual judges or types of judges affect the court and its
patent jurisprudence. It then discusses the internal dynamics of the court and how these
impact the decision-making process. This section concludes by examining the Federal
Circuit when it sits en banc.
A. Individual Judges
There is a common belief among patent practitioners that decision-making at the Federal
Circuit is highly judge-dependent (Field, 2014, p. 627). And although a distinction is
drawn between individual judges who happen to sit on the court and the court as an
institution (Duffy, 2004, p. 804), looking at common characteristics among judges and
structural issues affecting those judges and the court provides some insight into how the
Federal Circuit operates and the influence these characteristics and issues have on the
Federal Circuit’s patent jurisprudence.
For example, Paul Gugliuzza (2012) notes that as of 2012, Federal Circuit judges’
experiences heavily skew in favor of patents and international trade, but no one has a
background specializing in government contracts, veterans, or personnel law (Gugliuzza,
2012, pp. 1468–69). This overemphasis on patents and international trade may restrict the

court from having a broad perspective to draw from and may contribute to the court’s
formalism in its patent jurisprudence (Gugliuzza, 2012, pp. 1468–69). To the extent there
are, in the future, judges with more diverse backgrounds on the court, the effect may be
marginal. The traditional backgrounds in patent and international trade may have led to
a culture in which judges on the Federal Circuit see themselves as experts in patent law
and new judges will be acculturated to the social practices of the Federal Circuit (Pedraza-
Fariña, 2015, p. 119). And as discussed in more detail below, others have focused attention
on the background of the Federal Circuit judges and whether these backgrounds influ-
ence decision-making generally and application of particular doctrines.
Because of studies such as these, it seems appropriate to analyze the individual judges
comprising the Federal Circuit, given their roles influencing the patent system. The
remainder of this subsection discusses studies measuring outcomes based on the back-
grounds and characteristics of the judges, how turnover affects the court, and the role that
case distribution plays in the structure of the Federal Circuit and its patent jurisprudence.
1. Judicial backgrounds and characteristics

Many people have assumed that the Federal Circuit is nothing more than a patent court
and that the judges on it are “narrowly specialized technology lawyers who spent all
their pre-appointment years practicing only patent law” (Michel, 2010, pp. 1200–01).
This is not, and never has been, the case (Michel, 2010, pp. 1200–01). Instead, the court
hears a variety of different cases and the court is comprised of judges with a variety of
backgrounds, some technical and some not (Michel, 2010, pp. 1200–01). Judge Michel
(2010), upon his retirement, believed that having a mix of judges with and without patent
backgrounds is its greatest strength and hopes that the court continues to maintain the
same proportion of patent versus non-patent judges as it has had over the last 30 years
(Michel, 2010, p. 1202).
But what effect have these judges with their different backgrounds had on patent
jurisprudence and outcomes? Several empirical studies have undertaken to answer that
question. In an early study, Allison and Lemley (2000) analyzed the Federal Circuit’s
patent validity decisions from 1989–96 and examined individual judges and groups of
judges based on when they were appointed, whether they had patent backgrounds before
joining the court, and the party of the appointing President (Allison and Lemley, 2000,
pp. 746, 751–52). Of the 22 judges in this study, 11 were appointed to the bench before
creation of the Federal Circuit and 11 were appointed after its creation (Allison and
Lemley, 2000, p. 751). Of the 22 judges, six had patent law backgrounds prior to joining
the court (Allison and Lemley, 2000, p. 751). Of the 22, 14 were appointed by Republican
presidents and eight were appointed by Democratic presidents (Allison and Lemley, 2000,
p. 752).
Allison and Lemley’s (2000) results show that judges with patent backgrounds accounted
for 38.2 percent of the participations, but wrote 62.9 percent of the opinions (Allison and
Lemley, 2000, pp. 752–53). These results suggest that judges with patent backgrounds
were more likely to be assigned to write the majority opinions (Allison and Lemley, 2000,
p. 753). Drilling down, they show that specific judges author a disproportionate number
of patent validity opinions (Allison and Lemley, 2000, p. 753). For example, Judge Lourie
wrote the opinion in 73.8 percent of the cases he participated in (Allison and Lemley,
2000, p. 753). Judge Markey wrote the opinion in 52.4 percent, Judge Newman wrote the

opinion in 58.5 percent, and Judge Rich wrote the opinion in 52.2 percent (Allison and
Lemley, 2000, p. 753). As a result, they conclude that a small number of judges with patent
backgrounds have a significant influence on the law concerning patent validity (Allison
and Lemley, 2000, p. 753).
But when looking at whether judges with patent backgrounds upheld validity more or
less than those without patent backgrounds, the authors found that judges with patent
backgrounds upheld the validity of patents 57.9 percent of the time, while those without
patent backgrounds upheld the validity 52.1 percent of the time (Allison and Lemley,
2000, p. 754). Extending these results to the superpopulation, the authors’ Chi-squared
test resulted in a p-value of 0.466, indicating that they could not reject the null hypothesis
and not confidently predict that having a patent law background influenced validity
voting (Allison and Lemley, 2000, p. 755).
When looking at whether judges appointed after the creation of the Federal Circuit
were more “pro-patent” than those appointed beforehand, the authors found that there
was no significant difference in the validity outcomes between these two groups (Allison
and Lemley, 2000, p. 754). Extending these results to the superpopulation, the authors’
Chi-squared test resulted in a p-value of 0.621, indicating that they could not reject the
null hypothesis and not confidently predict that year of appointment influenced validity
voting (Allison and Lemley, 2000, p. 754).
Allison and Lemley (2000) also investigated voting patterns of individual judges to
measure whether they were “pro-patent” or “anti-patent” (Allison and Lemley, 2000,
p. 755). Most judges were right around the average, but there were a few outliers (Allison
and Lemley, 2000, p. 755). For example, Judges Mayer and Markey were more inclined
to hold a patent invalid, while Judges Newman, Plager, and Radar were more likely to
uphold the patent’s validity (Allison and Lemley, 2000, p. 755). Nonetheless, the authors
note that too much weight should not be placed on these outliers, given the small number
of cases for each judge (Allison and Lemley, 2000, p. 755). Extending these results to the
superpopulation, the authors’ Chi-squared test resulted in a p-value of 0.404, indicating
that they could not reject the null hypothesis and not confidently predict that individual
judges vote differently in validity cases (Allison and Lemley, 2000, p. 755). However, when
looking only at authored opinions, the results were different (Allison and Lemley, 2000,
pp. 756–57). Applied to the superpopulation, the authors’ Chi-squared test resulted in a
p-value of 0.001, indicating that they could reject the null hypothesis and predict with a
high degree of confidence that who writes the majority opinion is related to whether the
court finds the patent valid (Allison and Lemley, 2000, p. 757).
Looking at affirmation rates and applying them to the superpopulation, the authors’
data show that when the judges were appointed and whether they had patent backgrounds
were not statistically significant in predicting affirmation of validity (Allison and Lemley,
2000, pp. 758–59). Allison and Lemley (2000) also investigated voting patterns of indi-
vidual judges to measure their affirmation rates (Allison and Lemley, 2000, p. 759). Most
judges were right around the average, but there were a few outliers (Allison and Lemley,
2000, p. 759). Extending these results to the superpopulation, the authors’ Chi-squared
test resulted in a p-value of 0.069, indicating that they could reject the null hypothesis
with only 90 percent confidence (Allison and Lemley, 2000, p. 760). Thus, they could only
weakly predict that individual judges vary significantly in the likelihood that they will vote
to affirm (Allison and Lemley, 2000, p. 760).

When looking only at authored opinions, the results are similar (Allison and Lemley,
2000, pp. 760–61). Applied to the superpopulation, the authors’ Chi-squared test resulted
in a p-value of 0.249, indicating that they could not reject the null hypothesis and not
confidently predict the court would vote differently in affirming or reversing the district
court when different judges wrote the majority opinion (Allison and Lemley, 2000,
p. 761).
Finally, Allison and Lemley (2000) analyzed whether Federal Circuit judges were more
likely to affirm findings of juries or judges (Allison and Lemley, 2000, p. 762). They found
that there was little difference between pre-1982 appointed judges and post-1982 judges
(Allison and Lemley, 2000, pp. 762–63). However, judges with patent backgrounds were
quite different than those without (Allison and Lemley, 2000, p. 763). Judges with patent
backgrounds were less likely to defer to jury findings than their peers without patent
backgrounds (Allison and Lemley, 2000, p. 763). However, when it came to affirming
district court judges, there was no appreciable difference between judges with and without
patent backgrounds (Allison and Lemley, 2000, pp. 763–64).
In a more recent study, Barnes (2013) examined whether judges with technical expertise
analyze Section 112’s enablement and written description requirements differently than
other judges on the Federal Circuit (Barnes, 2013, p. 972). This study uses the Federal
Circuit’s Section 112 cases from 1997–2011 (Barnes, 2013, p. 974). Because of the way
Barnes conducted his search for these cases, Rule 36 affirmances were not included and
some cases may have been omitted (Barnes, 2013, pp. 988–89). He measured technical
expertise by noting whether the judge had a bachelor’s degree in a technical subject or had
practical engineering or scientific experience and passed the Fundamentals of Engineering
test (Barnes, 2013, p. 981). Seven Federal Circuit judges were classified as having a techni-
cal background and 27 were classified as non-technical (Barnes, 2013, p. 983–85). Barnes
also included district court judges who sat by designation on the Federal Circuit, but all
were non-technical (Barnes, 2013, pp. 985–86). To the extent we are only interested in
the technical or non-technical backgrounds of active and senior Federal Circuit judges,
Barnes’ inclusion of district judges sitting by designation muddies the waters.
When comparing reversal rates on Section 112’s disclosure requirements, Barnes (2013)
finds that technical judges reverse in 56.8 percent of the cases, whereas non-technical
judges reverse only 24.6 percent of the time (Barnes, 2013, pp. 995–96). He breaks the
data down further to compare reversal rates when appeals stem from district courts and
the BPAI and similar disparities exist (Barnes, 2013, pp. 995–96). Technical judges reverse
district courts 60 percent of the time, whereas non-technical judges only reverse 29.2
percent of the time (Barnes, 2013, p. 996). For BPAI appeals, technical judges reverse in
42.9 percent of the cases, in comparison to 12.1 percent for non-technical judges (Barnes,
2013, p. 996). All of these results were statistically significant at p < .05 using Pearson’s
chi-squared test (Barnes, 2013, p. 996).
Similar analyses were done comparing patent invalidation rates and the results were less
stark (Barnes, 2013, p. 996). Overall, technical judges found patents invalid on enablement
and written description in 56.7 percent of the cases, while non-technical judges did so in
59.8 percent of the cases (Barnes, 2013, p. 996). The differences between technical and
non-technical judges’ invalidation rates from district court and BPAI appeals were also
modest (Barnes, 2013, p. 996). None of these results was statistically significant (Barnes,
2013, p. 996).

When limiting the study to whether the authoring judge had a technical background,
Barnes (2013) shows that reversal rates were higher when the judge had a technical
background (47.1 percent) than authors without a technical background (23.4 percent)
(Barnes, 2013, p. 1000). Invalidation rates, however, were similar, with technical authoring
judges invalidating the patent in 55.2 percent of the cases and non-technical authoring
judges invalidating the patent in 58.2 percent of the cases (Barnes, 2013, p. 1000). These
results are similar to those found when looking at entire panels (Barnes, 2013, p. 1000).
Petherbridge’s (2009) study of Federal Circuit judges and their views on the doctrine
of equivalents is another example of investigating judicial characteristics (Petherbridge,
2009, pp. 444–55). In this study, he rejected the null hypotheses that there were no
differences in the likelihood of success on appeal, dispositive success on appeal, and
affirmation rates when different judges heard the appeals with p-values of .000, .000, and
.001, respectively (Petherbridge, 2009, pp. 445–46). Logistic regression models illustrated
similar results and showed that the judges had varying levels of influence on the measured
outcome variables (Petherbridge, 2009, pp. 446–49).
As described earlier, his study showed that certain judges, such as Bryson, Rich,
Skelton, and Smith, had strong positive effects on affirmation of district court decisions
concerning the doctrine of equivalents, while Judges Linn, Prost, and Schall had strong
negative effects (Petherbridge, 2009, p. 447). Looking at judges appointed contemporane-
ously and at the most active judges on the court also allowed Petherbridge (2009) to
reject the null hypotheses that their involvement did not affect the measured outcomes
(Petherbridge, 2009, pp. 450–51). Likewise, grouping judges together as “proceduralists,”
“swings,” and “holistics” showed a statistically significant effect on the measured out-
comes (Petherbridge, 2009, p. 455). In contrast, grouping together particular judges by the
appointing President or whether the judge had a patent background before appointment
to the Federal Circuit had no statistically significant effect on two of the three measured
outcomes (Petherbridge, 2009, p. 455). Similarly, when grouping judges by whether they
had a patent background before appointment to the court, Petherbridge (2009) failed to
reject the null hypotheses that there were no differences in the likelihood of the measured
outcomes when these judges heard the appeals (Petherbridge, 2009, pp. 452–54).
Finally, he broke down the analysis into particular sub-doctrines within the doctrine
of equivalents and found that judges’ methodology, appointing President, and patent
background sometimes had statistically significant effects on patentees’ likelihood of
success on appeal (Petherbridge, 2009, pp. 462–63). In sum, Petherbridge concluded that
doctrine of equivalent outcomes depend on judges and groups of judges (Petherbridge,
2009, p. 465).
As described earlier with respect to certainty, Wagner and Petherbridge’s (2004) study
of claim construction decisions from April 23, 1996 through November 1, 2002 shows
a division between procedural and holistic judges on the Federal Circuit (Wagner and
Petherbridge, 2004, pp. 1111–12). The addition of two strongly procedural judges in 2000
(Judges Linn and Dyk) caused a shift toward a more procedural approach at the court
(Wagner and Petherbridge, 2004, p. 1153). Using binary logistic regression analysis, they
found that particular judges participating on the panel had a statistically significant effect
on the claim construction approach used (Wagner and Petherbridge, 2004, pp. 1163–66).
Stiernberg (2013) conducted a study examining whether Federal Circuit judges with
technical backgrounds reversed district courts on claim construction more frequently

than nontechnical judges (Stiernberg, 2013, pp. 286–88). Importantly, Stiernberg (2013)

considered technical expertise by determining whether the judge had scientific or technical
expertise relevant to the claim at issue in the case (Stiernberg, 2013, p. 290). In this study,
Stiernberg collected all Federal Circuit claim construction cases from April 30, 2007
through April 30, 2012, including Rule 36 affirmances, and randomly chose 100 cases
for further examination, resulting in 159 claim construction instances (Stiernberg, 2013,
pp. 288–89).
The data show that when a Federal Circuit judge has technical expertise related to the
patent in suit, the judge is more likely to find error in the district court’s claim construction
(Stiernberg, 2013, p. 295). The predicted probability of a Federal Circuit judge reversing
on claim construction increases by 53.42 percent when the Federal Circuit judge has a
technical background related to the patent in suit (Stiernberg, 2013, p. 295). The data show
that the number of years a Federal Circuit judge has been on the bench has no impact
on the probability of reversal of the district court’s claim construction (Stiernberg, 2013,
p. 294). Also, differences in political ideology between district judges and Federal Circuit
judges are not significant (Stiernberg, 2013, p. 290).
Finally, Field (2014) measures judicial hyperactivity of Federal Circuit judges using
reversal rates, activism differentials, and scaled activism scores, which are categorized by
the different standards of review (Field, 2014, pp. 628–30). Judicial hyperactivity measures
the frequency appellate judges substitute their judgments for those of the district court
judge below (Field, 2014, p. 631). Field (2014) used 299 cases from the first half of 2010,
including 110 patent cases and 189 non-patent cases; these cases had 828 different issues
(Field, 2014, p. 644).
When examining patent cases and using simple reversal rates, Field (2014) shows that
there is a wide range of reversal rates among the Federal Circuit judges (Field, 2014,
pp. 656–57). Grouping all patent cases together (regardless of their standard of review),
Judge Plager was at one extreme with a 66.7 percent reversal rate, while Judge Mayer
had a 20 percent reversal rate (Field, 2014, pp. 656–57). In comparison, when examining
simple reversal rates in non-patent cases, Field (2014) shows a wide range, but it is much
lower than in the patent case (ranging from 6.1 percent to 36.8 percent) (Field, 2014,
pp. 660–61).
Using activism differentials and scaled activism scores, developed by Corey Yung,
Field (2014) uses de novo review reversal rates as a baseline to better compare judicial
activism (Field, 2014, p. 664). Using these methods, Field (2014) shows the level of judicial
hyperactivity for each judge for patent cases, with Judge Michel being the least hyperactive
and Judge Radar being the most (Field, 2014, pp. 668–69). In non-patent cases, Judge
Schall was the least hyperactive, while Judge Plager was the most (Field, 2014, pp. 672–74).
This study shows that Federal Circuit judges with patent backgrounds (i.e., who regularly
practiced patent law or had scientific or technical expertise) are more judicially hyperac-
tive than those without, but only in patent cases; these same judges are not hyperactive in
non-patent cases (Field, 2014, pp. 669, 673).
Field (2014) also analyzed the data to determine whether the party of the appointing
president affected the judges’ levels of activism (Field, 2014, pp. 682–85). Similarly to
what Petherbridge (2009) demonstrated for the doctrine of equivalents, the party of the
appointing president is not statistically significant when it comes to determining judicial
hyperactivity (Field, 2014, pp. 682–85).

Given the age of some of these studies and the numerous new appointments to the
court over the last few years, follow-up research can be undertaken to see what changes,
if any, have taken place to the composition of the judges on the court and whether these
changes have had an impact on specific doctrines or methodologies.
2. Turnover
Turnover of judges at the Federal Circuit also has the potential to disrupt patent law
development and its application. Judge Michel (2010) recounted that when he joined the
court, Chief Judge Markey told him that it usually takes five years for a new judge to
hit full stride (Michel, 2010, pp. 1204–05). In the last six years, there have been four new
judges appointed to the court. In the last nine years, seven of the 12 active judges on the
court have been added. Upon his retirement, Judge Michel (2010) noted that the most
dramatic development in the court’s evolution was that it faced a large, sudden change in
membership and that these new members would undoubtedly introduce uncertainty and
change in both the balance of the court’s members and possibly patent doctrine (Michel,
2010, pp. 1203–04). Judge Michel’s concern, as we saw earlier, has been with certainty
and predictability. He warned that sudden changes in precedent would upset expectations
of businesses and that this could cause harm to the broader economy (Michel, 2010,
pp. 1204–05). Whether it be from new jurisprudential views or inexperience, there is some
risk of upsetting the status quo (Michel, 2010, pp. 1204–05).
Bock (2014), however, suggests that turnover at the Federal Circuit could be a positive
development (Bock, 2014, pp. 227–36). Although he suggests having a pool of rotating
district court judges temporarily serving on the Federal Circuit, some of the same
principles would apply to turnover of full-time Federal Circuit judges. Such turnover may
help hasten the percolation process and lead to higher-quality doctrinal developments
(Bock, 2014, p. 236). Because of the recent spate of appointments, further research on the
impact of these new judges and their impact on specific doctrines and methods should
be undertaken.
3. Case distribution

The final area of interest with respect to individual judges is how case distribution affects
the development and application of patent law. Federal Circuit Rule 47.2(b) and Section
103(b)(3) of the Federal Courts Improvement Act of 1982 aim to provide each Federal
Circuit judge with a representative cross-section of the court’s jurisdiction (Froats, 2010,
pp. 80, 83). In theory, under this random assignment method, Federal Circuit judges
would, over time, sit on the same number of panels in the various fields the court has
jurisdiction over (Froats, 2010, p. 80). If, however, the system does not result in the judges
having a representative cross-section of cases, then some judges could be exercising more
influence on some areas (Froats, 2010, p. 80). To implement these rules, a computer in the
clerk’s office randomly generates three-judge panels (Froats, 2010, pp. 80, 83–84).
Froats’ (2010) study examines all patent appeals from the BPAI from 2005–09 to see
whether the randomization process results in the judges having a representative cross-
section of the BPAI appeals (Froats, 2010, pp. 81, 84–85, 88). The author notes that one
limitation of this study is that it fails to account for recusals and instances where judges
are unavailable because of illness, emergency, or designation to another court (Froats,
2010, p. 90). Furthermore, the author contends that patent prosecution is a distinct field

from patent litigation and should be examined separately (Froats, 2010, p. 99). It is not
clear whether the clerk’s office shares this view.
Froats’ (2010) data show that the current system results in great disparities in the
number of appeals from the BPAI heard by each judge (Froats, 2010, pp. 93–94). For
example, Judge Prost heard 29.1 percent of the BPAI appeals (32 panels), whereas Judges
Schall and Dyk only participated in 17.3 percent of these appeals (19 panels each) (Froats,
2010, p. 93). The author suggests that this disparity calls into question the success of the
randomization process (Froats, 2010, p. 94).
In measuring the effects of an uneven case distribution, Froats (2010) also studied the
subset of precedential opinions issued by the Federal Circuit during the relevant time
period (Froats, 2010, p. 105). If there is an uneven distribution of precedential cases,
then this could be problematic because certain judges would have a larger role in shaping
patent jurisprudence as it relates to prosecution (Froats, 2010, p. 105). The data show that
Judge Moore (controlling for the fact that she was appointed during the relevant time
period) participated in 31 percent of the appeals, Judge Newman participated in 30.8
percent, and Judges Michel, Gajarsa, and Prost participated in 28.2 percent (Froats, 2010,
p. 106). On the other end of the spectrum (not counting senior judges), Judges Lourie
and Schall participated in 12.8 percent and 15.4 percent of the cases, respectively (Froats,
2010, p. 106). The author notes that although it is possible for certain judges to have more
influence on patent law jurisprudence because of their participation on these panels, he
does not go so far to say that one particular judge has more influence than another based
on these results, or that the random assignment system leads to this (Froats, 2010, p. 107).
Finally, the author notes that it is likely impossible to know in advance which cases will
result in precedential opinions, so designing an assignment system to achieve these results
is unlikely (Froats, 2010, p. 107).
B. Internal Dynamics
The internal dynamics of the Federal Circuit may also affect the development of patent
law and whether the Federal Circuit achieves the goals it was established to create. Chief
Judge Markey (1992) noted that part of the Federal Circuit’s success in bringing uniform-
ity to patent law was that the judges unanimously agreed to subject their precedential
opinions to review by their colleagues and the Senior Technical Advisor (STA) (Markey,
1992, p. 578). This review was limited to language that appeared to create a conflict with
Federal Circuit precedent; it did not involve a review of the decision itself (Markey, 1992,
p. 578). Although this process added to each judge’s workload, Markey (1992) noted
that it was an example of the court’s dedication to its mission and resulted in its success
(Markey, 1992, p. 578). The effectiveness of peer and STA review depends on the willing-
ness of the non-panel judges to review and comment on draft opinions and the receptivity
of the judges on the panel (Bock, 2014, p. 207). The STA’s conflict check process has been
scaled back in recent years to only review opinions upon request (Bock, 2014, p. 207). The
effect of scaled-back STA review is an opportunity for further research, as are studies of
how well the peer review system works.
A concern has been raised that the Federal Circuit suffers from the “curse of expertise,”
which yields a resistance to correct misguided precedent (Bock, 2014, p. 216). In short,
because of their expertise in patent law, judges on the Federal Circuit approach their work

with overconfidence, which renders them unreceptive to reconsider their prior analyses
(Bock, 2014, p. 216). The rollback of the STA’s conflict check and limited attention to
scholarship may evidence this (Bock, 2014, p. 216), although Schwartz and Petherbridge
(2012) cast doubt on the latter issue.
Bock (2014) also points to the “knowing-doing gap” as a potential source for the
Federal Circuit’s misguided precedent (Bock, 2014, p. 219). In other words, the Federal
Circuit judges may well be aware that their precedent is suboptimal, but they have dif-
ficulty garnering support to break free from the precedent-setting institutional inertia
(Bock, 2014, pp. 219–20). The potential causes of the knowing-doing gap include judges
wanting to act consistently with their prior decisions, maintaining a collegial working
environment, and avoiding the costs of correction when there is no perceived urgency
(Bock, 2014, pp. 220–27). If these behaviors exist, then the effect is an ossification of
precedent (sometimes suboptimal).
To overcome the curse of expertise and the knowing-doing gap, Bock (2014) suggests
changing the internal dynamics of the Federal Circuit by using a group of rotating
district judges to serve on the court for staggered terms of limited duration (Bock, 2014,
pp. 227–29). Adding non-expert district judges adds new voices and could help the court
focus on suboptimal precedent and understand the difficulties this precedent presents
(Bock, 2014, p. 229). In addition, temporarily adding district judges could help overcome
the knowing-doing gap because there would be fewer prior opinions by these judges and
hence fewer opportunities to act inconsistently (Bock, 2014, pp. 232–23). Also, because
of the frequent turnover of these district judges, they may feel freer to “rock the boat”
to reevaluate prior precedent (Bock, 2014, pp. 233–34). Finally, these district court judges
may feel the urgency in correcting suboptimal precedent in circumstances where they
will be forced to grapple with the precedent upon their return to the district court (Bock,
2014, p. 235).
To be sure, the internal dynamics discussed here are preliminary observations and
theoretical arguments that may affect the development and application of patent law.
Empirical research to justify or reject these theories would be a welcome addition to the
literature.
C. En Banc Review
The final aspect of how the Federal Circuit’s structure and staffing affect its role as an
institution and the goals it was established to achieve is when the court hears and decides
cases en banc. Cotropia (2010) examined the Federal Circuit’s en banc rate in an effort
to determine whether the court was “willing to innovate new legal rules, percolate these
concepts, and then eventually adopt them in future decisions” (Cotropia, 2010, p. 822).
In his study, Cotropia (2010) measures the proportion of en banc reviews to written
decisions (Cotropia, 2010, p. 813). His dataset consisted of data from the AO reports
from 1998–2009 as well as a Westlaw search for the same period (Cotropia, 2010, p. 814).
His results show that the Federal Circuit decided cases en banc 0.18 percent of the time
(Cotropia, 2010, p. 817). He then compared this rate with those of the Third, Fifth,
Ninth, Tenth, and DC Circuits, all of which had higher en banc rates (0.21 percent, 0.21
percent, 0.38 percent, 0.32 percent and 0.24 percent, respectively) (Cotropia, 2010, p. 817).
Cotropia (2010) notes that although the Federal Circuit’s en banc rate was the lowest of

the selected circuits, the differences between it and the next three lowest circuits were
not statistically significant (Cotropia, 2010, p. 817). The differences between the Federal
Circuit and the Ninth and Tenth Circuits were statistically significant (Cotropia, 2010,
p. 817). Based on these results, Cotropia (2010) suggests that the Federal Circuit does not
exhibit much willingness to change the law through en banc review, but that it is not alone
(Cotropia, 2010, p. 822).
In examining how these results explain or impact percolation and entrenchment,
Cotropia (2010) offers a few suggestions. One is simply that although there is not uni-
formity within the court, the law remains entrenched (Cotropia, 2010, p. 822). Another
explanation could be that the disputes provoking dissenting opinions are minor and not
sufficiently important for en banc review (Cotropia, 2010, pp. 822–23). The final explana-
tion is that the Federal Circuit could be harmonizing the law without using the transpar-
ent en banc procedure, but instead uses the comment period on precedential opinions to
resolve potential conflicts (Cotropia, 2010, p. 823). As a result, the law does not become
entrenched; instead it evolves, but does so in a less transparent way than en banc review
provides (Cotropia, 2010, p. 823). Importantly, Cotropia’s (2010) study was not limited to
patent decisions; it included all areas under the Federal Circuit’s jurisdiction.
In contrast, Vacca (2011) examined all of the Federal Circuit’s en banc patent cases
decided from 1982–2010 (40 cases in total) and calculated the en banc rate as a propor-
tion of the court’s total patent cases (Vacca, 2011, pp. 736–38). Vacca’s results show the
Federal Circuit hears 0.29 percent of its patent cases en banc (Vacca, 2011, pp. 736–38).
This measure of the en banc rate is higher than Cotropia’s (2010) finding of en banc cases
generally (0.18 percent). When looking only at the en banc cases decided from 2001–09,
the Federal Circuit’s en banc rate is 0.30 percent or 0.23 percent, depending on whether
one includes cases ordered to be heard en banc, but not yet decided (Vacca, 2011, p. 738).
In comparison with all of the other circuits’ en banc rates for all cases, the Federal Circuit
had the highest rate (Vacca, 2011, p. 738).
Vacca (2011) also explores the process in which the Federal Circuit hears patent cases
en banc (Vacca, 2011, p. 739). First, he notes that the Federal Circuit has ordered cases
to be heard en banc sua sponte in at least 48 percent of the cases (Vacca, 2011, p. 739).
Second, he describes how the Federal Circuit freely permits amici curiae to file briefs in
its en banc cases without leave of the court or seeking consent from the parties (Vacca,
2011, p. 743). In some cases, the court specifically calls for the United States or USPTO to
file an amicus brief (Vacca, 2011, p. 743). A survey of en banc orders from other circuits
did not reveal such a practice (Vacca, 2011, p. 744). The Federal Circuit’s rationale for
soliciting the views of stakeholders is so it can make informed decisions on how to shape
and interpret patent law, taking multiple viewpoints and interests into consideration
(Vacca, 2011, p. 744).
Given these practices, and the broad scope of the questions the Federal Circuit seeks to
answer when it sits en banc, Vacca (2011) argues that the Federal Circuit’s en banc process
is similar to an administrative agency engaging in substantive rulemaking (Vacca, 2011,
p. 744). Just like an administrative agency issues a notice of proposed rulemaking, the
Federal Circuit issues an order for rehearing en banc, setting forth when it will hear the
case and what issues it will examine (Vacca, 2011, pp. 747–48). And similar to an agency
giving interested parties an opportunity to participate in the rulemaking process by sub-
mitting their views or arguments, the Federal Circuit’s practice of freely permitting amici

to file briefs mimics this commenting process (Vacca, 2011, p. 749). Finally, an administra-
tive agency must publicize the final rule and explain its basis and purpose (Vacca, 2011,
p. 749). The Federal Circuit engages in similar conduct when it issues its en banc opinion
announcing its rule (Vacca, 2011, p. 749). Vacca (2011) concludes that although other
institutions can and should play a role in shaping patent policy, the Federal Circuit seems
willing to do so and does so in an open and inclusive manner (Vacca, 2011, pp. 757–58).
Finally, Taylor (2013) also briefly examines the Federal Circuit’s en banc decisions
as they relate to formalism. Taylor (2013) suggests that when the court decides cases en
banc, it oftentimes does so in a way that analyzes the basic underlying purpose of the
rule, but also considers innovation policy (Taylor, 2013, pp. 638, 659–64). But as with
most general statements, there are exceptions and the Federal Circuit’s use of policy in
en banc decisions is no different (Taylor, 2013, pp. 653–54, 656–57). Despite the general
practice of using policy in en banc decisions, Taylor suggests the Federal Circuit could do
a better job incorporating policy into its decisions (Taylor, 2013, p. 681). One suggestion is
encouraging amici to file briefs relating to policy concerns in en banc cases (Taylor, 2013,
p. 681). Vacca’s (2011) review of en banc orders indicates that the court is doing exactly
this (Vacca, 2011, p. 743).
As noted earlier, to complement the Federal Circuit’s current en banc practices, Laura
Pedraza-Fariña (2015) suggests that the court make use of advisory panels on issues of
patent law and policy (Pedraza-Fariña, 2015, pp. 154–59). These advisory panels, with a
mixture of technological, economic, and sociological expertise, could advise the Federal
Circuit on areas of the law in need of clarification and provide input on how a decision
may impact innovation in particular fields (Pedraza-Fariña, 2015, pp. 154–55).
VII. RELATIONSHIPS WITH OTHER INSTITUTIONS
As John Duffy (2004) has pointed out, when studying the success or failure of the Federal
Circuit as an institution, we cannot look at the Federal Circuit in isolation (Duffy, 2004,
p. 806). Instead, “a complete analysis of any institution encompasses that institution’s
relation to other relevant actors” (Duffy, 2004, p. 806). Therefore, to fully understand
the Federal Circuit, we must consider the other institutions that play a role in the patent
system (Duffy, 2004, p. 806).
This section examines the Federal Circuit’s interactions with other institutions and, in
appropriate situations, examines which institutions are and should be performing what
role in the development of patent law. In particular, this section looks at the historical
disinterest and recent hyper-interest by the Supreme Court, its treatment of Federal
Circuit opinions, and the Federal Circuit’s response. This section explores the role of the
USPTO and its relationship with the Federal Circuit. Next, this section looks at the role
district courts can and should play in the framework of patent law. The following subsec-
tion briefly discusses the role of the Solicitor General and how this office has affected
the Federal Circuit’s development of patent law. Finally, this section concludes with an
examination of the role Congress has and should play in the development of patent policy
and how the Federal Circuit has fit into this relationship.
As will be evident, unlike the rest of this chapter, the sections below do not focus on
prior empirical studies. Instead, the scholarship on the interaction of these institutions

with the Federal Circuit largely comprises theoretical and descriptive accounts. There is a
dearth of sophisticated empirical literature supporting or rejecting these theories.
A. Supreme Court
The Supreme Court has, over the last ten to 15 years, had an increased interest in patent
law (Dreyfuss, 2010, pp. 829–30). Duffy (2010) illustrates that prior to the mid-twentieth
century, the Supreme Court had a relatively large number of patent cases on its docket
in any given year (Duffy, 2010, p. 520). But starting around 1950, the Supreme Court’s
interest in patent cases dropped dramatically (Duffy, 2010, p. 520). From 1950–82, the
Supreme Court heard, on average, just over one patent case per term (Duffy, 2010, p. 522).
Immediately after the Federal Circuit’s creation, the Supreme Court showed even less
interest in patent law, hearing only five cases in the first 12 years (Duffy, 2010, p. 522; Lee,
2010, p. 42). Moreover, four of the five cases did not involve substantive patent rules, but
were procedural or jurisdictional issues (Duffy, 2010, p. 522). This lack of Supreme Court
intervention in patent cases arguably showed deference to the Federal Circuit’s expertise
in the field (Lee, 2010, p. 42).
But starting in 1994, the Supreme Court began to show more interest in patent law
(Duffy, 2010, p. 523; Lee, 2010, p. 43). From 1995–98, the Court decided five cases and all
but one involved important patent policy matters (Duffy, 2010, pp. 523–24). And over the
next ten terms (1999–2008), the Court heard 11 more patent cases (Duffy, 2010, p. 524). At
this point, the Court had returned to its pre-Federal Circuit rate of hearing patent cases
(Duffy, 2010, p. 524). Recently, the Supreme Court has been highly interested in patent
cases. From the 2009 term through the 2017 term, the Supreme Court decided 31 patent
cases (Written Description, Supreme Court Patent Cases—http://writtendescription.
blogspot.com/p/patents-scotus.html).
And because of this increased interest in patent cases, a struggle over which i nstitution—
the Supreme Court or Federal Circuit—is best suited to manage patent jurisprudence has
arisen. Over the last ten to 15 years, the Supreme Court has indicated that it is concerned
with the Federal Circuit’s performance (Dreyfuss, 2008, p. 791). During this time, the
Supreme Court has reversed or vacated the Federal Circuit’s decision in virtually all of the
cases (Dreyfuss, 2008, p. 791; Dreyfuss, 2010, pp. 829–30). What appears to be occurring
is that the Federal Circuit has posited itself as an expert institution in patent law and
has become disobedient in following Supreme Court precedent (Pedraza-Fariña, 2015,
p. 124). As Laura Pedraza-Fariña (2015) notes:
There have been no quantitative empirical studies comparing Federal Circuit disobedience of
Supreme Court decisions to disobedience by other circuits, or assessing whether the Federal
Circuit is more likely to defy the Supreme Court in its attributed area of expertise (patent law)
than in any of the other cases that make up its docket . . . [but] qualitative evidence suggests that
this is the case (Pedraza-Fariña, 2015, p. 124).
Moreover, as mentioned earlier, Federal Circuit is perceived as being too formalistic

and this may be another source of contention between the two courts (Lee, 2010, p. 46).
Regardless of the cause of this tension, as Peter Lee concludes, “[t]he Supreme Court’s
deference to Federal Circuit jurisprudence, as well as its general indifference to patent
matters, appears to have ended” (Lee, 2010, p. 43).

Of course, by institutional design, the Supreme Court has the final say on patent issues.
But there are several limitations to the Supreme Court serving as the steward of patent
law. First, the Court cannot intervene too frequently because of the limited number of
cases it hears each term (Dreyfuss, 2010, p. 839). In essence, hearing a small handful
of cases would not allow the Court to effectively be the steward of patent law. Second,
and related to the first, is that the Court does not hear enough patent cases to develop
expertise in the subject matter (Dreyfuss, 2008, p. 806; Dreyfuss, 2010, p. 839). Although
the Supreme Court has shown a heightened interest in patent cases over the last several
years, hearing three to six patent cases per year seems incomparable to the hundreds of
patent cases the Federal Circuit hears each year. Moreover, because the Federal Circuit
hears all patent appeals, the Supreme Court cannot rely on the experience of different
circuit courts to guide its rule-making and policy-setting (Dreyfuss, 2008, p. 808).
Notwithstanding these limitations, it is not clear that the Supreme Court is as anti-
formalistic as the Federal Circuit is formalistic. As described earlier, Taylor (2013) defends
the Federal Court against charges that it suffers from an excessive level of formalism and
illustrates several examples where the court engages in policy analysis, especially in en
banc and dissenting opinions (Taylor, 2013, pp. 638, 657–63).
Taylor also evaluates the Supreme Court’s perceived anti-formalism in patent cases.
His analysis reveals that the Supreme Court has a checkered history of policy analysis in
patent cases (Taylor, 2013, p. 638). He lists a host of patent cases over the last decade or
more where the Supreme Court’s analyses did not rest on policy considerations, including
J.E.M. Ag Supply, Inc. v. Pioneer Hi-Bred International, Inc., 534 U.S. 124 (2001) (relying
on precedent and other governing statutes); eBay, Inc. v. MercExchange, LLC (2006) (rely-
ing on “well-established principles of equity”); Quanta Computer, Inc. v. LG Elec., Inc.,
553 U.S. 617 (2008) (relying on precedent); Kappos v. Hyatt, 566 U.S. 431 (2012) (relying
on precedent); Microsoft Corp. v. i4i Ltd. P’ship, 564 U.S. 91 (2011) (relying on precedent);
Board of Trustees of Leland Stanford Junior University v. Roche Molecular Systems, Inc.,
563 U.S. 776 (2011) (relying on statutory and case law); and Global-Tech Appliances, Inc.
v. SEB S.A., 563 U.S. 754 (2011) (relying on “the long history of willful blindness and its
wide acceptance in the Federal Judiciary”).
Nonetheless, Taylor does acknowledge that the Supreme Court frequently uses policy
in its analyses of patent cases (Taylor, 2013, p. 666). For example, the Court in Merck
KGaA v. Integra Lifesciences I, Ltd., 545 U.S. 193 (2005) interpreted the statutory text in
light of the purposes of the statute and regulations and the realistic effect of the narrow
interpretation of the rule by the Federal Circuit (Taylor, 2013, pp. 666–67). Likewise, in
Pfaff v. Wells Electronics, Inc. (1998), the Court interpreted the on-sale bar rule in the
context of the “basic policies underlying the statutory scheme” (Taylor, 2013, p. 667).
The Supreme Court’s use of policy considerations is even more prevalent in areas where
the statutory language is more open-ended (e.g. non-obviousness and patentable subject
matter) or non-existent (doctrine of equivalents) (Taylor, 2013, pp. 638, 667–73).
Because patent law, as currently structured, involves a balancing of rules, standards,
incentives, and exemptions, policy is necessarily implicated. And, as Taylor posits, regard-
less of whether the Federal Circuit wants to engage in the balancing act or not, the balanc-
ing act is ongoing (Taylor, 2013, p. 674). As such, he encourages the Federal Circuit and
the Supreme Court to engage in more policy discussion and create an ongoing dialogue
between the courts (Taylor, 2013, pp. 675–82). Doing so may also help solve the problem

that patent law is not percolating at an optimal rate because the Federal Circuit does not
“compete” with any other circuits (Nard and Duffy, 2007, pp. 1620–21, 1644, 1645).
One way of fostering this dialogue and having the Supreme Court play a significant
role in the percolation process is for the Supreme Court to view itself not as the “final law
sayer,” but as the “prime percolator” (Golden, 2009, p. 662). Because the Federal Circuit
has a wealth of experience with patent law, its contributions to the development of patent
doctrine and administration are invaluable (Golden, 2009, p. 660). If the Supreme Court
decided cases in this “prime percolator” role, it would only provide partial direction in
its opinions and leave the lower courts and USPTO to flesh out the details and allow for
continued growth (Golden, 2009, p. 663). In particular, Golden suggests that the Supreme
Court consider three traits that indicate if a case is good for review on the merits:
(1) the substantive question involved is not currently subject to meaningful debate in the courts
below;
(2) there is good reason to suspect that the Federal Circuit’s settled approach to that question
is substantially inferior to a legally permissible alternative; and
(3) the case at hand is a good vehicle for addressing the substantive question as part of
determining the outcome of a dispute between the specific parties involved (Golden, 2009,
pp. 709–10).
And although the Supreme Court could undertake this approach on its own, Golden
urges the Federal Circuit to work with the Supreme Court to help percolate patent law
by not writing or reading opinions broadly and hearing more cases en banc (Golden,
2009, p. 717). As noted above, the Supreme Court has become more active in patent
cases over the last decade or so and the Federal Circuit has been active in hearing cases
en banc (Vacca, 2011). Additional study of whether these developments mimic Golden’s
suggestions and whether they have impacted percolation and entrenchment should be
undertaken.
B. USPTO
Some have argued that a major problem with patent law is that the Federal Circuit is not
in the best position to develop the law, and that instead the USPTO and its staff of patent
specialists would be better (Dreyfuss, 2004, pp. 791–92; Rai, 2003, p. 1065). But there are a
couple of major limitations on the USPTO serving in this role. First, the USPTO may not
do a better job than the Federal Circuit because of a lack of resources and the lack of a
longstanding culture managing the law (Dreyfuss, 2004, pp. 792–93). Second, unlike many
administrative agencies, the USPTO has no substantive rulemaking authority (Dreyfuss,
2004, p. 793; Dreyfuss, 2010, p. 838; Gugliuzza, 2013, p. 1820).
Even so, the USPTO’s role in granting patents and adjudicating challenges to patent-
ability gives it some potential power. However, the Federal Circuit has given minimal def-
erence to the USPTO’s fact-finding, which weakens the USPTO’s influence (Gugliuzza,
2013, pp. 1820–22; Rai, 2003, p. 1065). Moreover, the Federal Circuit enhances its power
by minimizing the USPTO’s role and autonomy by deciding questions on appeal that
were not addressed by the USPTO and then remanding the case for further consideration
(Gugliuzza, 2013, p. 1822). As Pedraza-Fariña (2015) explains, these are examples of an
expert institution seeking to maintain control over a body of knowledge (Pedraza-Fariña,

2015, pp. 109, 131). But this is not always the case. There are instances where the Federal
Circuit has given weight to the USPTO’s longstanding internal practices that do not have
the force of law (Gugliuzza, 2013, p. 1824).
But the AIA’s adoption of more meaningful ways to administratively challenge issued
patents, such as post-grant review, inter partes review, and covered business methods
review, may give the USPTO additional power to interpret and shape patent law and
policy (Gugliuzza, 2013, pp. 1825–27; Dreyfuss, 2015, pp. 235–42). Because these new
procedures at the Patent Trial and Appeal Board (PTAB) are cheaper and quicker, they
seem to be attractive options for patent stakeholders (Dreyfuss, 2015, p. 239). And because
of the timing, the PTAB is likely to be the first institution to implement Supreme Court
law and consider substantive changes under the AIA (Dreyfuss, 2015, p. 240).
Further, because PTAB judges have significant expertise in patent law and there is an
effort to have at least one judge with expertise in the field of invention on each panel, the
PTAB has the potential to contribute to the development of patent law policy (Dreyfuss,
2015, pp. 240, 261). For example, Rochelle Dreyfuss (2015) has taken a look at initial
PTAB cases and argues that the PTAB’s efforts have been successful at laying out a frame-
work for analyzing issues related to definiteness and patentable subject matter in light of
the Supreme Court’s pronouncements in Nautilus, 134 S. Ct. 2120 (2014) and Alice Corp.
Ltd. v. CLS Bank Int’l, 134 S. Ct. 2347 (2014) (Dreyfuss, 2015, pp. 261–62). In contrast, the
Federal Circuit has struggled with interpreting these cases (Dreyfuss, 2015, pp. 262–63).
Dreyfuss (2015) also points out that the PTAB decisions could go a long way in chang-
ing the nature of the Federal Circuit’s jurisprudence (Dreyfuss, 2015, pp. 266–67). For
example, one explanation for the Federal Circuit’s rigid and formal analyses is that the
court is trying to enhance predictability and uniformity (Dreyfuss, 2015, p. 266). However,
the Supreme Court has routinely rejected these rigid rules (Dreyfuss, 2015, p. 266).
Dreyfuss (2015) argues that because the PTAB is providing detailed and technology-
specific analyses, bright-line rules and less deference may be less necessary (Dreyfuss,
2015, p. 266).
Despite its potential, Dreyfuss (2015) recognizes some limitations. One such limitation
is that the PTAB, like the Federal Circuit, does not frequently refer to underlying policy in
deciding cases (Dreyfuss, 2015, p. 271). Engaging with underlying policies could facilitate
the dialogue with the Federal Circuit and eventually the Supreme Court (Dreyfuss, 2015,
p. 271). Of course, given the hierarchical structure of these institutions, the ability of the
PTAB to play an important role in shaping patent law and policy depends on the Federal
Circuit’s willingness to engage in a meaningful dialogue with the PTAB (Dreyfuss, 2015,
pp. 277–78). Because the PTAB and its new processes are still in their early formative
years, this is an area ripe for further theoretical and empirical study.
C. District Courts
As in most areas of the law, facts play a critical role in patent law (Rai, 2002, p. 881).
In patent law, facts affect assessments of infringement, validity, and patent scope (Rai,
2002, p. 881). And the Federal Circuit’s purported expertise involves, to a large extent,
comprehending complex technologies covered in patents, the prior art, and the allegedly
infringing protects or processes, and being able to assess the relationship between them
(Dreyfuss, 2008, p. 802). Much of this turns on factual determinations (Rai, 2002, p. 881).

But because it is an appellate court, the Federal Circuit is, in theory, required to give
deference to district courts on factual matters (Dreyfuss, 2008, pp. 802–03; Gugliuzza,
2013, p. 1831). But similar to what was shown with respect to the USPTO, the Federal
Circuit’s review of district court decisions contravenes the traditional analysis (Rai, 2003,
p. 1065). To overcome what the Federal Circuit sees as a deferential hurdle, it has recast
some issues into questions of law (Dreyfuss, 2008, pp. 802–03; Rai, 2002, pp. 885–86;
Gugliuzza, 2013, pp. 1831–32; Pedraza-Fariña, 2015, pp. 98–102, 122–23). And when the
Supreme Court has occasionally stepped in to stop this practice, the Federal Circuit has
responded by requiring that district courts engage in a particular analytical approach
(Dreyfuss, 2008, pp. 802–03). These practices fall in line with Pedraza-Fariña’s (2015)
theory that the Federal Circuit behaves like a weak expert institution—maintaining
control over a body of knowledge by, inter alia, rejecting what others do and how they do
it (Pedraza-Fariña, 2015, p. 109).
One concern with the Federal Circuit wrestling fact-finding control away from the
district courts is that the Federal Circuit may not be any better at this because the Federal
Circuit’s expertise is in patent law, not the factual complexities of a particular technology
(Rai, 2002, p. 888; Rai, 2003, pp. 1088–89). Another concern is that even if the Federal
Circuit is a superior factfinder, having it perform this function is inefficient (Rai, 2003,
pp. 1089–90). As Rai (2003) explains, both trial court and Federal Circuit resources are
“expended on questions that will typically matter in only a single case” and certainty
about the patent rights is delayed (Rai, 2003, pp. 1089–90).
Despite these concerns, it is questionable whether district judges do a better job (Rai,
2002, p. 889). To address this problem, Rai (2002, 2003) suggests specialized patent trial
courts (Rai, 2002, p. 894; Rai, 2003, pp. 1065–66, 1099–101). To a certain extent, some
district courts have become de facto specialized courts by hearing a disproportionate
amount of patent cases (Kesan and Ball, 2011; Gugliuzza, 2013, p. 1839).
Kesan and Ball (2011) undertook a study to test whether increasing patent specializa-
tion resulted in “better” adjudication—that is, more accurate and efficient patent litiga-
tion (Kesan and Ball, 2011, p. 396). In their study, Kesan and Ball (2011) first looked to
see how patent litigation is concentrated (Kesan and Ball, 2011, p. 420). They find that
patent litigation is highly concentrated, but “most patent cases are still presided over by
judges with little or no patent experience” (Kesan and Ball, 2011, p. 420).
Next, they examine how judicial experience (patent and otherwise) relates to the
efficiency of handling patent cases (Kesan and Ball, 2011, p. 420). They find that judicial
experience “seem[s] to reduce case duration, thereby increasing the efficiency with which
patent cases are adjudicated” (Kesan and Ball, 2011, p. 420). In particular, they find that
recent patent experience (three years or less before the case is filed) has a bigger impact
than total patent experience, but the impact is only moderate (Kesan and Ball, 2011,
p. 443).
Finally, Kesan and Ball (2011) ask how experience (patent and otherwise) relates to
accuracy in patent cases (as measured by the probability of being overturned on appeal)
(Kesan and Ball, 2011, p. 420). They find that more experience with patent cases does
increase the accuracy of patent rulings (Kesan and Ball, 2011, p. 420). In particular,
they “find that the probability that a case will be at least partially reversed is related to
the specialized patent experience of the district court judge”, and that “general judicial
experience has no impact on the reversal rate” (Kesan and Ball, 2011, pp. 443–44). Drilling

down a bit more into specific patent issues, Kesan and Ball (2011) show that recent and
cumulative patent experience reduces the probability of reversal on grounds other than
claim construction, and that recent patent experience reduces the probability of reversal
on claim construction (Kesan and Ball, 2011, pp. 443–44). However, their results also
show that cumulative patent experience has no impact on the probability of reversal on
claim construction (Kesan and Ball, 2011, pp. 443–44).
Congress’s creation in the AIA of a pilot specialized trial judge program may comple-
ment the de facto specialization taking place at the trial court level (Kesan and Ball,
2011, p. 396) and could help alleviate some of the institutional deficiencies identified
by scholars. And although benefits may flow from specialized trial courts or judges, the
Federal Circuit’s recent willingness to transfer cases from specialized districts, such as the
Eastern District of Texas, could be seen as an effort to “retain its status as the only expert
patent court” (Gugliuzza, 2013, p. 1839) and potentially undermines the patent system of
these benefits. Further research into these practices and the power struggle between the
Federal Circuit and the district courts is warranted.
D. Solicitor General
As described earlier in this section, the Supreme Court began its renewed interest in
patent cases in 1994. And although the merits of the Court’s 1994 case were not of much
importance, Duffy (2010) argues that this case was important because it marks the first
time the Court had issued an invitation for the Solicitor General to file an amicus brief
on whether the Court should grant certiorari (also known as CVSG orders) (Duffy, 2010,
p. 525).
Duffy’s data shows that the Supreme Court had used CVSG orders fairly regularly since
the 1960s and hit its peak in the 1980s (Duffy, 2010, pp. 525–26). And starting in 1999, the
Court began to use them in patent cases more regularly (Duffy, 2010, p. 526). However,
during the interim five-year period, the Solicitor General argued as an amicus or a party
in four of the five cases (Duffy, 2010, p. 528). In these cases, the Solicitor General fared
well, with the Court adopting its position or arguments in most of the cases (Duffy, 2010,
p. 528).
But from 2000–09, the Solicitor General’s role in patent cases grew stronger (Duffy,
2010, p. 529). During this period, the Court issued 19 CVSG orders in patent or patent-
antitrust cases (Duffy, 2010, p. 530). These CVSG orders comprised 10 percent of the
Court’s total CVSG orders during this period (Duffy, 2010, p. 530).
The importance of the Solicitor General’s role in patent cases stems from the fact that
the Supreme Court traditionally has relied on circuit splits to decide whether to grant
certiorari (Duffy, 2010, p. 530). Given the Federal Circuit’s exclusive jurisdiction over
patent cases, this was not a likely indicator of the importance of the issue (Duffy, 2010,
p. 530). And although the Court has used split decisions when the Federal Circuit decides
cases en banc, the other proxy it uses to decide whether a case warrants review is to rely on
CVSG orders (Duffy, 2010, p. 536). In short, the Supreme Court has come to rely on the
Solicitor General to advise it on the worthiness of patent cases (Duffy, 2010, pp. 536–37).
From 2000–09, the Solicitor General fared well in having its position adopted by
the Supreme Court in patent cases (Duffy, 2010, p. 538). In only one case, eBay, Inc. v.
MercExchange, L.L.C. (2006), did the party supported by the Solicitor General lose the

case (Duffy, 2010, p. 538). And even in eBay, the Court adopted the Solicitor General’s
analysis (Duffy, 2010, p. 538). Comparing the success rate of the Solicitor General to the
Federal Circuit’s is also telling (Duffy, 2010, p. 540). Duffy shows that since 1996, the
Solicitor General participated in 13 patent cases and supported a different result than
the Federal Circuit in nine of those (Duffy, 2010, p. 540). In all nine of those cases, the
Supreme Court agreed with the Solicitor General’s position over the Federal Circuit’s
(Duffy, 2010, p. 540). And in two of the cases where the Solicitor General agreed with the
Federal Circuit’s end result, but disagreed with the reasoning, the Supreme Court adopted
a rule closer to the Solicitor General’s position than the Federal Circuit’s (Duffy, 2010,
p. 540).
The Supreme Court’s increased reliance on the Solicitor General has given it increased
power to help formulate patent law (Duffy, 2010, p. 543). The importance of this develop-
ment has been that the creation of the Federal Circuit may have inadvertently shifted the
power to shape patent law and policy away from the judiciary and toward the executive
branch (Duffy, 2010, p. 543).
Duffy’s study of the role of the Solicitor General in shaping patent law and policy
is now several years old and the Supreme Court has continued its increased interest
in patent law during this time. Not only could a researcher conduct a follow-up study
to determine whether the Solicitor General’s role has remained prominent and success
rate has remained high, but further analysis could be undertaken to see how these rates
compare with other fields of law.
E. Congress
Although Congress has largely delegated patent policy-making to the courts (Rai, 2003,
pp. 1040–41) and the Federal Circuit has staked out a claim as the expert institution
(Pedraza-Fariña, 2015), Congress clearly has the power to change patent law and policy. If
Congress does not like what the Federal Circuit is doing, it can change the law (Dreyfuss,
2010, p. 840; Gugliuzza, 2013, p. 1827).
But there are concerns that Congress is an ineffective institution when it comes to
patent policy. First, there is a concern that Congress has not been interested enough in
patent policy (Dreyfuss, 2010, p. 840; Anderson, 2014, p. 965). From 1952–2011, there
were more than 50 years of disinterest, followed by several years of failed attempts at
reform (Dreyfuss, 2010, p. 840; Anderson, 2014, p. 965). It was not until 2011, upon
passage of the AIA, that Congress mustered sufficient political will to enact real change.
Second, scholars argue, even if political will exists, intervention by Congress is undesir-
able (Rai, 2003, pp. 1127–29). As Arti Rai points out, innovation is unpredictable, so
detailed statutes enacted by Congress will often be wrong (Rai, 2003, p. 1127). And once
codified, these laws become intractable and only perpetuate the problems (Rai, 2003,
p. 1127). In addition, policy-making by Congress is subject to opportunistic behavior
(Rai, 2003, p. 1130; Anderson, 2014, p. 966).
Notwithstanding these concerns, scholars have recently begun to illustrate the interesting
relationship between Congress and the Federal Circuit (Anderson, 2014; Gugliuzza, 2013,
pp. 1827–28). Rather than writing detailed statutes to effectuate change, Congressional
hearings on patent law have acted as catalysts to spur the Federal Circuit into action to
revisit certain aspects of its patent jurisprudence (Anderson, 2014, pp. 966–67; Gugliuzza,

2013, pp. 1827–28). Jonas Anderson (2014) and Paul Gugliuzza (2013) describe how
Congress began seriously examining patent reform in 2005, and that the Federal Circuit
responded to several of these legislative proposals by publicly commenting on them,
exercising its infrequently used discretionary powers, urging the bar to appeal these issues
to the Federal Circuit, and ultimately altering its precedent to moot parts of the legislation
(Anderson, 2014, pp. 967–68; Gugliuzza, 2013, pp. 1827–28). Anderson (2014) argues that
playing a catalytic role, which drives the Federal Circuit into action, permits Congress to
focus its reform efforts on areas where the Federal Circuit has little or no control, such as
administrative reform (Anderson, 2014, p. 1006).
Despite the claim that this relationship between Congress and the Federal Circuit has
the potential to strike an appropriate balance between competing policy considerations,
this potential rests on the assumption that Congress will continue overseeing patent
policymaking (Anderson, 2014, p. 1018). There has been a decade of such oversight, but
future scholarship may want to continue monitoring this trend and the Federal Circuit’s
responses to see whether it persists and to determine whether patent policy is optimized
by this relationship.
VIII. CONCLUSION
In a sense, the Federal Circuit has been an experiment in institutional design (Duffy,
2004, p. 803). Thus, it is fitting that this court has been the subject of an extraordinary
amount of empirical study to determine whether the experiment has been a success or
failure. As this chapter has illustrated, we still do not have a clear answer to that broad
question. Given that the goals of the Federal Circuit sometimes work at cross purposes
and that there are a multitude of variables affecting the development of patent law and
the underlying policies, it is doubtful that this question can ever be answered with finality.
Nonetheless, the empirical studies of the Federal Circuit serve an important purpose.
We may never be able to design a perfectly functioning system to administer patent law.
In fact, we may not even be able to design a perfectly functioning court to resolve patent
appeals. However, continued research into the effectiveness of particular areas and how
those areas interact with others and the underlying policies can help those charged with
designing and implementing the system to do so with a grounded understanding of what
the strengths, weaknesses, and challenges are.
The extensive and sophisticated empirical scholarship on the Federal Circuit is a great
start. But there is much more work to be done. The addition of several new judges at the
Federal Circuit, several years of a Supreme Court with an active interest in patent law,
Congressional action to reform patent law, and changes in procedures and policies at the
USPTO will drive the need to re-evaluate previous studies, undertake completely new
analyses, and determine what changes, if any, are necessary.
REFERENCES
Allison, John R. and Mark A. Lemley. 2000. ‘How Federal Circuit Judges Vote in Patent Validity Cases’, 27
Florida State University Law Review 745–66.

Anderson, Jonas. 2014. ‘Congress as a Catalyst of Patent Reform at the Federal Circuit’, 63 American University
Atkinson, Scott E., Alan C. Marco, and John L. Turner. 2009. ‘The Economics of a Centralized Judiciary:
Uniformity, Forum Shopping, and the Federal Circuit’, 52 Journal of Law & Economics 411–43.
Barnes, Dunstan H. 2013. Comment, ‘Technically Speaking, Does it Matter? An Empirical Study Linking
the Federal Circuit Judges’ Technical Backgrounds to How They Analyze the Section 112 Enablement and
Written Description Requirements’, 88 Chicago-Kent Law Review 971–1012.
Bock, Jeremy W. 2014. ‘Restructuring the Federal Circuit’, 3 New York University Journal of Intellectual Property
& Entertainment Law 197–249.
Burk, Dan L. and Mark A. Lemley. 2003. ‘Policy Levers in Patent Law’, 89 Virginia Law Review 1575–696.
Cotropia, Christopher A. 2010. ‘Determining Uniformity Within the Federal Circuit by Measuring Dissent and
En Banc Review’, 43 Loyola Los Angeles Law Review 801–25.
Dreyfuss, Rochelle Cooper. 1989. ‘The Federal Circuit: A Case Study in Specialized Courts’, 64 New York
Dreyfuss, Rochelle Cooper. 2004. ‘The Federal Circuit: A Continuing Experiment in Specialization’, 54 Case
Western Reserve Law Review 769–801.
Dreyfuss, Rochelle Cooper. 2008. ‘In Search of Institutional Identity: The Federal Circuit Comes of Age’, 23
Dreyfuss, Rochelle Cooper. 2010. ‘The Federal Circuit as an Institution: What Ought We to Expect?’, 43 Loyola
Los Angeles Law Review 827–42.
Dreyfuss, Rochelle Cooper. 2015. ‘Giving the Federal Circuit a Run for its Money: Challenging Patents in the
PTAB’, 91 Notre Dame Law Review 235–300.
Duffy, John F. 2004. ‘Comment, Experiments After the Federal Circuit’, 54 Case Western Reserve Law Review
803–10.
Dunner, Donald R. 2010. ‘The U.S. Court of Appeals for the Federal Circuit: Its Critical Role in the Revitalization
of U.S. Patent Jurisprudence, Past, Present, and Future’, 43 Loyola Los Angeles Law Review 775–84.
Field, Ted L. 2014. ‘Hyperactive Judges: An Empirical Study of Judge-Dependent “Judicial Hyperactivity” in
the Federal Circuit’, 38 Vermont Law Review 625–86.
Froats, Thomas. 2010. ‘Unrepresentative Randomization: An Empirical Study of Judging Panels of USPTO
Appeals to the CAFC’, 19 Texas Intellectual Property Law Journal 79–110.
Fromer, Jeanne C. 2010. ‘Patentography’, 85 New York University Law Review 1444–519.
Golden, John M. 2009. ‘The Supreme Court as “Prime Percolator”: A Prescription for Appellate Review of
Questions in Patent Law’, 56 UCLA Law Review 657–724.
Gruner, Richard S. 2010. ‘How High is Too High?: Reflections on the Sources and Meaning of Claim
Construction Reversal Rates at the Federal Circuit’, 43 Loyola Los Angeles Law Review 981.
Gugliuzza, Paul R. 2012. ‘Rethinking Federal Circuit Jurisdiction’, 100 Georgetown Law Journal 1437–505.
Gugliuzza, Paul R. 2013. ‘The Federal Circuit as a Federal Court’, 54 William & Mary Law Review 1791–864.
Henry, Matthew D. and John L. Turner. 2006. ‘The Court of Appeals for the Federal Circuit’s Impact on Patent
Litigation’, 35 Journal of Legal Studies 85–117.
Hill, Russell B. and Frank P. Cote. 2002. ‘Ending the Federal Circuit Crapshoot: Emphasizing Plain Meaning
in Patent Claim Interpretation’, 42 IDEA 1–26.
Kesan, Jay P. and Gwendolyn G. Ball. 2011. ‘Judicial Experience and the Efficiency and Accuracy of Patent
Adjudication: An Empirical Analysis of the Case for a Specialized Patent Trial Court’, 24 Harvard Journal
Lee, Peter. 2010. ‘Patent Law and the Two Cultures’, 120 Yale Law Journal 2–82.
Lefstin, Jeffrey A. 2007. ‘The Measure of the Doubt: Dissent, Indeterminacy, and Interpretation at the Federal
Circuit’, 58 Hastings Law Journal 1025–94.
Lefstin, Jeffrey A. 2010. ‘The Constitution of Patent Law: The Court of Customs and Patent Appeals and the
Shape of the Federal Circuit’s Jurisprudence’, 43 Loyola Los Angeles Law Review 843–92.
Lourie, Alan D. 2006. ‘Keynote Address at the Joint Patent Practice Seminar of the Connecticut, New York, New
Jersey, and Philadelphia Intellectual Property Law Associations’, 72 Patent, Trademark & Copyright Journal 41.
Lunney, Glynn S. 2004. ‘Patent Law, the Federal Circuit, and the Supreme Court: A Quiet Revolution’, 11
Markey, Howard T. 1992. ‘The Federal Circuit and Congressional Intent’, 41 American University Law Review
577–9.
Michel, Paul R. 1994. ‘The Challenge Ahead: Increasing Predictability in Federal Circuit Jurisprudence for the
New Century’, 43 American University Law Review 1231–57.
Michel, Paul R. 2010. Afterword, ‘Past, Present, and Future in the Life of the U.S. Court of Appeals for the
Federal Circuit’, 59 American University Law Review 1199–212.
Moore, Kimberly A. 2001. ‘Forum Shopping in Patent Cases: Does Geographic Choice Affect Innovation?’, 79
North Carolina Law Review 889–938.

Moore, Kimberly A. 2005. ‘Markman Eight Years Later: Is Claim Construction More Predictable?’, 9 Lewis &
Clark Law Review 231–47.
Mullally, Kelly Casey. 2010. ‘Legal (Un)Certainty, Legal Process, and Patent Law’, 43 Loyola Los Angeles Law
Review 1109–60.
Nard, Craig Allen. 1995. ‘Deference, Defiance, and the Useful Arts’, 56 Ohio State Law Journal 1415–509.
Nard, Craig Allen. 2002. ‘Toward a Cautious Approach to Obeisance: The Role of Scholarship in Federal
Circuit Patent Law Jurisprudence’, 39 Houston Law Review 101–26.
Nard, Craig Allen and John F. Duffy. 2007. ‘Rethinking Patent Law’s Uniformity Principle’, 101 Northwestern
Newman, Pauline. 1993. ‘The Federal Circuit: Judicial Stability or Judicial Activism?’, 42 American University
Newman, Pauline. 2002. ‘Origins of the Federal Circuit: The Role of Industry’, 11 Federal Circuit Bar Journal
541–3.
O’Malley, Kathleen M. 2012. ‘An Expanded “Slim Volume” on the Limited Role of Courts in Shaping Patent
Policy’, 22 Federal Circuit Bar Journal 91–100.
O’Malley, Kathleen M., Patti Saris, and Ronald H. Whyte. 2004. ‘A Panel Discussion: Claim Construction from
the Perspective of the District Judge’, 54 Case Western Reserve Law Review 671–90.
Osborn, Lucas S. 2012. ‘Instrumentalism at the Federal Circuit’, 56 St. Louis University Law Journal 419–63.
Pedraza-Fariña, Laura G. 2015. ‘Understanding the Federal Circuit: An Expert Community Approach’, 30
Petherbridge, Lee. 2009. ‘Patent Law Uniformity?’, 22 Harvard Journal of Law & Technology 421–73.
Petherbridge, Lee. 2010. ‘On the Development of Patent Law’, 43 Loyola Los Angeles Law Review 893–944.
Plager, S. Jay. 2009. ‘Frontiers in Empirical Patent Law Scholarship—Keynote Address’, 87 North Carolina Law
Review 1323–39.
Plager, S. Jay and Lynne E. Pettigrew. 2007. ‘Rethinking Patent Law’s Uniformity Principle: A Response to Nard
and Duffy’, 101 Northwestern University Law Review 1618–76.
Rai, Arti K. 2002. ‘Specialized Trial Courts: Concentrating on Expertise on Fact’, 17 Berkeley Technology Law
Journal 877–97.
Rai, Arti K. 2003. ‘Engaging Facts and Policy: A Multi-Institutional Approach to Patent System Reform’, 103
Rantanen, Jason and Lee Petherbridge. 2015. ‘Disuniformity’, 66 Florida Law Review 2007–43.
Schwartz, David L. 2008. ‘Practice Makes Perfect?: An Empirical Study of Claim Construction Reversal Rates
in Patent Cases’, 107 Michigan Law Review 223–84.
Schwartz, David L. 2010. ‘Pre-Markman Reversal Rates’, 43 Loyola Los Angeles Law Review 1073–108.
Schwartz, David L. and Lee Petherbridge. 2012. ‘Legal Scholarship and the United States Court of Appeals for
the Federal Circuit: An Empirical Study of a National Circuit’, 26 Berkeley Technology Law Journal 1561–602.
Sichelman, Ted. 2010. ‘Myths of (Un)Certainty at the Federal Circuit’, 43 Loyola Los Angeles Law Review
1161–94.
Stiernberg, Charlie. 2013. ‘Note, Science, Patent Law, and Epistemic Legitimacy: An Empirical Study of
Technically Trained Federal Circuit Judges’, 27 Harvard Journal of Law & Technology 279–300.
Taylor, David O. 2013. ‘Formalism and Antiformalism in Patent Law Adjudication: Precedent and Policy’, 66
Southern Methodist University Law Review 633–83.
Thomas, John R. 2003. ‘Formalism at the Federal Circuit’, 52 American University Law Review 771–810.
Vacca, Ryan. 2011. ‘Acting Like an Administrative Agency: The Federal Circuit En Banc’, 76 Missouri Law
Review 733–61.
Wagner, R. Polk and Lee Petherbridge. 2004. ‘Is the Federal Circuit Succeeding? An Empirical Assessment of
Judicial Performance’, 152 University of Pennsylvania Law Review 1105–80.
Wood, Diane P. 2014. ‘Keynote Address: Is it Time to Abolish the Federal Circuit’s Exclusive Jurisdiction in
Patent Cases?’, 13 Chicago-Kent Journal of Intellectual Property 1–10.
Cases
Alice Corp. Ltd. v. CLS Bank Int’l (2014), 134 S. Ct. 2347.
AT&T Corp. v. Excel Communications, Inc. (1999), 172 F.3d 1352.
Bilski v. Kappos (2010), 561 U.S. 593.
Bd. of Trustees of Leland Stanford Jr. Univ. v. Roche Molecular Sys., Inc. (2009), 583 F.3d 832.
Bd. of Trustees of Leland Stanford Jr. Univ. v. Roche Molecular Sys., Inc. (2011), 563 U.S. 776.
Christianson v. Colt Industries Operating Corp. (1988), 486 U.S. 800.
Cybor Corp. v. FAS Tech., Inc. (1998), 138 F.3d 1448.
eBay, Inc. v. MercExchange, L.L.C. (2006), 547 U.S. 388.

Festo Corp. v. Shoketsu Kinzoku Kogyu Kabushiki Co., Ltd. (2000), 234 F.3d 558.
Festo Corp. v. Shoketsu Kinzoku Kogyu Kabushiki Co., Ltd. (2002), 535 U.S. 722.
Global-Tech Appliances, Inc. v. SEB S.A. (2011), 563 U.S. 754.
Gottschalk v. Benson (1972), 409 U.S. 63.
Graham v. John Deere Co. (1966), 383 U.S. 1.
Graver Tank & Mfg. Co., Inc. v. Linde Air Prod. Co. (1950), 339 U.S. 605.
Holmes Group, Inc. v. Vornado Air Circulation Systems, Inc. (2002), 535 U.S. 826.
In re Bilski (2008), 545 F.3d 943.
In re Indep. Serv. Org. Antitrust Litigation CSU, L.L.C. v. Xerox Corp. (2000), 203 F.3d 1322.
In re Winslow (1966), 365 F.2d 1017.
Integra Lifesciences I Ltd. v. Merck KGaA (2003), 331 F.3d 860.
J.E.M. Ag. Supply, Inc. v. Pioneer Hi-Bred Int’l, Inc. (2001), 534 U.S. 124.
Johnson & Johnson Associates, Inc. v. R.E. Service Co., Inc. (2002), 285 F.3d 1046.
Kappos v. Hyatt (2012), 566 U.S. 431.
KSR Int’l Co. v. Teleflex Inc. (2005), 119 Fed. Appx. 282.
KSR Int’l Co. v. Teleflex Inc. (2007), 550 U.S. 398.
LG Elec., Inc. v. Bizcom Elec., Inc. (2006), 453 F.3d 1364.
Madey v. Duke University (2002), 307 F.3d 1351.
Markman v. Westview Instruments, Inc. (1995), 52 F.3d 967.
Markman v. Westview Instruments, Inc. (1996), 517 U.S. 370.
Merck KGaA v. Integra Lifesciences I, Ltd. (2005), 545 U.S. 193.
Microsoft Corp. v. i4i Ltd. P’ship (2011), 564 U.S. 91.
Nautilus, Inc. v. Biosig Instruments, Inc. (2014), 134 S. Ct. 2120.
Pfaff v. Wells Elec., Inc. (1997), 124 F.3d 1429.
Pfaff v. Wells Elec., Inc. (1998), 525 U.S. 55.
Phillips v. AWH Corp. (2004), 363 F.3d 1207.
Phillips v. AWH Corp. (2005), 415 F.3d 1303.
Quanta Computer, Inc. v. LG Elec., Inc. (2008), 553 U.S. 617.
Schering Corp. v. Geneva Pharms., Inc. (2003), 339 F.3d 1373.
South Corp. v. United States (1982), 690 F.2d 1368.
State Street Bank & Trust Co. v. Signature Fin. Group, Inc. (1998), 149 F.3d 1368.
Teva Pharm. USA, Inc. v. Sandoz, Inc. (2015), 135 S. Ct. 831.
Commission on Revision of the Federal Court Appellate System. 1975. ‘Structure and Internal Procedures:
Recommendations for Change’, 67 Federal Rules Decisions 195–408.
Federal Court Improvements Act of 1982, H.R. 4482, 97th Cong., Pub. L. No. 97-164, 96 Stat. 25.
S.R. No. 97-275, 97th Cong., 1st Sess. 1981, reprinted in 1982 U.S.C.C.A.N. 11.
H.R. Rep. No. 97-312, 97th Cong., 1st Sess. 1981.

7. Empirical studies of claim construction
J. Jonas Anderson* and Peter S. Menell**
7
Contents
I. Introduction
II. Economic, Legal, and Institutional Background
III. Hypotheses and Research Design
A. Reversal Rate Studies
B. Claim Construction Methodology
C. Judicial Behavior: Learning and Collegial Effects
IV. Jurisprudential and Policy Ramifications
V. Future Directions
References
I. INTRODUCTION
Patent claims define the scope of the patent right and hence are central to the operation
of the patent system. Patent prosecutors devote substantial effort to crafting patent
claims so as to maximize the scope of their right without “reading on” prior art (and
thereby defeating novelty). Businesses seeking to enter a technology marketplace must be
careful to avoid encroaching patent claims. Thus, when patentees enforce their rights, the
interpretation of claim boundaries guides both validity and infringement analysis.
With the steep rise in the use of jury trials in patent cases in the 1970s, the construction
of patent claims receded from view as judges delegated claim construction to juries with
general instructions. Following the Supreme Court’s decision in Markman v. Westview
Instruments, 517 U.S. 370 (1996), which held that “the construction of a patent, including
terms of art within its claim, is exclusively within the province of the court” (p. 372),
district judges began the practice of construing patent claims in advance of trial following
so-called “Markman” hearings. These constructions became subject to appellate review
after the trial or summary judgment ruling.
The Markman decision thus opened a valuable window into an important facet of
patent law and the litigation process. Beginning with the work of then-Professor Kimberly
Moore (2001), scholars have collected data on claim construction as a means of evaluating
a variety of hypotheses about the functioning of the patent system and judicial behavior
more generally.
** Professor of Law, American University, Washington College of Law.

** Koret Professor of Law and Director, Berkeley Center for Law & Technology, Berkeley
School of Law, University of California.
159

This chapter surveys this evolving literature. Section II summarizes the economic,
legal, and institutional background of this research. Section III describes the hypotheses,
research design, data, empirical methods, principal results, and limitations of these
studies. Section IV explores the jurisprudential and policy ramifications of these studies.
Finally, Section V suggests areas for future research.
II. ECONOMIC, LEGAL, AND INSTITUTIONAL

BACKGROUND
The patent system relies upon the delineation of boundaries to establish rights to intangible
resources—processes, machines, articles of manufacture, and compositions of matter. The
Patent Act of 1870 formalized use of patent claims by requiring applicants to “particularly
point out and distinctly claim the part, improvement, or combination which he claims as his
invention or discovery” (ch. 230, § 26, 16 Stat. 198, 201 (emphasis added)). Such claims pro-
vide the basis to examine applications (eligible subject matter, novelty, non-obviousness, and
disclosure), inform the public of the scope of rights, and enable courts to adjudicate patent
validity and infringement. Clarity of patent claims and transparency of patent institutions
play a critical role in the efficacy of patent law and institutions (Menell and Meurer, 2013).
Under current law, patent applicants must conclude their specification (written descrip-
tion of the invention and the manner and process of making and using the invention)
“with one or more claims particularly pointing out and distinctly claiming the subject
matter which the inventor . . . regards as the invention” (35 U.S.C. § 112(b)). A typical
patent contains multiple claims relating to the invention.
U.S. patent claims typically follow a “peripheral” claim format whereby the drafter
delineates the outer boundaries of the claimed invention. As an example, U.S. Patent No.
5,205,473 claims an insulated recyclable beverage container sleeve as:
A recyclable, insulating beverage container holder, comprising a corrugated tubular member

comprising cellulosic material and at least a first opening therein for receiving and retaining
a beverage container, said corrugated tubular member comprising fluting means for contain-
ing insulating air; said fluting means comprising fluting adhesively attached to a liner with a
recyclable adhesive.
Like other peripheral claims, this claim begins with a preamble (“A recyclable, insulating
beverage container holder”) followed by a transitional phrase (“comprising”) and the
claim body setting forth the claim restrictions (or elements). The open transitional phrase
“comprising” signifies that the patentee claims all structures containing the claim restric-
tions and anything else. By contrast, the closed transitional phrase “consisting of ” limits
the claim to structures containing the claim restrictions and nothing else.
The Patent Act authorizes applicants to prepare claims in various formats, including
dependent form (which incorporates by reference limitations of a prior or multiple
prior claims) and means-plus-function form. The Patent Act provides a specific rule for
interpreting claims using the means-plus-function format:
An element in a claim for a combination may be expressed as a means or step for performing a
specified function without the recital of structure, material, or acts in support thereof, and such

Empirical studies of claim construction 161
claim shall be construed to cover the corresponding structure, material, or acts described in the
specification and equivalents thereof.
(35 U.S.C. § 112(f)). Thus, the claim restriction “said corrugated tubular member
comprising fluting means for containing insulating air” from the ‘473 patent would be
limited to the “corresponding structure, material, or acts described in the specification
and equivalents thereof.”
As a noted patent jurist remarked, “the name of the game is the claim” (Rich, 1990,
p. 499). The incentives of those claiming intellectual property can diverge from the social
interest (Menell, 2019; Menell and Meurer, 2013). Inventors can sometimes benefit from
obfuscating the scope of rights and keeping others in the dark about their intellectual
property. (See, e.g., Sheldon, 2005, pp. 6–114; Fish, 2007, pp. 7–35). The Supreme Court
has recently sought to tighten claim drafting standards to better promote the claim clarity
required by the Patent Act (Nautilus, Inc. v. Biosig Instruments, Inc., 134 572 U.S. 898 (2014)).
Courts have developed specialized doctrines governing claim construction (Menell,
Powers, and Carlson, 2010). As a starting point, courts interpret claim terms based on
their ordinary and accustomed meaning to those having ordinary skill in the art (scientific
or engineering field of the invention) as of the filing date of the invention. Since few
judges have scientific or technical training, this standard places district judges in a chal-
lenging situation.
Various other rules and doctrines can come into play depending upon the circumstances.
The court must place principal reliance on intrinsic evidence (the patent specification and
the prosecution history), but may consider extrinsic evidence (e.g., dictionaries, treatises,
and expert testimony) so long as it does not contradict the intrinsic evidence. (Menell et
al., 2016, ch. 5). Patent claims are generally not limited to the preferred embodiments
set forth in the specification except with respect to means-plus-function claim elements,
which are limited to the “corresponding structure, material, or acts described in the
specification and equivalents thereof ” (35 U.S.C. § 112(f)).
Moreover, patentees can serve as their own lexicographers. Thus, courts generally
follow the definitions of claim terms set forth in the specification. The doctrine of claim
differentiation provides that each claim in a patent is presumptively different in scope.
Therefore, independent claims are generally interpreted more broadly than dependent
claims that incorporate the independent claim and add a limitation. Courts will also
read disclaimers reflected in the prosecution history (the correspondence between the
applicant and the Patent Office) into the claim terms where the disclaimers clearly and
unmistakably disavow claim scope.
Following the rise in the use of jury trials for patent cases in the 1970s (Anderson and
Menell, 2014, p. 20) (over 70 percent of patent cases were tried to juries by the early 1990s),
the interpretation of patent claims became subsumed in the jury’s verdict. This left the
trial judge and appellate court with very limited basis for ascertaining the correctness of
any underlying claim constructions. The Court of Appeals for the Federal Circuit, which
has exclusive jurisdiction over patent appeals, came to see the lack of transparency as an
impediment to the predictability and reviewability of patent adjudication (Michel, 1994,
pp. 1238–39).
The Federal Circuit addressed this problem in Markman v. Westview Instruments, 52
F.3d 967 (Fed. Cir. 1995) (en banc) (“Markman I”). Some of the Federal Circuit judges

assumed that if claim construction involved factual issues, then the Seventh Amendment
right stood in the way of trial judges in cases in which a party requested a jury trial
from construing patent claim terms. The majority opinion worked around the Seventh
Amendment impediment by holding that the construction of a patent claim is a pure
question of law, and hence outside of the jury’s purview, based on the “fundamental prin-
ciple of American law that ‘the construction of a written evidence is exclusively with the
court’” (Markman I, p. 978) (quoting Levy v. Gadsby, 7 U.S. (3 Cranch) 180, 186 (1805)).
Under this principle, the responsibility to construe patent claims fell, in the first instance,
to the district judge. It also meant that district judges’ construction of patent claim terms
was subject to de novo review. The decision finessed the question of how district judges,
lacking scientific and technical training, could step into the shoes of skilled artisans. The
court masked the factual nature of disputes about how terms would be understood by
skilled artisans by reasoning that although the trial judge could use extrinsic evidence
in construing claims, “the court is not crediting certain evidence over other evidence or
making factual evidentiary findings. Rather, the court is looking to the extrinsic evidence
to assist in its construction of the written document, a task it is required to perform”
(Markman I, p. 981).
The Supreme Court affirmed the Federal Circuit’s conclusion that the Seventh
Amendment did not require claim construction to be allocated to juries, but through
different reasoning (Markman v. Westview Instruments, 517 U.S. 370 (1996) (“Markman
II”)). The Court held that the district judge should be responsible for claim construction
based on judges’ “training in exegesis [of written instruments],” notwithstanding what it
characterized as the “mongrel [or mixed fact/law] practice” of patent claim construction.
(pp. 388, 378 respectively). In contrast to the Federal Circuit’s analysis, the Supreme
Court did not deem patent claim construction purely a question of law. Rather, consistent
with its characterization of claim construction as a “mongrel practice,” the Court noted
merely that claim construction was a matter “exclusively within the province of the court”
(p. 372).
The distinction between characterizing claim construction as a mixed question of fact
and law as opposed to a pure question of law has important ramifications for appellate
review. If claim construction is a mixed question, then the Federal Circuit would have to
accord deference to trial judges’ factual determinations underlying their claim construc-
tion rulings.1 If claim construction is a pure question of law, then the Federal Circuit
would review claim construction rulings de novo—that is, without any deference to the
district court’s factual findings.
The Federal Circuit ultimately adhered to the view expressed in its original Markman I
opinion that claim construction is a pure question of law (Cybor Corp. v. FAS Technologies,
Inc., 138 F.3d 1448 (Fed. Cir. 1998) (en banc)), although with vigorous disagreement (pp.
1465–66, Mayer, C.J., joined by Newman, J., concurring in the judgment). The Federal
Circuit continued to tinker with presumptions and the hierarchy of sources in an effort
to make claim construction more predictable. In Texas Digital Systems, Inc. v. Telegenix,
1
According to Rule 52(a)(6) of the Federal Rules of Civil Procedure, appellate courts must
“give due regard to the trial court’s opportunity to judge the witnesses’ credibility” and defer to the
trial court’s factual determinations unless “clearly erroneous.”

Inc., 308 F.3d 1193 (Fed. Cir. 2002), a panel sought to further clarify the claim construc-
tion framework by recognizing dictionaries, encyclopedias, and treatises as “particularly
useful resources to assist the court in determining the ordinary and customary meanings
of claim terms” due to their public availability and objectivity (p. 1202). The court noted
that, unlike expert testimony, these reference sources are not “colored by the motives
of the parties” or “inspired by litigation” (p. 1203). “Indeed, these materials may be the
most meaningful sources of information to aid judges in better understanding both the
technology and the terminology used by those skilled in the art to describe the technol-
ogy” (p. 1203).
Elevating dictionaries may well have exacerbated the unpredictability of claim
construction, eventually leading the Federal Circuit to grant en banc review to another
case to reconsider many aspects of claim construction methodology, including the
standard of review (Phillips v. AWH Corp., 415 F.3d 1303 (Fed. Cir. 2005) (en banc)).
The Phillips majority emphasized the role of intrinsic evidence. While authorizing trial
courts to consider extrinsic evidence, the Federal Circuit deemed such evidence “less
significant” and “less reliable” in determining the scope of claim terms. While continuing
to recognize the objectivity of dictionaries, the majority nonetheless backed away from
Texas Digital’s presumption that a dictionary meaning would apply unless the term
in question was explicitly defined in the specification or where the intrinsic evidence
disavowed or disclaimed such meaning. The majority declined to revisit the scope of
appellate review of claim construction rulings, thereby leaving Cybor’s de novo standard
in place. Nonetheless, lingering division among the members of the court remained.
While adhering to the view that claim construction is a question of law subject to de novo
review, Judge Lourie, joined by Judge Newman, wrote separately to urge his colleagues
“to lean toward affirmance of a claim construction in the absence of a strong conviction
of error” (p. 1330). In dissent, Judge Mayer, joined by Judge Newman, reiterated his
continued frustration with “the futility, indeed the absurdity, of this court’s persistence
in adhering to the falsehood that claim construction is a matter of law devoid of any
factual component” (p. 1330).
Division over the standard of review continued to surface after the Phillips decision.2
Anecdotal evidence indicated that the Cybor rule steered district judges away from using
extrinsic evidence and fully explaining the basis for their constructions, produced a
2
Amgen Inc. v. Hoechst Marion Roussel, Inc., 469 F.3d 1039, 1040 (Fed. Cir. 2006) (Moore,
C.J., joined by Rader, J., dissenting from denial of rehearing en banc); Medegen MMS, Inc. v. ICU
Med., Inc., 317 F. App’x 982, 988–91 (Fed. Cir. 2008) (Walker, Chief District Judge (N.D. Cal.)
dissenting) (urging greater deference); Trading Techs. Int’l, Inc. v. eSpeed, Inc., 595 F.3d 1340,
1363–64 (Fed. Cir. 2010) (Clark, District Judge (E.D. Tex.), concurring and writing separately
to “suggest that the current de novo standard of review for claim construction may result in the
unintended consequences of discouraging settlement, encouraging appeals, and, in some cases,
multiplying the proceedings”); Retractable Techs., Inc. v. Becton, Dickinson & Co., 659 F.3d 1369,
1373 (Fed. Cir. 2011) (Moore, J., joined by Rader, C.J., dissenting from denial of rehearing en
banc); Lighting Ballast Control LLC v. Philips Elecs. N. Am. Corp., 744 F.3d 1272, 1297 (Fed.
Cir. 2014) (en banc) (O’Malley, J., joined by Rader, C.J, Reyna, J., and Wallach, J. dissenting)
(castigating the majority for “misapprehend[ing] the Supreme Court’s guidance, contraven[ing]
the Federal Rules of Civil Procedure, and add[ing] considerable uncertainty and expense to
patent litigation”).

notable claim construction reversal rate, and demoralized district judges (Anderson and
Menell, 2014, pp. 45, 68–69; Mazumdar, 2005, p. 537).3
III. HYPOTHESES AND RESEARCH DESIGN
The controversy over the Cybor rule and its effects on reversal rates motivated the first
empirical studies of claim construction. The second group of studies explores the sources
and methodologies that judges employ in construing patent claims. A third group of
studies looks at claim construction as a proxy for appellate behavior more generally. This
section examines the hypotheses underlying these studies, the data used, the empirical
methods deployed, and the principal results.
A. Reversal Rate Studies
Moore (2001) presents the first systematic evaluation of appeals of district court claim
constructions. Moore’s paper measures the reversal rate as a basis for assessing the capac-
ity of district judges to adjudicate complex technical patent issues. In that latter respect,
her project built upon her work assessing the role of juries in patent cases (Moore, 2000).
Moore (2001) collected a database of all claim construction appeals to the Federal
Circuit from 1996 to 2000. It includes every Federal Circuit case, whether published,
unpublished, or summarily affirmed (Fed. Cir. Rule 36), in which claim construction
issues were appealed. She identified potential cases using a Boolean search for “patent
& claim/s interp! or constru!” on Westlaw (Moore, 2001, p. 8, n. 36). Her initial queries
retrieved 515 cases. After eliminating false positives, she identified 323 cases reviewing
construction of 496 separate claim terms. Since the Rule 36 summary affirmances did
not result in opinions, she reviewed the briefing in each of these (161) cases to determine
whether claim constructions had been appealed.
As Moore (2001, pp. 9–10) notes, her database could not directly assess the appellate
“correctness” of all district court claim construction decisions because not all claim
constructions were appealed. Some cases settled prior to appeal. And even for those
cases that were appealed, not all claim construction rulings were appealed. The critical
question is whether her sample is random or reflects selection bias. She notes two possible
influences: (1) observed reversal rate could understate the appellate correctness of district
court claim constructions because parties are less likely to appeal correctly decided cases
and/or issues; and (2) the transaction costs of appellate review.4
Moore found that the Federal Circuit reversed district court judges on at least one
claim construction issue in 33 percent of all the appealed patent cases between 1996 and
3
O’Malley et al., 2004, p. 682; Merck & Co. v. Teva Pharm. USA, Inc., 395 F.3d 1364, 1381
(Fed. Cir. 2005) (Rader J., dissenting) (noting that the Federal Circuit “often hears criticism from
district court judges that its reversal rate on claim construction issues far exceeds that of other
circuit courts”).
4
See also Priest and Klein, 1984, predicting that litigated outcomes should tend toward even
division under various assumptions; Kobayashi, 1996; Kessler et al., 1996, noting that the cases
that are appealed are most likely close cases.

2000. The Federal Circuit reversed 33 percent of means-plus-function constructions. On

15 percent of these constructions, the district court and the Federal Circuit disagreed over
whether the claim term was properly categorized within the § 112(f) category. Of those
that the Federal Circuit held to fall properly within that category, the Federal Circuit
reversed the district courts on 30 percent of the constructions. Where the Federal Circuit
found an error in claim construction, it reversed or vacated the trial court decision in 81
percent of the cases, resulting in an overall case reversal/vacate rate of 27 percent attribut-
able to claim construction errors. The 33 percent claim construction error rate exceeded
the error rate for appeals of other patent issues by a significant level.
Moore surmised that the higher reversal rate for claim construction determinations was
most likely attributable to the lack of deference under the Cybor rule, but could also reflect
district judges’ limited experience with claim construction. Since the reversal rate trended
upward during the period of study, she surmised that the Federal Circuit’s muddled claim
construction jurisprudence contributed to the problem.
Moore also tested whether the reversal rate was affected by variation among the
Federal Circuit judges, as opposed to objective error, by examining dissent rates in claim
construction decisions among Federal Circuit judges. She found, however, that Federal
Circuit judges dissented in only 3 percent of claim construction appeals. Using linear
regression, she also concluded that the Federal Circuit claim construction rulings were
driven to a statistically significant degree by Federal Circuit judges’ scientific or technical
training, political affiliation (based on the political party of the appointing President), or
prior patent experience.
Several follow-on studies—Chu (2001); Bender (2001); Zidel (2003)—found higher
reversal rates (approaching 50 percent) than those reported by Moore (2001), but these
higher rates were attributable to the omission of summary affirmance cases. Moore (2005)
followed up her initial study and found that the claim construction reversal rate continued
to rise in the three years (through 2003) following her initial study.
Schwartz (2010) added additional years to Moore’s studies, looking back to 1991 and
forward to 2008. His study separated the reversal rate over distinct time periods: January
1991 to Markman I (April 1995); Markman I to Markman II (April 1996); Markman II
to Cybor (March 1998); Cybor to Phillips (July 2005); and Phillips to December 2008.
He found that the reversal rate rose from a low of 17.7 percent in the period between
Markman I and Markman II to 32.1 percent in the period between Cybor and Phillips.
Schwartz also found that the reversal rate of claim construction cases in the pre-Markman
era was much lower than that of the post-Markman era. Sichelman (2009) analyzed
reversal rates from 2000 through 2007 on a host of patent litigation issues. Compared to
the reversal rates of other issues, claim construction was reversed 33 percent of the time,
higher than all but two of the issues (novelty and indefiniteness).
The debate over the standard of appellate review continued to fester after Phillips,
motivating Anderson and Menell (2014) to undertake a comprehensive assessment of the
history of claim drafting, the historical jurisprudence of claim construction and appellate
review of claim construction, and the empirical effects of the Phillips decision on a broad
range of claim construction measures, including reversal rates and evidentiary sources, by
technology area and Federal Circuit judge.
Our data set covered the period January 1, 2000 until December 31, 2011, which placed
the Phillips 2005 decision squarely at the midpoint. We used Boolean queries to identify

all Federal Circuit claim construction decisions (including Rule 36 summary affirmances)
and worked with a team of research assistants to collect and code a database covering
1930 individual claim terms from 1067 cases (pp. 78–79). The database comprised the fol-
lowing variables by claim term reviewed by the Federal Circuit: lower court; decision date;
posture on appeal (jury trial, bench trial, summary judgment, mandamus, other); Federal
Circuit panel; dissenting judge (if any); precedential nature of disposition (precedential,
nonprecedential, summary affirmance); final disposition on appeal; technological field(s)
(up to four); final disposition of claim construction (affirmed, reversed, avoided); whether
the Federal Circuit discussed the characteristics of the person having ordinary skill in the
art; the type of evidence discussed (intrinsic evidence—specification, same claim, other
claims, prosecution history; extrinsic evidence—expert testimony/tutorial, dictionary/
treatises, other).
Our data confirmed the results of Moore (2001, 2005) and Schwartz (2010) for the years
in which our studies overlapped. Of most significance, the data revealed a significant drop
(from 37.2 percent to 24 percent on a per term basis) in the claim construction reversal
rate. We found that the rate fell across all Federal Circuit judges and across all technology,
except business methods. The drop occurred immediately following the Phillips oral en
banc argument, five months before the decision was rendered. Thus, we were able to rule
out learning effects by district courts, since the drop occurred prior to the decision in
Phillips being announced and approximately two years before any post-Phillips district
court decisions could have reached the Federal Circuit. A simple linear regression of
the variables that we collected confirmed that the oral argument date was statistically
significant in explaining the change in reversal rate. Anderson (2015) confirmed that the
reversal rate remained low through 2013 (p. 177).
Notwithstanding our empirical demonstration that the Federal Circuit had “infor-
mally” shifted its standard of review toward greater deference, we nonetheless advocated
for a formal change to a hybrid standard of review (Anderson and Menell, 2014,
pp. 74–76). Beyond the jurisprudential point that the Federal Circuit had misapprehended
the Supreme Court’s reference to the mongrel (mixed law/fact) nature of claim construc-
tion in its Markman decision, we contended that the de novo standard of review stood in
the way of district courts developing an adequate record for understanding specialized
claim terms from the perspective of skilled artisans and discouraged clearly articulated
decisions.
Miller (2015) and Cotropia (2014) have found that claim construction reversals are
more likely in high technology areas. Miller (2015) defined software patents as those in
which “at least one claim element in the patent consists of data processing—the actual
manipulation of data—regardless of whether the code carrying out that data processing
is on a magnetic storage medium or embedded in a chip.” Using a database of Federal
Circuit opinions issued between January 1, 2002 and December 31, 2012, Miller found a
reversal rate of 40 percent for software patents and 24 percent for non-software patents.
Overall, he found that software patents were responsible for 40 percent of the difference
between the Federal Circuit’s overall claim construction reversal rate (29 percent) and its
18 percent reversal rate on all other patent issues.
Similarly, Cotropia (2014) found that claim construction of electronics, information
technology, and business methods (as a group) were more likely to be reversed if the
patentee won below, but not if the accused infringer won below between 2010 and 2013.

These results, however, were not statistically significant. Cotropia grouped patents in one
of three groups (Biologic/Chemical, Electronics/IT/Business Methods, and Mechanical),
unlike previous studies that broke down patents into more individualized groupings.
Lefstin (2007) studied dissenting opinions at the Federal Circuit to gauge the legal
indeterminacy that surrounds claim construction. He analyzed all published opinions
from 1998–2005 (Markman to Phillips) to determine how often they provoked dissenting
opinions. Although there are numerous ways to analyze the meaning of dissents, compar-
ing dissents within the same court reveals whether they occur because of uncertainty in
the law or other, court-specific reasons. Using dissent as a sign of indeterminacy, Leftsin
found the dissent frequency for claim construction to be similar to that of a number of
issues in patent law (infringement, invalidity, inequitable conduct, other), at 8.3 percent.
Thus, he concluded that there is no more legal indeterminacy in claim construction than
other areas of patent law.
Lefstin also compared claim construction dissent rates at the Federal Circuit to dissent
rates in contract interpretation cases at the Circuit Courts of Appeal. He concluded
that while claim construction is more likely to attract a dissent, this result can largely be
explained by the Federal Circuit’s propensity to dissent. Ultimately, he concluded that
claim construction was not substantially different from other doctrines, yet “[t]he relative
indeterminacy of claim construction appeals fluctuates with little apparent pattern over
the period 1998 to 2005 although the marked increase of indeterminacy in 2004 remains
prominent” (p. 1088).
Krause and Auyang (2014) focused their analysis on the directionality of reversal.
Using a dataset of post-Phillips claim construction reversals, they found that a higher
percentage of reversals (70.2 percent) result in broadening of patent claims, from which
they concluded that district courts systematically err in the narrowing direction. This
inference, however, does not account for selection effects that could explain the observed
mix of broadening/narrowing errors, such as the fact that the number of appealed pat-
entee wins is much lower than the number of appealed patentee losses (Cotropia, 2014).
Furthermore, the mix of appeals based on alleged infringement versus validity errors
could also skew the mix of broadening and narrowing reversals.
B. Claim Construction Methodology
The second group of studies identifies the sources examined by appellate judges in
performing claim construction and how they weigh those sources. These studies look for
judge-level effects to explain the claim construction process.
Wagner and Petherbridge (2004) focused on the methodological split among Federal
Circuit judges. Their study grouped judges into three categories: (1) proceduralists, who
give primary weight to the claim language (and the ordinary meaning, often derived
from dictionaries); (2) holistics, who interpret patent claim terms using an open-ended
methodology drawing upon the full range of interpretive tools—claim language, speci-
fication, prosecution history, dictionaries, and expert testimony; or (3) “swing” judges, a
middle group that does not subscribe to either the procedural or holistic methodology.
They looked at the Federal Circuit opinions from April 26, 1996 (Markman II) through
November 1, 2002 which “offer[ed] an observable patent claim construction analysis”
(p. 1145). Thus, their database included 393 cases, but did not include Rule 36 cases

or a number of opinions which did not lend themselves to claim construction analyses
(p. 1146, n. 143).
Wagner and Petherbridge found that the proceduralist methodology was employed
almost twice as often as the holistic approach, a trend that seemed to be increasing with
time. But, they found that claim construction appeal outcomes were highly dependent
on the composition of the appellate panel, and thus Federal Circuit claim construction
appeals were largely determined by one’s draw of judges. They concluded that the court
was divided into “factions” of proceduralists and holistics. In a follow-up study, Wagner
and Petherbridge could not confirm that Phillips had a “significant impact on the stability
and predictability” of claim construction: rather, they concluded that Phillips had further
divided the Federal Circuit on claim construction (Wagner and Petherbridge, 2011, p. 142).
With the Federal Circuit’s 2002 Texas Digital decision, attention shifted to the use
of dictionaries as a tool for making claim construction more predictable. Miller and
Hilsenteger (2005) examined the use of dictionaries in claim construction decision. They
used Boolean searches on Westlaw to flag potential claim construction decisions employ-
ing dictionaries between 1995 and 2004. (p. 845). They found significant increases in the
citation of dictionaries in both district court and Federal Circuit decisions using both
narrow and broad search criteria as a percentage of cases (pp. 845–50).
For those terms for which the Federal Circuit consulted a dictionary during the study
period, Miller and Hilsenteger also measured the number of dictionaries referenced per
claim term. They found that in approximately 20 percent of instances, the Federal Circuit
referenced more than one dictionary (pp. 860–61). They also analyzed the type of diction-
ary consulted, finding that general purpose dictionaries (70.5 percent), as opposed to
specialized technical dictionaries (29.5 percent), dominated the data set (p. 861). Focusing
more specifically on how dictionaries were used, Miller and Hilsenteger found examples
of serial dictionary use—that is, where dictionaries were used to interpret dictionaries
(pp. 863–64). They also found that Federal Circuit judges with patent law or technical
backgrounds were more likely to cite specialized dictionaries than their generalist col-
leagues (pp. 864–65).
Based on these findings, Miller and Hilsenteger came to see that Texas Digital’s
elevation of dictionaries to a higher status in the claim construction framework may be
illusory due to the wide range of definitions within and among dictionaries available (pp.
866–83). They note, for example, that the outcome of particular appeals—affirmance or
reversal—can depend on which general purpose dictionary the Federal Circuit chooses
to reference (pp. 876–79). They conclude their analysis by recommending that the Patent
Office solve the dictionary choice problem by requiring applicants to designate their
default dictionary preferences (general purpose and specialized) in the patent application
(pp. 883–904).
Just as Miller and Hilsenteger (2005) came to press, the Federal Circuit’s en banc
Phillips decision down-played the role of dictionaries in claim construction. The decision
noted that “[t]he main problem with elevating the dictionary to such prominence is that
it focuses the inquiry on the abstract meaning of words rather than on the meaning of
claim terms within the context of the patent” (Phillips v. AWH Corp., 415 F.3d 1303, 1321
(Fed. Cir. 2005) (en banc)). As a result, “too often [the Texas Digital] line of cases has
been improperly relied upon to condone the adoption of a dictionary definition entirely
divorced from the context of the [patent’s] written description” (p. 1321).

Anderson and Menell (2014) looked at the evidence examined in claim construction
appeals. The study found that while reliance on claim language and the specification has
remained largely unchanged, reliance on prosecution history has declined since 2003,
in agreement with Wagner and Petherbridge’s findings. Similarly, reliance on extrinsic
evidentiary sources, dictionaries in particular, has diminished since Phillips. This is
consistent with Phillips’s emphasis on intrinsic evidence. Furthermore, Anderson and
Menell (2014) found a large increase in cases decided by summary affirmance (Rule 36).
C. Judicial Behavior: Learning and Collegial Effects
The final group of studies looks at claim construction as a proxy for appellate behavior
more generally.
Schwartz (2008) examined whether district court judges improve their performance of
claim construction over time. He constructed two databases, one of claim construction
appeals and the second of biographical information on the individual judge. The appellate
database is similar to the databases created by Moore (2001, 2005) and Anderson and
Menell (2014). The biographical database contains the size of each district judge’s patent
docket, from 1995 to 2005. Using the databases, Schwartz assessed whether the judges’
performance (as measured by reversal rate) improved after each subsequent Federal
Circuit appeal.
Schwartz found no evidence that district judge reversal rates dropped as judges gained
experience. His study indicated that neither the reversal rate nor judges’ ability to “learn”
how to construe claims had improved over time. That is, judges were no more likely to
be reversed on their first appealed claim construction case than they were on their fifth.
Furthermore, Schwartz found that a district court judge’s first reversal does not have
a substantive or statistically significant effect on the future performance of the judge.
Namely, it does not appear that district court judges had improved their claim construc-
tion accuracy after their first reversal from the Federal Circuit. Using the biographical
database, Schwartz also analyzed a number of judge-specific factors (age, legal experience,
patent-specific experience) to present summary statistics of district court judge reversal
rates. He concluded that experience did not affect claim construction reversals. Age, on
the other hand, may play a role as the reversal rate drops until age 70, at which time it
spikes to nearly 40 percent.
Schwartz suggested three possible explanations for his findings: (1) inherent indetermi-
nacy of claim construction; (2) lack of guidance from the Federal Circuit; or (3) failure of
district court judges to properly apprehend claim construction analysis.
By contrast, Kesan and Ball (2011) and Stiernberg (2013) suggest that experience may
in fact benefit district court judges tasked with performing claim construction. Kesan and
Ball studied all cases filed from 1995 to 2003. Similar to Schwartz, they found that neither
general judicial experience nor cumulative patent-specific judicial experience increased the
likelihood of affirmance on appeal. However, they found that “recent” experience with
patent cases (defined as cases within the previous three years) did increase affirmance
odds in claim construction cases. Stiernberg (2013) analyzed claim construction appeals
from 2007–12 and found that district court judges with patent-specific experience were less
likely to have their claim construction decisions reversed. Furthermore, in agreement with
all previous studies, he found that time on the bench did not affect a judge’s reversal rate.

In a follow-up study, Schwartz (2009) examined the performance of International

Trade Commission administrative law judges (ALJs). He found that ALJs are reversed
as frequently as district court judges from patent-heavy jurisdictions—approximately 31
percent. Furthermore, he concluded that ALJs do not get better at claim construction
with more experience. Schwartz highlights the potential downsides to conducting such a
study, including selection effects, low numbers of cases appealed, and differing standards
for claim construction practice.
Lemley and Miller (2016) used claim construction data to evaluate hypotheses about
learning and familiarity among jurists. They utilized the Federal Circuit’s program of
inviting district court judges (as well as a few circuit court judges) to sit for a week on the
court. Their study employed Miller’s claim construction database and compared whether
the judge’s claim construction opinions were more likely to be reversed after sitting with
the Federal Circuit than before. They concluded that district judges who sit with the
Federal Circuit by invitation are thereafter significantly less likely to be reversed in their
claim construction decisions. They attributed the improvement in district judges’ reversal
rates to the personal relationships district judges develop with appellate judges while
sitting at the Federal Circuit as opposed to a learning effect.
IV. JURISPRUDENTIAL AND POLICY RAMIFICATIONS
Claim construction studies have played a substantial role in the development of claim
construction jurisprudence and patent policy relating to claim clarity. Shortly after she
published her 2005 study on claim construction reversal rates, Professor Kimberly Moore
was nominated to the Federal Circuit. She was confirmed by the Senate in September
2006. As noted above, her research criticized the Cybor rule and she appeared to favor
overturning it (Retractable Techs., Inc. v. Becton, Dickinson & Co., 659 F.3d 1369, 1373
(Fed. Cir. 2011) (Moore, J., joined by Rader, C.J., dissenting from denial of rehearing en
banc)). She ultimately declined the opportunity in Lighting Ballast Control LLC v. Philips
Elecs. N. Am. Corp., 744 F.3d 1272 (Fed. Cir. 2014) (en banc), joining the majority opinion
upholding Cybor on stare decisis grounds. The majority, concurring, and dissenting opin-
ions in the Lighting Ballast decision extensively discussed the jurisprudential, empirical,
and normative analysis in Anderson and Menell (2014).
On March 31, 2014, the Supreme Court granted certiorari in Teva Pharmaceuticals
USA, Inc. v. Sandoz, Inc., 134 S. Ct. 1761 (2014), a case addressing the standard of appel-
late review of patent claim construction determinations. Drawing on the jurisprudential
and empirical analysis in Anderson and Menell (2014), Menell, Anderson, and Rai (2014)
called for overturning the Cybor rule and adopting a hybrid standard of review based on
the mixed fact/law character of claim construction.
The Supreme Court’s decision in Teva Pharmaceuticals USA, Inc. v. Sandoz, Inc.,
572 U.S. 1033 (2015) overturned the Cybor rule, holding that the Federal Rule of Civil
Procedure 52(a)(6) requires a court of appeals to uphold a district court’s findings of fact
unless they are clearly erroneous (p. 835; Anderson and Menell, 2015). The Court noted
that its Markman decision “neither created, nor argued for, an exception to Rule 52(a),”
and recognized that “subsidiary factfinding is sometimes necessary” in patent claim
construction, directly contradicting nearly two decades of Federal Circuit jurisprudence.

The majority recognized that while “[c]onstruction of written instruments often presents
a ‘question solely of law,’ at least when the words in those instruments are ‘used in
their ordinary meaning,’” extrinsic evidence may help when “a written instrument uses
‘technical words or phrases not commonly understood.’” “And in that circumstance, the
‘determination of the matter of fact’ will ‘preced[e]’ the ‘function of construction.’ . . .
This factual determination, like all other factual determinations, must be reviewed for
clear error” (pp. 837–38). The Supreme Court’s decision noted the relatively high rate of
claim construction reversals in its analysis (p. 839, citing Menell et al., 2014)).
Empirical studies of claim construction may also have contributed to the broader policy
questions surrounding claim clarity. The Supreme Court’s decision in Nautilus, Inc. v.
Biosig Instruments, Inc., 572 U.S. 898 (2014), tightening the standards for invalidating
claims for indefiniteness (35 U.S.C. § 112(b)), partially addresses the problem of claim
construction by subjecting vague or ambiguous claims to greater risk of invalidation
(Menell, 2014).
Along similar lines, in June 2013, the Executive Office of the President called upon
the Patent Office to promote “clearer patents with a high standard of novelty and
non-obviousness” and address problems of functional claiming (Executive Office of
the President, 2013; Lemley, 2013; Menell, 2012; Menell, 2013). The Patent Office has
pursued several initiatives along these lines, including promoting the use of glossaries
in patent applications (U.S. Patent and Trademark Office, 2014; Miller and Hilsenteger,
2005), and “making claim construction explicit in the record, including the scope of claim
terms, and functionally defined clauses” (U.S. Patent and Trademark Office, 2015).
In its recent decision Williamson v. Citrix Online, LLC, 792 F.3d. 1339 (Fed. Cir. 2015)
(en banc), the Federal Circuit has also tightened the standard for subjecting claim terms
using generic language to the more limited scope applicable to means-plus-function
claim formats. Prior to this decision, the Federal Circuit applied a “strong” presumption
that a term lacking the word “means” does not invoke § 112(f). In re-characterizing the
presumption, Williamson explained that a term lacking “means” will nonetheless be con-
strued under § 112(f) if the “challenger demonstrates that the claim term fails to ‘recite[]
sufficiently definite structure’ or else recites ‘function without reciting sufficient structure
for performing that function’” (p. 1349) (quoting Watts v. XL Sys., Inc., 232 F.3d 877,
880 (Fed. Cir. 2000). The focus is on the claim language as a whole, not just the isolated
term that is akin to “means.” Generic terms such as “mechanism,” “element,” “device”
and other such terms that do not connote sufficiently definite structure in the context
of the overall claim are tantamount to stating “means,” and therefore may be construed
pursuant to § 112(f) if nothing else in the claim provides sufficient structure.
V. FUTURE DIRECTIONS
These jurisprudential developments and administrative reforms open up new avenues for
empirical research on claim construction and claim drafting.
The Supreme Court’s decision in Teva Pharmaceuticals USA, Inc. v. Sandoz, Inc., 135
S. Ct. 831 (2015) opens a new chapter in claim construction case management and juris-
prudence (Anderson and Menell, 2015). It remains to be seen the extent to which district
judges will make greater use of extrinsic evidence in claim construction proceedings and

prepare more careful orders explaining the basis for their claim construction. Holding
evidentiary hearings and preparing detailed factual findings adds to the already substan-
tial workloads of federal district judges. These workloads are especially burdensome in
patent-heavy districts. It is also unclear the extent to which the Federal Circuit will view
extrinsic evidence relied upon by district courts as worthy of deference. The Teva decision
affords the Federal Circuit the ability to find that intrinsic evidence trumps extrinsic
evidence.
Other hypotheses that could be tested following the Teva decision are whether the more
deferential regime results in greater cost and delay as parties engage in escalating battles
of expert witnesses and whether greater deference increases the finality of district court
decisions—that is, whether more cases settle prior to appeal. Measuring these effects
could prove difficult as a result of roughly contemporaneous changes in patent litigation
driven by changes in patent eligibility doctrines (Bilski v. Kappos, 561 U.S. 593 (2010);
Mayo Collaborative Services v. Prometheus Laboratories, Inc., 566 U.S. 66 (2012); Ass’n
for Molecular Pathology v. Myriad Genetics, Inc., 569 U.S. 576 (2013); Alice Corp. v. CLS
Bank International, 573 U.S. 208 (2014)), and the rise in the use of post-grant proceed-
ings at the Patent Trial and Appeal Board (PTAB) to invalidate patents. Nonetheless, the
greater availability of trial court records provides a window into studying how the Teva
decision reverberates throughout the patent litigation ecosystem.
The Patent Office initiatives—in conjunction with greater judicial scrutiny of patent
claiming following the Teva, Nautilus v. Biosig Instruments, Inc., 572 U.S. 898 (2014), and
Williamson v. Citrix Online, LLC, 792 F.3d 1339 (2015) (en banc) decisions—provide an
opportunity to investigate claim drafting, reissuance, and prosecution practices. These
decisions and initiatives have the potential to discourage the use of ambiguous claim
language and promote the use of glossaries. In addition, in conjunction with the patent
eligibility cases, they are likely to alter the types of patents sought.
The PTAB provides a further opportunity for studying claim construction. The PTAB
uses the “broadest reasonable interpretation” standard for construing patent claims in
evaluating patent validity (In re Cuozzo Speed Techs., LLC, 778 F.3d 1271 (Fed. Cir.
2015)). It will be useful to evaluate the extent to which the broadest reasonable interpreta-
tion standard compares to the actual meaning applied in district proceedings.
Similarly, the Williamson decision provides another natural experiment for evaluating
patent litigation behavior. It appears likely that a significant number of software patents
(and perhaps others) will be more carefully scrutinized in the coming years. This might
also lead to an increase in reissuance practice as patentees seek to clarify the scope of their
claims in advance of enforcement.
REFERENCES
Anderson, J. Jonas. 2015. “Specialized Standards of Review”, 18 Stanford Technology Law Review 151–88.
Anderson, J. Jonas and Menell, Peter S. 2014. “Informal Deference: A Historical, Empirical, and Normative
Analysis of Patent Claim Construction”, 108 Northwestern University Law Review 1–83.
Bender, Gretchen A. 2001. “Uncertainty and Unpredictability in Patent Litigation: The Time is Ripe for a
Consistent Claim Construction Methodology”, 8 Journal of Intellectual Property Law 175–222.
Chu, Christian A. 2001. “Empirical Analysis of the Federal Circuit’s Claim Construction Trends”, 16 Berkeley

Cotropia, Christopher A. 2014. “Is Patent Claim Interpretation Review Deference or Correction Driven?”, 2014
Brigham Young University Law Review 1095–120.
Kesan, Jay P. and Ball, Gwendolyn G. 2011. “Judicial Experience and the Efficiency and Accuracy of Patent
Krause, Thomas W. and Auyang, Heather F. 2014. “What Reversals and Close Cases Reveal about Claim
Construction: The Sequel”, 13 John Marshall Review of Intellectual Property Law 525–54.
Lefstin, Jeffrey A. 2007. “The Measure of the Doubt: Dissent, Indeterminacy, and Interpretation at the Federal
Circuit”, 58 Hastings Law Journal 1025–94.
Lemley, Mark A. and Miller, Shawn P. 2016. “If You Can’t Beat ’Em, Join ’Em? How Sitting by Designation
Affects Judicial Behavior”, 94 Texas Law Review 451–84.
Miller, Scott Joseph and Hilsenteger, James A. 2005. “The Proven Key: Roles and Rules for Dictionaries and
the Patents Office and the Courts”, 54 American University Law Review 829–939.
Miller, Shawn P. 2015. “‘Fuzzy’ Software Patent Boundaries and High Claim Construction Reversal Rates”, 17
Moore, Kimberley A. 2000. “Judges, Juries, and Patent Cases—An Empirical Peek inside the Black Box”, 99
Moore, Kimberly A. 2001. “Are District Court Judges Equipped to Resolve Patent Cases?”, 15 Harvard Journal
Moore, Kimberley A. 2005. “Markman Eight Years Later: Is Claim Construction More Predictable?”, 9 Lewis
& Clark Law Review 231–47.
Schwartz, David L. 2008. “Practice Makes Perfect?: An Empirical Study of Claim Construction Reversal Rates
in Patent Cases”, 107 Michigan Law Review 223–84.
Schwartz, David L. 2009. “Courting Specialization: An Empirical Study of Claim Construction Comparing
Patent Litigation Before Federal District Courts and the International Trade Commission”, 50 William and
Schwartz, David L. 2010. “Pre-Markman Reversal Rates”, 43 Loyola of Los Angeles Law Review 1073–108.
Sichelman, Ted. 2009. “Myths of (Un)certainty at the Federal Circuit”, 43 Loyola of Los Angeles Law Review
1161–94.
Stiernberg, Charlie. 2013. “Science, Patent Law, and Epistemic Legitimacy: An Empirical Study of Technically
Trained Federal Circuit Judges”, 27 Harvard Journal of Law & Technology 279–300.
Wagner, R. Polk and Petherbridge, Lee. 2004. “Is the Federal Circuit Succeeding? An Empirical Assessment of
Judicial Performance”, 152 University of Pennsylvania Law Review 1105–80.
Wagner, R. Polk and Petherbridge, Lee. 2011. “Did Phillips Change Anything? Analysis of the Federal Circuit’s
Claim Construction Jurisprudence”, in Shyamkrishna Balganesh, ed., Intellectual Property and the Common
Law. New York, NY: Cambridge University Press.
Zidel, Andrew T. 2003. “Patent Claim Construction in the Trial Courts: A Study Showing the Need for Clear
Guidance from the Federal Circuit”, 33 Seton Hall Law Review 711–71.
General References
Anderson, J. Jonas and Menell, Peter S. 2015. “Restoring the Fact/Law Distinction in Patent Claim
Construction”, 109 Northwestern University Law Review Online 187–200.
Executive Office of the President, White House. June 2013. “Patent Assertion and U.S. Innovation”, www.
whitehouse.gov/sites/default/files/docs/patent_report.pdf.
Fish, Robert D. 2007. Strategic Patenting. IN, Bloomington: Trafford Publishing.
Kessler, Daniel, Meites, Thomas and Miller, Geoffrey. 1996. “Explaining Deviations from the Fifty-Percent Rule:
A Multimodal Approach to the Selection of Cases for Litigation”, 25 The Journal of Legal Studies 233–59.
Kobayashi, Bruce H. 1996. “Case Selection, External Effects, and the Trial Settlement Decision”, in David A.
Anderson, ed., Dispute Resolution: Bridging the Settlement Gap. London: JAI Press.
Lemley, Mark A. 2013. “Software Patents and the Return of Function Claiming”, 2013 Wisconsin Law Review
905–64.
Mazumdar, Anandashankar. 2005. “Federal District Courts Need Experts that Are Good ‘Teachers,’ Judges Tell
Bar”, 70 (1736) BNA’s Patent, Trademark & Copyright Journal 536–37.
Menell, Peter S. 2012. Promoting Patent Claim Clarity, Solutions to the Software Patent Problem, UC Berkeley
Public Law Research Paper, No. 2171287, http://papers.ssrn.com/sol3/papers.cfm?abstract_id=2171287.
Menell, Peter S. 2013. “It’s Time to Make Vague Software Patents More Clear”, Wired, Feb. 7, www.wired.
com/2013/02/its-time-to-make-vague-software-patents-more-clear/.
Menell, Peter S. 2014. Brief Amicus Curiae of Professor Peter S. Menell In Support Of Neither Party, Nautilus,
Inc., v. Biosig Instruments, Inc., Supreme Court of the United States, No. 13-369.

Menell, Peter S. 2019. “Economic Analysis of Intellectual Property Notice and Disclosure”, in Ben Depoorter
and Peter S. Menell, eds., Research Handbook on the Economics of Intellectual Property Law, Vol. I.
Menell, Peter S. and Meurer, Michael J. 2013. “Notice Failure and Notice Externalities”, 5 Journal of Legal
Analysis 1–59.
Menell, Peter S., Anderson, Jonas and Rai, Arti K. 2014. Brief of Professors Peter S. Menell, J. Jonas Anderson,
and Arti K. Rai As Amici Curiae In Support Of Neither Party, Teva Pharmaceuticals USA, Inc., et al., v.
Sandoz, Inc., et al., Supreme Court of the United States, No. 13-854.
Menell, Peter S., Powers, Matthew D. and Carlson, Steven C. 2010. “Patent Claim Construction: A Modern
Synthesis and Structured Framework”, 25 Berkeley Technology Law Journal 711–829.
Menell, Peter S., Powers, Matthew, Pasahow, Lynn, Pooley, James, Carlson, Steven, Homrig, Jeffrey, Pappas,
George, Chang, Carolyn, Reiner Mayer, Colette and David Peters, Marc. 2016. Patent Case Management
Judicial Guide. Washington, D.C.: Federal Judicial Center. 3rd ed.
Michel, Paul R. 1994. “The Challenge Ahead: Increasing Predictability in Federal Circuit Jurisprudence for the
New Century”, 43 The American University Law Review 1231–57.
O’Malley, Kathleen M., Saris, Patti and Whyte, Ronald H. 2004. “A Panel Discussion: Claim Construction from
the Perspective of the District Judge”, 54 Case Western Reserve Law Review 671–90.
Priest, George L. and Klein, Benjamin. 1984. “The Selection of Disputes for Litigation”, 13 The Journal of
Rich, Giles S. 1990. “The Extent of the Protection and Interpretation of Claims-American Perspectives”, 21
International Review of Industrial Property and Copyright Law 497–510.
Sheldon, Jeffrey G. 2005. How to Write a Patent Application. NY: New York City, Practicing Law Institute.
The Patent Act of 1870, ch. 230 § 26, 16 Stat. 198.
U.S. Patent and Trademark Office (USPTO), Glossary Pilot Program, 59 Fed. Reg. 17137 (Mar. 27, 2014), www.
gpo.gov/fdsys/pkg/FR-2014-03-27/pdf/2014-06792.pdf.
U.S. Patent and Trademark Office, Request for Comments on Enhancing Patent Quality, 89 Fed. Reg. 6475 (Feb.
5, 2015), www.gpo.gov/fdsys/pkg/FR-2015-02-05/pdf/2015-02398.pdf.
U.S. Patent No. 5,205,473 (filed Mar. 19, 1992).
Case References
Alice Corp. Pty. Ltd. v. CLS Bank Int’l, 573 U.S. 208 (2014).
Amgen Inc. v. Hoechst Marion Roussel, Inc., 469 F.3d 1039 (Fed. Cir. 2006).
Ass’n for Molecular Pathology v. Myriad Genetics, Inc., 569 U.S. 576 (2013).
Cybor Corp. v. FAS Techs., Inc., 138 F.3d 1448 (Fed. Cir. 1998) (en banc).
In re Cuozzo Speed Techs., LLC, 778 F.3d 1271 (Fed. Cir. 2015).
Levy v. Gadsby, 7 U.S. 180 (1805).
Lighting Ballast Control LLC v. Philips Elecs. N. Am. Corp., 744 F.3d 1272 (Fed. Cir. 2014) (en banc).
Markman v. Westview Instruments, 52 F.3d 967 (Fed. Cir. 1995) (en banc).
Markman v. Westview Instruments, 517 U.S. 370 (1996).
Mayo Collaborative Servs. v. Prometheus Labs., Inc., 566 U.S. 66 (2012).
Medegen MMS, Inc. v. ICU Med., Inc., 317 F. App’x 982 (Fed. Cir. 2008).
Merck & Co. v. Teva Pharms. USA, Inc., 395 F.3d 1364 (Fed. Cir. 2005).
Nautilus, Inc. v. Biosig Instruments, Inc., 572 U.S. 898 (2014).
Phillips v. AWH Corp., 415 F.3d 1303 (Fed. Cir. 2005) (en banc).
Retractable Techs., Inc. v. Becton, Dickinson & Co., 659 F.3d 1369 (Fed. Cir. 2011).
Teva Pharms. USA, Inc. v. Sandoz, Inc., 572 U.S. 1033 (2014).
Teva Pharms. USA, Inc. v. Sandoz, Inc., 135 S. Ct. 831 (2015).
Tex. Digital Sys. v. Telegenix, Inc., 308 F.3d 1193 (Fed. Cir. 2002).
Trading Techs. Int’l, Inc. v. eSpeed, Inc., 595 F.3d 1340 (Fed. Cir. 2010).
Watts v. XL Sys., Inc., 232 F.3d 877 (Fed. Cir. 2000).
Williamson v. Citrix Online, LLC, 792 F.3d 1339 (Fed. Cir. 2015) (en banc).
35 U.S.C. § 112(b).
35 U.S.C. § 112(f).
Fed. R. Civ. P. 52(a)(6).

8. Empirical studies of the International Trade
Commission
Colleen V. Chien* and David L. Schwartz**
Contents
I. Introduction
II. How the ITC’s Section 337 Authority Works
III. How Section 337 Investigations Proceed
IV. Existing Empirical Research on Section 337
A. Filing and Adjudication Trends
B. The ITC Versus District Court
C. Legal Doctrines and Outcomes
V. Directions for Future Research
References
I. INTRODUCTION
The International Trade Commission (ITC) is an independent, quasi-judicial agency of

the Federal government that hears patent cases1 through the exercise of its authority
under Section 337 of the Tariff Act of 1930 to regulate unfair trade practices. Section 337
provides relief from unfair methods of competition and unfair acts in the importation or
sale of articles into the United States if it harms a U.S. industry—for example, through
patent infringement. While the ITC was designed to provide a special venue to deal with
the special problem of infringing imports, as more and more goods are manufactured
overseas, its jurisdiction has come to overlap more and more with that of U.S. district
courts. Indeed, over the past 20 years, around 70 percent of cases that are filed in Section
337 have a district court counterpart (Chien and Cotter, forthcoming). The opposite is not
true: most district court patent cases have no Section 337 counterpart.
Several features of the Section 337 docket have made it well suited for empirical patent
scholarship and the subject of policy inquiry. First, the Section 337 adjudication at the
ITC has a number of unique features that are designed to enable prosecution of alleged
foreign infringers that would otherwise evade U.S. district courts. Patentees enjoy more
permissive jurisdiction, faster adjudication, and broader (and narrower) remedies in the
** Professor of Law, Santa Clara University Law School.

** Stanford Clinton Sr. and Zylpha Kilbride Clinton Research Professor of Law, Northwestern
University School of Law.
*1 We are aware of only three public interest-based denials of exclusion orders, and six
Presidential vetoes.
175

ITC as compared to the district court. Unlike federal district courts, the ITC, pursuant
to its Section 337 authority, does not hear counterclaims, almost always grants injunc-
tive relief in the form of exclusion or cease and desist orders to prevailing patentees,
and does not recognize certain defenses to infringement available to accused infringers
in district courts. These characteristics, and the relatively high rate of parallel filings
and of adjudication (44 percent of cases proceed to trial, as compared to 11 percent in
district court) (Chien, 2008), provide a useful opportunity to probe the impact of various
institutional design features on party behaviors and outcomes. Second, the long-standing
overlap between the jurisdiction of the district courts and ITC provide a case study in
the coordination between parallel venues litigating the same dispute. Patent disputes
are increasingly adjudicated in multiple venues—including the district court and Patent
Trial and Appeal Board at the United States Patent and Trademark Office (USPTO),
where the majority of recently introduced inter partes review proceedings have a parallel
district court filing, and the courts of first instance within the U.S. and abroad. Given
the issue of multiple venues, there is increased policy relevance to questions of how to
coordinate the activities of different venues in the same dispute. Policy issues include, for
instance, how to approach tribunals with different remedies, defenses, and requirements
for pursuing claims on the same patent, such as in the ITC and district court. Finally,
because the ITC’s rules require a domestic injury and the frequent grant of cease and
desist orders or exclusion orders (which operate similar to district court injunctions), the
ITC has come under policy scrutiny. Some argue that the ITC is unfairly protectionist,
while others accuse NPEs of moving to the ITC seeking the injunctions that the Supreme
Court’s eBay decision has made harder to obtain from district court.
For all of these reasons, while the ITC is only one of 96 venues nationwide in which
patent infringement cases can initially be brought,2 more ITC patent investigations reach
trial than district court patent cases. The settlement rate of ITC investigations was just
under 50 percent for the period Q1 2014 to May 2016 (ITC, 2014b). This chapter describes
the features of the ITC and the main policy and theoretical questions that have animated
study of the ITC’s 337 authority, what we do and don’t know, and possible directions for
future research.
II. HOW THE ITC’S SECTION 337 AUTHORITY WORKS
The ITC is tasked with investigating the impacts of import practices on domestic
industries, a role it executes through five principal functions: import injury investiga-
tions, intellectual property-based import investigations pursuant to Section 337, industry
and economic analysis, tariff and trade information services, and trade policy support
(ITC website, 2015a and 2015b).3 It is led by six Commissioners who are appointed for
nine-year terms by the President based on alternating political party affiliations with the
2
There are 94 district court venues, infringement cases can also be brought in the ITC and the
Court of Federal Claims (Anderson, 2015).
3
www.usitc.gov/press_room/about_usitc.htm (last accessed Apr. 27, 2015); www.usitc.gov/
press_room/documents/usitc_organization_chart.pdf (last accessed Apr. 27, 2015).

Empirical studies of the International Trade Commission 177
advice and consent of the Senate (ITC website, 2015c), and includes five administrative
law judges (ALJs), serving lifetime appointments.
The ITC’s Section 337 authority was created by the Tariff Act of 1922 and then the
Smoot-Hawley Tariff Act of 1930, which also raised tariffs on a wide array of goods.
However, it was not until the 1974 Trade Act, which gave the ITC the power to order
injunctions, formalized Section 337 processes in accordance with the Administrative
Procedures Act, and codified a variety of other features tailored to the characteristics
of imports, that the ITC became practically useful to patent plaintiffs. To allow injured
domestic industries to bring cases against foreign defendants which might be able to evade
court service, as well as the enforcement of remedies, the ITC’s jurisdiction became in
rem, or against the “thing” or accused imported article, rather than in personam, against
a particular defendant which may be located overseas. In rem jurisdiction also provided
the basis for general exclusion orders—a special remedy available only at the ITC that
excludes infringing items regardless of source, avoiding the need to file a series of
separate complaints against multiple foreign manufacturers and thwarting an anecdotally
common infringer tactic of closing up shop under one name only to reopen under another
one. To provide efficient and effective relief to domestic industries, ITC investigations
were required by statute to be completed within 12 to 18 months.
Despite these improvements, obstacles remained to the use of Section 337 (Wilson,
1986). The Omnibus Trade and Competitiveness Act of 1988 addressed these by eliminat-
ing the prerequisite to suit of proof of injury, in effect enabling any patentee alleging an
intellectual property-infringing importation to initiate an investigation. The Act also
relaxed the “domestic industry” requirement, no longer requiring engineering, research,
69
56
41 42
40
35
33
31
29
26
24
17 17 18
13 13
11 11
9
6
94
95
96
97
98
99
00
01
02
03
04
05
06
07
08
09
10
11
12
13
19
19
19
19
19
19
20
20
20
20
20
20
20
20
20
20
20
20
20
20
Calender year
Source: (ITC, 2014a).
Figure 8.1 Number of ITC investigations (1994–2013)

or even production of a product—and in at least one case, mere licensing activities were
deemed to be sufficient to meet this prong.4 But while some of these features made the
ITC a favorable venue for domestic industries, the perception that they, in doing so,
discriminated against other countries led to charges of illegal protectionism. In 1988, a
General Agreement on Tariffs and Trade panel ruled that aspects of Section 337 violated
international law, and in 1995 the statute was changed again. Under the auspices of the
Uruguay Round Agreements, the ITC’s strict time limits were relaxed to resolution on
“the earliest practicable time.” ITC respondents were given the ability to request and
obtain a stay of a concurrent district court action, and it became harder to get a general
exclusion order.
III. HOW SECTION 337 INVESTIGATIONS PROCEED
A Section 337 investigation begins with the filing of a formal complaint by a company that
alleges the importation of products that infringe U.S. intellectual property rights. While
the ITC can also hear cases alleging trademark, trade dress, copyright, false advertising,
passing off, antitrust and trade secret misappropriation,5 historically, approximately 80
percent to 90 percent of its caseload has been allegations of patent infringement (Chien,
2009). For instance, in fiscal year 2009, 79 of the 85 active Section 337 investigations
included a patent infringement claim (ITC, 2009a).
Within 30 days of receiving a complaint, the ITC normally issues an institution deci-
sion, declining to proceed only in “rare circumstances” (ITC Trial Lawyers Association,
2017), and the case is assigned a presiding ALJ. In another unique feature of the ITC, an
investigative (or “staff ”) attorney is assigned to each Section 337 investigation and par-
ticipates in discovery, motions, and trial on behalf of the public interest. In addition, the
presiding ALJ sets a target date for completion of the investigation, typically 16 months or
less following initiation of the investigation (ITC, 2009b, p. 16). This aggressive timeframe
for completion of cases means that the ITC, in contrast to the district court, has generally
been unwilling to grant stays for inter partes review proceedings at the USPTO (Quinn
Emanuel webpage, 2015).
Section 337 proceedings are governed by “Ground Rules” from each ALJ that closely
resemble the Federal Rules of Evidence and the Federal Rules of Civil Procedure used
in district court adjudication. Complainants can also seek and receive temporary relief,
in the form of a “preliminary decision,” within 70 to 120 days after the investigation is
instituted, followed, in successful cases, by a temporary exclusion order 20 or 30 days
later (ITC, 2009b, Question 17). Unlike district court judges, ALJs at trial allow hearsay
evidence and most do not require live direct witness testimony. Discovery and investiga-
tion proceed on a compressed timeframe as the ALJ conducts a hearing typically within
4
More recently, the ITC has made it more difficult for complainants to meet the domestic
industry requirement, holding that it is only in “exceptional circumstances” that domestic industry
includes the manufacture of a “downstream article” that incorporates a patented article. See In
re Certain Wireless Standard Compliant Electronic Devices, Including Communication Devices and
Tablet Computers, Inv. No. 337-TA-953 (Dec. 2015), 2015 WL 9875534.
5
Tianrui Group Co. v. ITC, 661 F.3d 1322, 1337 (Fed. Cir. 2011).

eight to 10 months after the initiation of a Section 337 investigation. Not less than four
months before the target date, the ALJ issues an Initial Determination (ID) on all of the
issues presented. The ITC reviews the ID and issues a Final Determination about four
months later. Following the Final Determination, the ITC determines the remedy, if any.
Unlike district courts, the ITC cannot award monetary awards. Because there can be
no money damages, even for findings of past infringement by the ITC, the patent owner
frequently files a parallel lawsuit in district court to seek monetary remedies. However,
based on a finding of unfair competition, the ITC has the power to issue three types of
injunctions: (1) a limited exclusion order; (2) a general exclusion order (which applies to
imports regardless of source), to keep infringing goods outside the United States; and (3)
a cease and desist order, to prevent the dissemination of imports that are already within
U.S. borders. Though the Federal Circuit has ruled that the eBay factors do not apply to
remedy decisions by the ITC in Section 337 proceedings,6 the ITC is required by law to
take them into account to determine whether an exclusion order is inconsistent with the
public interest, with input from other regulatory agencies: 19 U.S.C. § 1337(b)(2). The ITC
has declined to issue an exclusion order to a prevailing patentee three times in its history
(Chien and Lemley, 2012).
Each ITC order is subject to a 60-day Presidential Review period during which
the United States Trade Representative, a component of the Executive Office of the
President, may disapprove of the order for policy reasons. The President’s veto power
has been exercised six times (Chien and Lemley, 2012; Froman, 2013), the latest time in
2013 when the President vetoed an exclusion order against Apple iPhones on the basis of
infringement of a standards-essential patents held by Samsung (Froman, 2013).
IV. EXISTING EMPIRICAL RESEARCH ON SECTION 337
Researchers have studied Section 337 cases based on several lines of inquiry. Almost all
the empirical research has focused on patent disputes, with a single article addressing
trade secret claims (Riley and Stroud, 2013). First, several studies have considered filing
and adjudication trends at the ITC, by plaintiff nationality and type, in part motivated
by charges of protectionism and the opportunistic use of the ITC. A second group of
studies compares processes and outcomes at the ITC and district court, with respect to,
for example, claim construction and remedies. A third group of studies has considered
specific legal doctrines and how the ITC has applied them.
A. Filing and Adjudication Trends
While Section 337 explicitly encodes a preference for domestic industries, several studies
have considered the question of whether the ITC is impermissibly biased in favor of U.S.
companies in its application of Section 337. Domestic plaintiffs dominate the filings,
but cases have been brought by as or more frequently against domestic defendants than
foreign ones (Chien, 2008, 2012). Plaintiffs do win a higher percentage of their ITC cases
6
Spansion, Inc. v. International Trade Comm’n, 629 F.3d. 1331 (Fed. Cir. 2010).

than they do their district court cases (Chien, 2008; Hahn and Singer, 2008), but this may
reflect selection effects—when cases filed in both venues are considered, most of this dif-
ference disappears (Chien, 2008). In addition, parties are much more likely to adjudicate
their cases to a final outcome in the ITC than they are in district court (Chien, 2008; Hahn
and Singer, 2008; ITC, 2014).
From 1995 to 2014, China was the country that was most frequently alleged to be
the original source or middleman of infringing imports, followed by Taiwan, and Japan
(Chien, 2008; Chien and Cotter, forthcoming). Almost three-quarters of allegedly
infringing goods were from Asia. Ran (2010) has performed an analysis of Section 337
investigations involving Chinese entities from 2007–09.
A number of studies have considered the strategic issues relating to the selection of
selecting whether to initiate an action at the ITC. One study has analyzed ITC investiga-
tions through the lens of non-market actions taking advantage of institutional targeting
(Somaya and McDaniel, 2012). Most of the attention, however, has been focused on
case filings in the ITC by non-practicing entities (NPEs) and other patent holders whose
odds of getting an injunction in the ITC (close to 100 percent) far exceed their rate of
0 percent in district court (7 percent, when requests are contested) (Chien and Lemley,
2012). As with district court filings, the numbers are contested. While Lemley and Chien
(2012) found that in 2011, 51 percent of respondents had been named by an NPE, the
ITC’s own report, which broke down NPEs into two types, reported the NPE total to be
closer to 41 percent in 2011 and 53 percent in 2012, although the vast majority of these
were brought by “Category 1 NPEs”—that is, inventors which conduct R&D, research
institutions, start-ups, and other manufacturers whose own products do not practice the
patents at issue (ITC, 2014b). Category 2 NPEs—that is, entities which do not manufac-
ture products and whose business model primarily focuses on purchasing and asserting
patents—appeared more frequently than Category 1 NPEs in 2013, 2014, and 2015,
with six, three, and two ITC investigations, respectively, brought by Category 1 NPEs in
those years (ITC, 2016). Concerns about the use of the ITC to seek injunctions based on
standards-essential patents subject to a commitment to license on reasonable and non-
discriminatory terms have led to various policy actions (Froman, 2013; DOJ-PTO, 2013).
These and other developments may be responsible for the fluctuation in ITC case
filings, which, after rising from 2009–11, declined through 2015 (Chien and Cotter,
forthcoming; Davis, 2014) (see also Cotropia (2011), for a discussion of how in particular
one case that narrowed the availability of general exclusion orders, the Kyocera case, did
not significantly impact filing trends).
While the studies described above focused on firm filings, one older study considered
the characteristics of those who bring cases, and the impact of adjudicated outcomes.
While complaining firms tend to be larger and more diverse, Mutti and Yeung’s (1996)
results suggest, their profits also are significantly depressed when there is a finding of no
violation.
B. The ITC Versus District Court
Eighty five percent of the ITC’s Section 337 docket is comprised of patent cases, as
compared to 15 percent of a typical district court (Chien, 2008). Does this specialization
result in different or better outcomes? Schwartz (2009) considered this issue in an empiri-

cal study of claim construction outcomes, finding no discernible difference in success

rates at the Federal Circuit between patent-experienced ALJs at the ITC and district court
judges. A parallel study also looked at appeals of claim construction and other ITC rulings
to the Federal Circuit, finding that most ITC errors were based on affirmations of ALJ
rulings (Diehl, 2011).
Empirical research has also quantified how different fora have different injunction
award rates. While the ITC’s injunction rate is estimated to be 95–100 percent, district
courts applying the discretionary standard articulated in the eBay case have awarded
injunctions at a lower rate, around 72.5 percent, with significant variability depending
on the plaintiff’s entity type (Seaman, 2016; Chien and Lemley, 2012; Hahn and Singer,
2008).
C. Legal Doctrines and Outcomes
A handful of empirical studies, mostly published in the trade publication ITC Reporter,
have considered the application of specific legal doctrines in Section 337 cases, usually
over a limited number of years. For example, Riley and Goulet (2010) have examined the
frequency and nature of public interest arguments in the evaluation of proposed exclusion
orders, advocating for greater representation of the public interest in ITC proceedings on
this basis. Namrow and Roeder (2010) studied inequitable conduct allegations at the ITC
and their outcomes, finding that, as in district court, “[i]nequitable conduct is frequently
asserted, yet rarely successful.”
Although it is seemingly easier to obtain an exclusion order at the ITC than to obtain
an injunction in district court, empirical research suggests that exclusion orders are not
always effective. Li finds, based on a review of exclusion orders issued from 2002–11,
that their scope is often imprecise and vague, and difficult to remedy in the case of non-
compliance. The ITC’s own surveys in 2005 and 2010 suggest that among patent owners
that received exclusion orders, the majority believed that infringing goods had been
imported despite the order (ITC, 2010).
V. DIRECTIONS FOR FUTURE RESEARCH
There are several directions for future research into the ITC that we think are ripe for
exploration. While several articles have tackled the differences between outcomes in the
ITC and federal district courts, there is no empirical research on how outcomes at the
ITC compare to those in international courts. More particularly, how do patent owners
decide whether to assert their patents in the U.S. district courts, the ITC, foreign courts,
or some combination thereof ? Even within the United States, the empirical research has
focused upon differences in outcomes at the various fora. A well-conducted survey of
litigants may yield fruitful information. While we recognize that selection concerns make
it difficult to untangle the effects of the fora from other effects, given the limited research
to date on the issue, further investigations are warranted.
Another direction for future research is the procedures used at the ITC. Currently,
most of the research consists of analyzing outcomes—who won, whether an exclusion
order was issued, etc. The full ITC dockets, along with public versions of the underlying

filings, are available for free on the ITC website. Using the dockets, more discrete aspects
of the ITC investigations can be considered—including, for instance, motion practice and
settlements.
Additional research can also focus on the substance of the infringement allegations.
One aspect worth studying is gray-market goods: goods which are authorized by the
intellectual property rights holders to be sold in a foreign country, but not authorized
to be imported into the United States. These disputes frequently involve allegations of
trademark and copyright infringement, though the availability of other, faster forms7
of relief depresses the use of the ITC. For these goods, the ITC investigations may
differ from a classic infringer-type scenario in which there is no agreement permitting
the alleged infringer to manufacture anywhere. Presently, we do not even have a count
on how many suits involve gray-market goods and how many are classic infringer-type
cases.
There is also scant research on the composition of the ITC. The ITC, by statute, has
an equal number of Republican and Democratic appointees. But a legal realist would
assume that the precise identification of Commissioners matters, with some being more
protectionists or pro-IP, and others being less so. Research into the effects, if any, of the
Commissioners on ITC investigations could yield interesting results. The same is true for
research into the ALJs who oversee the ITC investigations. Individual ALJs may have
preferences that differ from other ALJs.
Two aspects of ITC Section 337 cases that differ from U.S. district courts open doors
to potentially interesting empirical research. These two aspects can be studied internally
(within the ITC) and externally (compared to district courts), although external study is
more difficult.
First, the ITC staff attorney is one of the most distinctive features of Section 337
investigations, at least as compared to district courts. The staff attorney takes substantive
positions on issues in the investigation. There is little persuasive empirical evidence on
positions taken by the staff attorney during investigations, and even less on the effects
of the staff attorney on the investigation itself. Since the staff attorney is so different by
other courts in the United States, we may be able to learn lessons about the adversarial
system itself from studying how the staff attorney relates to accuracy and efficiency of
the adjudication process.
Second, Section 337 cases are decided by ALJs. There are no jury trials, and reduced
opportunities to seek summary judgment. This results in a more streamlined litigation
process, unlike the piecemeal litigation in the district courts. The more methodical and
piecemeal litigation in the district courts has advantages and disadvantages. A careful
study of trials at the ITC may be useful to understanding unitary trials, and perhaps even
shed light on whether unitary trials are more accurate or efficient than the mechanisms
typically employed by district courts. It can help us to know how effectively and how often
certain legal doctrines are being adjudicated at the ITC.
7
Pursuant, for example, to the Lanham Act of 1946, the Copyright Act of 1976, and the
Digital Millennium Copyright Act of 1998.

REFERENCES
Anderson, J. Jonas. 2015. “Court Competition for Patent Cases,” 163:3 University of Pennsylvania Law Review
631–35.
Chien, Colleen V. 2008. “Patently Protectionist? An Empirical Analysis of Patent Cases at the International
Trade Commission,” 50 William and Mary Law Review 63–114.
Chien, Colleen V. 2012. “The International Trade Commission and Patent Disputes,” Testimony before
Congress.
Chien, Colleen V. 2012. “Calling a Truce Over ITC Patent Data,” IP Watch Dog, Jul. 20, available at www.
ipwatchdog.com/2012/07/20/calling-a-truce-over-itc-patent-data/id=26640/.
Chien, Colleen V. and Mark A. Lemley. 2012. “Patent Holdup, the ITC, and the Public Interest,” 98 Cornell
Law Review 1–45.
Chien, Colleen V. and Tom Cotter. Redesigning Patent Law (forthcoming Oxford University Press).
Cotropia, Christopher A. 2011. “Strength of The International Trade Commission as a Patent Venue,” 20 Texas
Davis, Ryan. 2014. “ITC Becomes Ghost Town for Essential Patent Filings,” Law 360, Aug. 8, available at www.
law360.com/articles/565263/itc-becomes-ghost-town-for-essential-patent-filings.
Diehl, Michael. 2011. “Does ITC Review of Administrative Law Judge Determinations Add Value in Section
337 Investigations?” 21 Federal Circuit Bar Journal 119–63.
Froman, Michael. 2013. “Decision on the USITC’s Investigation of Certain Electronic Digital Media Devices,
Office of the United States Trade Representative,” Oct. 18, available at www.ustr.gov/about-us/press-office/
press-releases/2013/October/Froman-decision-USITC-investigation.
Hahn, Robert W. and Hal J. Singer. 2008. “Assessing Bias in Patent Infringement Cases: A Review of
International Trade Commission Decisions,” 21 Harvard Journal of Law and Technology 457–508.
ITC. 2009a. “Year in Review for Fiscal Year 2009.” Available at www.usitc.gov/publications/year_in_review/
pub4167.pdf.
ITC. 2009b. “Section 337 Investigations: Frequently Asked Questions.” Available at www.usitc.gov/intellectu
al_property/documents/337_faqs.pdf.
ITC. 2010. “Summary of Results of FY 2010 Survey Regarding Section 337 Exclusion Orders.” Available at
www.usitc.gov/intellectual_property/documents/summary_of_results_of_2010_survey.pdf.
ITC. 2014a. “Number of Section 337 Investigations Instituted by Calendar Year.” Available at www.usitc.gov/
intellectual_property/documents/cy_337_institutions.pdf.
ITC. 2014b. “USITC Section 337 Investigations—Facts and Trends regarding Caseload and Parties.” Available
at www.usitc.gov/press_room/documents/featured_news/337facts2014.pdf.
ITC. 2016. “Section 337 Statistics: Number of Section 337 Investigations Brought by NPEs.” Available at www.
usitc.gov/intellectual_property/337_statistics_number_section_337_investigations.htm.
ITC Trial Lawyers Association. 2017. FAQs. Available at www.itctla.org/resources/faqs.
ITC Website. 2015a. “United States International Trade Commission Office-Level Organizational Chart.”
Available at www.usitc.gov/press_room/about_usitc.htm; www.usitc.gov/press_room/documents/usitc_organi
zation_chart.pdf.
ITC Website. 2015b. “About the USITC.” Available at www.usitc.gov/press_room/about_usitc.htm.
ITC Website. 2015c. “Commissioner Bios.” Available at www.usitc.gov/press_room/bios.htm.
Mutti, John and Bernard Yeung. 1996. “Section 337 and the Protection of Intellectual Property in the United
States: The Complainants and the Impact,” 78:3 The Review of Economics and Statistics 510–20.
Namrow, Eric S. and Susan Roeder. 2010. “The Current State of the Inequitable Conduct Defense in ITC
Proceedings,” 337 Reporter XXXII 2–10.
Quinn Emanuel. 2015, “Inter Partes Review and the ITC: The Benefits and Risks of Filing IPR on Patents Asserted
in an ITC Investigation.” Available at www.quinnemanuel.com/the-firm/news-events/article-march-2015inter-
partes-review-and-the-itc-the-benefits-and-risks-of-filing-ipr-on-patents-asserted-in-an-itc-investigation/.
Ran, Ruixue. 2010. “Section 337 Investigations and the Chinese Market,” 337 Reporter Volume XXIX 2–6.
Riley, P. Andrew and Paul C. Goulet. 2010. “An Option for Public Interest Factfinding at the Commission,”
337 Reporter Volume XXIX 7–12.
Riley, P. Andrew and Jonathan R.K. Stroud. 2013. “A Survey of Trade Secret Investigations at the International
Trade Commission: A Model for Future Litigants,” 15 Columbia Science and Technology Law Review 41–89.
RPX Corp. 2013. “Data on the ITC.” Available at www.rpxcorp.com/wp-content/uploads/2014/01/RPX-2013-
NPE-Litigation-Report.pdf.
Schwartz, David L. 2009. “Courting Specialization: An Empirical Study of Claim Construction Comparing
Patent Litigation Before Federal District Courts and the International Trade Commission,” 50 William and
Seaman, Christopher B. 2016. “Permanent Injunctions in Patent Cases,” 101 Iowa Law Review 1949–83.

Somaya, Deepak and Christine A. McDaniel. 2012. “Tribunal Specialization and Institutional Targeting in
Patent Enforcement,” 23:2 Organization Science 869–87.
United States Department of Justice and United States Patent and Trademark Office. Jan. 8, 2014. “Policy
Statement on Remedies for Standards-Essential Patents Subject to Voluntary F/RAND Commitments.”
Available at www.justice.gov/sites/default/files/atr/legacy/2014/09/18/290994.pdf.
Wilson, Pete. 1986. “International Piracy Involving Intellectual Property: Hearing Before the Subcomm. on
Trade, Productivity, and Economic Growth of the Joint Economic Committee,” 99th Cong. 1028 (opening
statement).

9. Technical standards, standards-setting
organizations, and intellectual property: a survey
of the literature (with an emphasis on empirical
approaches)
Jorge L. Contreras* 8
Contents
I. Introduction
II. Standards and the Standard-Setting Landscape
A. Standards-Setting Organizations
1. The voluntary standard-setting ecosystem
2. SSO membership
3. SSO processes
B. Standards in Technology Markets
C. Open Standards
III. Patents and Standards
A. Patenting Standards
B. Quantifying Standards-Essential Patents
1. Number of disclosures—aggregate
2. SEP holders
3. Geographic distribution
4. Case studies
C. The Value of Standards-Essential Patents
1. Patent status
2. Patent transfers
3. Market share
4. Citations
5. Litigation
6. Firm performance
D. Acquisition and Timing of Declaration of Standards-Essential Patents
E. Effect of Patents on Standardization Activity
F. Standards as Prior Art
G. Potential Market Effects: Hold-up, Hold-out, and Stacking
1. Hold-up
2. Royalty stacking
H. Patent Pools and Standards
* Professor, University of Utah S.J. Quinney College of Law – jorge.contreras@law.utah.edu.
185

I. Patent Assertion Entities and Standards-Essential Patents

J. Patents and Non-ICT Standards
IV. Private Ordering and SSO Patent Policies
A. Cataloging SSO Patent Policies
B. SSO Policy Features: FRAND Licensing Terms
C. SSO Policy Evolution and Amendment
D. SSO Patent Policy Influences on Firm Membership Decisions
V. SSO Patent Policy Disputes
A. Disclosure Disputes
1. Intentional non-disclosure—patent ambush
2. Essentiality and disclosure
B. FRAND Licensing Disputes
1. Royalty rates
2. Royalty base and apportionment
3. Non-discrimination
4. Injunctions
VI. Other SSO Policy Features and Proposals
A. Ex Ante Licensing Disclosure
B. Aggregate Royalty Caps
C. Royalty-free Licensing
D. Alternative Dispute Resolution
E. Transfer of Commitments
VII. Non-Patent Intellectual Property and Standards
A. Copyright
1. Incorporation by reference
2. Software in standards
B. Trademarks and Certification Marks
VIII. Conclusions
References
I. INTRODUCTION
Technical interoperability or compatibility standards specify design features that enable

products manufactured by different vendors to work together in a manner that is largely
invisible to the consumer. Physical product configurations, from railroad gauges to
drill bits to electrical plugs, have been standardized for more than a century (Shapiro
and Varian, 1999; Ernst, 2012; Russell, 2014). More complex, but equally important,
are the many networking (USB, Wi-Fi, Bluetooth), Internet (TCP, IPv6, HTML), and
telecommunications (CDMA, GSM, UMTS, LTE) standards that are essential to the
global technology infrastructure. Broadly adopted standards such as these can produce
efficiency-enhancing network effects and other benefits (Katz and Shapiro, 1985; Shapiro
and Varian, 1999; Shapiro, 2001; Shapiro and Lemley, 2007).
Standards may be developed in a variety of settings. Many health, safety and envi-
ronmental standards are developed by governmental agencies. The large majority of
interoperability standards, however, are developed in the private sector. Individual firms

Technical standards, standards-setting organizations, and IP 187
may develop proprietary technologies that, through broad market adoption, become
de facto standards. In several well-known cases (e.g., Betamax vs. VHS, HD-DVD vs.
Blu-ray, Internet Explorer vs. Netscape Navigator), competing firms have engaged in so-
called “standards wars” to determine which of their proprietary formats would eventually
prevail and become essential features of products in the marketplace (Shapiro and Varian,
1999; Breshnahan and Yin, 2007). Over the past two decades, however, most interoper-
ability standards have been developed by groups of market participants that collaborate
within voluntary associations known as standards-development organizations (SDOs) or
standards-setting organizations (SSOs). The resulting standards are often referred to as
“voluntary consensus standards”, which will be the principal focus of this chapter.
Despite their potential benefits, voluntary consensus standards have over the past
decade become the subject of significant private litigation, regulatory enforcement and
policy debate. Much of the current controversy centers on the perceived proliferation
of patents covering standardized technologies, potentially abusive enforcement of such
patents against manufacturers and users of standardized products, and the terms on
which patent holders may be required to license the use of those patents to others. This
chapter offers an overview of the empirical, legal and economic literature concerning
the interaction of interoperability standards and SSOs with intellectual property rights
(primarily patents, with attention to copyrights and trademarks as well).1
II. STANDARDS AND THE STANDARD-SETTING

LANDSCAPE
A. Standards-Setting Organizations
1. The voluntary standard-setting ecosystem

SSOs vary greatly in size and composition. Some, which are sometimes referred to as
consortia or special interest groups, consist of just a few firms that collaborate on a
narrow set of technical specifications, sometimes for a single product. Standards for
consumer electronics devices and media such as the DVD were developed in this manner.
1
This chapter focuses principally on law, policy and organizations that are active in North
America and Europe, as well as case law developments in the United States and Europe. There is
a growing body of literature discussing standardization bodies and practices in Asia, particularly
China, Japan, Korea and India, that is beyond the scope of this chapter. Recent literature on
standardization in different Asian jurisdictions is summarized, among others, in assorted chapters
of NRC (2013) and Contreras, ed. (2017). Even with this degree of selectivity, the literature of
standards, standardization and IP is too large and rapidly developing to cover comprehensively in
this chapter. By way of illustration, between March 2013 and September 2016, nearly 600 papers
were distributed via the SSRN’s Law, Policy & Economics of Technical Standardization ejournal
(www.ssrn.com/link/Law-Policy-Econ-Tech-Standards.html), with more than 100 in 2016 alone. In
addition, in 2016 no fewer than four major reports commissioned by the European Commission in
the area of standards and IP were released (Pentheroudakis and Baron (2017), Pohlmann and Blind
(2016), JRC (2016) and Régibeau et al. (2016)). And this is just a subset of the rapidly growing
worldwide literature in this field.

Other SSOs have thousands of members and oversee multiple standardization activities
at any given time.
The European Commission (2014) places SSOs into three broad categories: (1) those
that are formally recognized by governments (e.g., the European Telecommunications
Standards Institute (ETSI) and the International Organization for Standardization (ISO));
(2) “quasi-formal” groups that are typically large and well organized and share many of the
characteristics of formally recognized groups, but lack official governmental recognition
(e.g., the Institute of Electrical and Electronics Engineers (IEEE) Standards Association
and the Internet Engineering Task Force (IETF)); and (3) smaller, privately organized con-
sortia (also known as special interest groups or fora). In addition, the work of individual
SSOs is sometimes coordinated at national and international levels. For example, in the
U.S., the American National Standards Institute (ANSI) oversees, accredits and establishes
policy for national SSOs that wish to develop American National Standards.
This international standard-setting “ecosystem” and its principal components are
described in greater detail by David and Shurmer (1996), Mattli and Büthe (2003),
Nickerson and zur Muehlen (2006), Cargill and Bolin (2007), Biddle et al. (2012), Ernst
(2012), EC (2014), Lundqvist (2014), Gandal and Régibeau (2015), and Biddle (2018).
Gandal and Régibeau (2015) identify 548 active SSOs worldwide across a range of indus-
tries. When less formal consortia are counted, Updegrove (2015) catalogs nearly 1000
standards-development groups operating in various fields. The increasing prominence of
consortia in the standards-development world, and the impact of consortia on innovation
in standardized technologies, are analyzed along various axes of accountability, transpar-
ency, efficiency, consensus, and flexibility by Updegrove (1995), Egyedi (2001a), Baron
and Pohlmann (2013), Bar and Leiponen (2014), Baron, Meniere and Pohlmann (2014),
and Delcamp and Leiponen (2014).
2. SSO membership

Standards development is conducted primarily by personnel employed by firms active in
relevant product markets, with the occasional involvement of academic and governmental
participants. Baron and Spulber (2018), using membership data from 195 different SSOs,
find that the median SSO had 114 members, and only five SSOs had membership levels
greater than 1000. They observe that several large firms in the computing, semiconductor
and electronics industries were, as of 2013, actively engaged in 50 or more different SSOs.
In 2003, Updegrove (2003) finds that two major computing firms were each involved in
more than 150 SSOs. Contreras (2014), observing rates of Asian participation in IETF,
finds that participation in Internet standardization by Japanese and Korean firms has
remained meaningful but steady over the years, while participation by Chinese firms has
increased from virtually nil in 2003 to a position in 2013 second only to U.S. firms.
3. SSO processes

The consensus standardization process was first modeled by Farrell and Saloner (1988)
as a war of attrition. Though this process was shown to result in greater coordination
than decentralized activity, it is cumbersome. Simcoe (2007b, 2012) explores the length
of the standardization process at IETF and the impact of Internet commercialization on
standards development. Farrell and Simcoe (2012) further explore and expand the war of
attrition model for standard-setting.

2
Though the processes for voluntary consensus standard-setting may vary among SSOs,
some common features are prescribed by law. Following a series of cases in the 1980s
involving abuses of the standardization process (Hydrolevel, 1982; Allied Tube, 1988),
the U.S. Supreme Court established that in order to avoid antitrust liability, SSOs should
observe a certain level of transparency, openness and due process. These requirements
have since been embodied in Office of Management and Budget Circular A-119 (1998),
which governs U.S. federal agency use of private sector standards, as well as guidance
from U.S. antitrust enforcement agencies (DOJ-FTC, 2000; DOJ-FTC, 2007). Similar
requirements have also been adopted in the ANSI Essential Requirements (2016), which
establish minimum policy and due process requirements for ANSI-accredited SSOs.
Baron et al. (2019, 4.1) systematically analyze these and other exogenous legal and regula-
tory constraints on SSO behavior.
Case studies of individual SSOs and their processes for standards and policy
development include the following:
Table 9.1 Case studies of SSOs
SSO Case studies

DVB Eltzroth (2008)
ECMA/JCT1 Egyedi (2001a, 2001b)
ETSI/3GPP Besen (1990), Shurmer and Lea (1995), Bekkers and Smits (1997), Bekkers
and Liotard (1999), Bekkers, Duysters, and Verspagen (2002), Iversen
(2002), Cowhey, Aronson, and Richards (2006), Leiponen (2008), Bekkers
and West (2009), Bar and Leiponen (2014), Baron and Gupta (2018),
Caviggioli et al. (2015)
HDTV2 Farrell and Shapiro (1992), Neil et al. (1995)
IEEE DeLacey et al. (2006), Wright (2008), Contreras (2013a)
IETF Lehr (1995), Froomkin (2003), Nickerson and zur Muehlen (2006), Simcoe
(2007b), Contreras and Housley (2008), Ernst (2012), Simcoe (2012),
Contreras (2013a, 2014, 2016a), Russell (2014), Wen et al. (2015)
INCITS Ernst (2012)
ISO/IEC Murphy and Yates (2009), Blind (2011), Choi and Jang (2014), Lundell et al.
(2015)
ITU (56k modems, Gandal, Gantman, and Genesove (2007), Greenstein and Rysman (2007),
H.265) Egyedi (2016)
MISMO Steinfield, et al. (2007)
OASIS Blind (2011), Ernst (2012)
SEMI Langlois (2007)
VITA Contreras (2013a)
W3C Egyedi (2001a), Russell (2011), Contreras (2016a), Gamalielsson and
Lundell (2016)
2
HDTV, the U.S. high-definition television standard adopted by the Federal Communications
Commission (FCC), was developed through a series of interactions among the FCC and different
U.S., European and Japanese SSOs.

B. Standards in Technology Markets
As noted above, standards are pervasive in technology markets today, particularly com-
puting, networking, semiconductors, and telecommunications. Baron and Spulber (2018)
collect data on more than 750 000 standards documents from 90 SSOs. They find that large
formal SSOs produced the majority of standards documents, with two groups (Comité
European de Normalization (CEN) and ISO) producing more than 100 000 documents
each. The largest single-sector SSO (ETSI) produced more than 85 000 documents in
the area of telecommunications standardization (id.). The authors recognize, however,
that the sheer quantity of standards documents produced may not be indicative of an
SSO’s commercial and technological significance, as major SSOs such as IETF (Internet
standards) and IEEE (Wi-Fi) have produced fewer than 5000 standards documents each
(id.), yet have shaped large areas of technical development.
In addition to aggregate studies of standards documents, studies have been conducted
regarding the prevalence of standards in particular product categories. Biddle, White, and
Woods (2010) identify more than 250 standards embodied in a single laptop computer;
and Armstrong, Mueller, and Syrett (2014) provide a catalog of the standards embodied
in the different functional subsystems of a smart phone.
C. Open Standards
The term “open standards” originally arose in the context of governmental procurement
regulations to describe policies whereby governmental agencies may procure software
or other technology systems only if they utilize open (as opposed to “proprietary” or
“closed”) standards. Various studies cited by Shah and Kesan (2007) predict significant
cost savings for agencies utilizing open standards, which can theoretically reduce costs of
document format incompatibility and conversion.3 Recently, however, use of the term
“open standards” has expanded beyond government procurement into general discus-
sions of standards and standardization (Russell, 2014).
Despite its widespread use, there is no generally accepted definition of “open stand-
ards” and numerous definitions exist (Ernst, 2012; Updegrove, 2012; Baron and Spulber,
2018; Lundell et al., 2015; Kesan, 2019). Shah and Kesan (2007) define open standards
by reference to three criteria: public availability, licensing on fair, reasonable, and non-
discriminatory (FRAND) terms, and development in a process open to public participa-
tion. West (2007) offers an economic analysis of openness along several dimensions
including access, competition, and cost. Krechmer (2011) identifies 17 related attributes
of open standards, classified according to the requirements of different interest groups
(SSOs, commercial implementers, end users, economists, and attorneys). And Baron and
Spulber (2018) consider any commonly available standard developed by an SSO, rather
than owned by a single firm, to be “open”.
In the procurement arena, a number of governmental bodies around the world have
mandated that open standards must be available on a royalty-free basis (Updegrove,
3
Nevertheless, Shah and Kesan (2012) find significant interoperability problems even with
document format standards touted as “open”.

2012). Particularly notable in this regard are recent policy initiatives by India and South
Africa (DeNardis, 2009; Rens, 2011). Ghosh (2011) presents evidence from a survey of
955 public officials in 13 European states regarding practices and attitudes in software
procurement. Notwithstanding EU policies favoring openness, he finds that a sub-
stantial percentage of respondents continue to support proprietary software solutions.
Governmental open standards policies are inherently political, and Shah and Kesan
(2007) present a case study of an attempt by the Commonwealth of Massachusetts to
implement an open standards policy in the face of significant opposition by proprietary
software vendors.4
III. PATENTS AND STANDARDS
A. Patenting Standards
While standards themselves are not generally patentable, products manufactured in

accordance with the protocols and parameters specified by standards (often referred to
as standards-compliant products) generally satisfy the statutory requirements for patent
protection. The owners of patents covering these standardized technologies are typically
the firms and institutions that employ individuals who make particular inventive contri-
butions to standards.5 Some of these contributions may be made jointly and owned by
multiple firms, but in most cases firms individually submit technical contributions to the
standard-setting process and own the resulting patents. Because standards documents
are often quite lengthy and complex, sometimes running to hundreds or thousands of
pages, multiple inventive concepts are frequently embodied in the same standard, leading
to the possibility of multiple patents covering any given standard. This being said, the
majority of standards are covered by few, if any, patents, and a small but important
minority are covered by hundreds if not thousands of different patents. Bekkers et al.
(2014, p. 19).6
B. Quantifying Standards-Essential Patents
Numerous efforts have been made to estimate the number of patents that cover particular
standards. Several SSOs permit or require their participants to disclose patents that are
likely to be necessary for the manufacture and use of a standards-compliant products (see
discussion of SSO disclosure requirements in section V.A below). SSO databases of these
“standards-essential patents” (SEPs) thus provide much of the raw data that is used to
calculate the patent coverage of standards developed at these SSOs.
4
The standards war between the ODF and OOXML open document formats is further
described in Blind (2011) and Ernst (2012).
5
SSOs themselves seldom if ever seek to claim any patent rights in standards that they produce.
6
Of 1486 standards studied, only seven were subject to more than 100 patent disclosures.
These seven standards represented 72 percent of all patent disclosure statements made. Bekkers et
al. (2014, p. 20).

1. Number of disclosures—aggregate

One of the first studies seeking to quantify aggregated SEP data was conducted by
Simcoe (2007a), who reviewed approximately 1300 patent disclosures at nine telecom-
munications-focused SSOs between 1981–2004. This data shows a dramatic increase in
patent disclosures beginning in the early 1990s, with the greatest number of disclosures
occurring at ETSI, IEEE, and the International Telecommunication Union (ITU). Blind
et al. (2011) reviewed approximately 8000 patent disclosures made at 11 large European
and U.S.-based SSOs. They find that the large majority of these disclosures were made at
SSOs focusing on telecommunications standards, of which the lion’s share were disclosed
at ETSI. Baron and Pohlmann (2018) collected more than 200 000 patent disclosures from
19 major SSOs. Of these, nearly 170 000 patents were disclosed at ETSI alone (relating to
more than 1300 different standards). Other SSOs identified by Baron and Pohlmann as
having more than 1000 patent disclosures each were Blue-ray, ISO, IEEE, ITU, and DVB
Forum (id.). Bekkers et al. (2014) observed that SDO patent disclosures have increased
over time (doubling every five years), in terms of both number of patents disclosed and
number of disclosure statements filed with the SDO.
2. SEP holders

Each of the authors described in the preceding section also studied the organizations
making SSO patent disclosures. Blind et al. (2011) identifies 292 such patent holders,
the top five being Nokia, Qualcomm, InterDigital, Ericsson, and Motorola. Baron and
Pohlmann (2018), utilizing their larger data set, identify more than 2000 different patent
holders, led by Qualcomm, InterDigital, LG Electronics, Nokia, Samsung, and Ericsson.
Simcoe (2007a) observes a significant increase in patent disclosures by small firms over
the period of his study (1981–2004), suggesting a rise in the activity of small, specialized
technology firms that participate in a few select SSOs. Bekkers et al. (2014, Table 8) clas-
sify the 334 largest organizations making SEP declarations at 13 major SDOs, finding that
approximately 71.2 percent of declarations are made by “downstream” product or service
suppliers, 12.8 percent are made by component suppliers, 5.7 percent by software firms,
and 7.6 percent by “upstream” technology or R&D providers, including universities.
3. Geographic distribution

Baron and Pohlmann (2018) tabulate the number of SEP declarations by country/region,
finding that in the aggregate, the greatest number of declared SEPs were filed in the U.S.
(over 50 000), followed by Japan, China, and Europe. Pohlmann and Blind (2016), analyz-
ing the same data set over time, observe that while the greatest number of declared SEPs
in the 1990s originated in the U.S., Europe, and Japan, since the early 2000s the number
of SEPs from China, Korea, and Taiwan have increased, with a comparative decline in
SEPs from Germany, Japan, and the U.S. Focusing on IETF, Contreras (2014) finds that
between 2001 and 2012, patent disclosures by Japanese firms remained relatively constant,
while disclosures by Chinese firms increased from zero before 2006 to 23 percent of all
IETF patent disclosures in 2012.
4. Case studies

The industry-wide surveys described above are complemented by a number of case stud-
ies quantifying patent disclosure at individual SSOs. Bekkers and West (2009) compare

patenting activity in ETSI’s GSM program (finalized in 1990) with its subsequent UMTS
standardization program (finalized in 1999). They find an eightfold increase in the number
of disclosed essential patents for UMTS (1227) over GSM (140), as well as a threefold
increase in the number of patent holders (23 to 72). Despite these increases, they find that
firm-level concentration of patents increased materially between the GSM and UMTS
projects, with CR4 concentration ratios7 increasing from 52 percent for GSM to 72
percent for UMTS, and CR8 ratios increased from 73 percent for GSM to 92 percent for
UMTS. A similar case study was conducted by Caviggioli et al. (2015) on approximately
13 000 patents disclosed with respect to later versions of the 4G LTE standard. In contrast
to the earlier study, they find a lower degree of concentration with respect to LTE SEPs
than earlier ETSI standards (CR4 concentration of 47 percent), which they attribute to
recent market entry by firms ranging from Apple to smaller Asian manufacturers. Cyber
Creative (2013), a Japanese firm, further analyzes data on nearly 6000 LTE SEP declara-
tions by SEP holder, country of origin, and year of declaration.
Another observation made by several researchers is the striking predominance of spe-
cific SEP disclosures at ETSI over all other SDOs. As noted above, Baron and Pohlmann
(2018) found that from more than 200 000 patent disclosures at 19 major SSOs, nearly
170 000 patents were disclosed at ETSI. Similar results were obtained by Bekkers et al.
(2014) and Bekkers et al. (2017). Bekkers et al. (2017) theorize that the high number of
SEP disclosures at ETSI is attributable, at least in part, to ETSI’s prohibition on “blanket”
licensing statements which permit a patent holder to avoid specific disclosures so long as
it commits to license all of its patents essential to a particular standard on specified terms
(e.g., royalty-free or FRAND).
C. The Value of Standards-Essential Patents
As the above studies indicate, the disclosure of standards-essential patents at SSOs, at

least in certain high-technology sectors, has increased over the past two decades. This
increase can potentially be explained if SEPs are perceived to have a higher value than
non-SEPs. Several authors have hypothesized that SEPs are valuable both as bargaining
chips in cross-licensing negotiations and as a basis for direct licensing revenue (Bekkers,
Bongard, and Nuvolari, 2011; Pohlmann and Blind, 2016). Accordingly, numerous
studies have sought to measure the value of SEPs based on a variety of metrics.
1. Patent status

Bekkers et al. (2014, Fig. 10), in reviewing European SEPs declared at 13 major SDOs,
found that, as of 2013, approximately one-third of such declared SEPs were not yet
issued (i.e., pending patent applications), 15 percent had lapsed due to non-payment of
maintenance fees, and 5 percent had expired.8 However, Pohlmann and Blind (2016),
using a sample of more than 200 000 declared SEPs and a matched sample of non-SEPs
7
CR4 and CR8 concentration ratios measure the total output produced in a given market by
a specified number of firms (four in the case of CR4 and eight in the case of CR8). Thus, a higher
ratio indicates that the top firms in the market produce a greater share of total output.
8
Similar results were observed for U.S. SEPs.

from corresponding patent offices, observe that holders of SEPs are more likely to pay
required maintenance fees to keep SEPs in force (67 percent of SEPs versus 76 percent
of non-SEPs lapsed due to non-maintenance). This finding suggests that SEPs are, on
average, viewed as more valuable than non-SEPs, thereby warranting the expenditure of
fees to maintain them.
This conclusion is also supported by the finding by Bekkers et al. (2014, pp. 33–34)
that patent applications declared as SEPs have a significantly greater chance of being
allowed and issued as patents than comparable non-SEPs. These findings apply across
substantially all SDOs and technology areas studied.
2. Patent transfers

Another metric of the value of a patent is whether it is sold or otherwise transferred to a
new owner. According to sources cited by Bekkers et al. (2014), 13 percent and 5 percent
of patents issued in the U.S. and Europe, respectively, are transferred. Bekkers et al. (2014)
studied patent transfers in Europe between 1997–2009 and found that of approximately
11 500 SEPs declared at five major SDOs, approximately 350 SEPs were subject to a
transfer of ownership between independent entities (i.e., excluding reassignments among
related entities). Of these, 93 were effected in connection with a larger corporate merger
or acquisition transaction, while 253 SEP transfers were effected independently of a larger
corporate transaction. The large majority of these transfers occurred in the last year of
the period studied (2009), with almost no transfers occurring prior to 2005.
Pohlmann and Blind (2016) find that SEPs are somewhat more likely than non-SEPs
to be transferred by their owners (12 percent of SEPs versus 10 percent of non-SEPs
transferred at least once during their lifetimes), suggesting a higher value or utility for
SEPs, at least in the perception of the transferees. Pohlmann and Blind also identify the
most active purchasers (Google, Qualcomm, Apple, BlackBerry, and Intel) and sellers
(Motorola, Nokia, Ericsson, Interdigital, and Panasonic) of SEPs in their sample.9
3. Market share

Bekkers, Duysters, and Verspagen (2002) observe a positive correlation between the own-
ership of patents covering the 3G GSM standard and firms’ share of the European market
for telecommunications equipment. Bekkers, Verspagen, and Smits (2002) also analyze
the impact of patents on development of the GSM standard, noting in particular the
ability of Motorola to enter and exert influence over the largely European standardization
effort due to its strong patent portfolio. They attribute the intensification of patenting
activity in the telecommunications sector, in large part, to Motorola’s demonstrated
success at influencing the GSM standardization process in the early 1990s. These studies
suggest that, at least in the case of GSM, patents contributed to the commercial success
of firms in the market for standardized products.
4. Citations
A number of studies seek to assess the quality of disclosed SEPs on the basis of forward
citations, or the number of times that a particular SEP is cited as prior art by later
9
See also section VI.E discussing transfer of FRAND and other SSO licensing commitments.

patent applications.10 Rysman and Simcoe (2008) study 724 patents disclosed as essen-
tial at four major SSOs. They find that, even before disclosure at the SSO, SEPs receive
twice as many citations as non-SEPs in the same technical field and year of applica-
tion, suggesting that SSOs select important technologies for standardization. After
disclosure at the SSO, SEPs gained an additional 19–47 percent increase in citations,
suggesting that the fact that a technology becomes standardized itself increases the
value of the underlying patents.11 Consistent results showing greater citation frequency
for SEPs are obtained by Simcoe, Graham, and Feldman (2009), Layne-Farrar (2011),
Bekkers et al. (2014), Baron and Pohlmann (2018), Caviggioli et al. (2015), Pohlmann
and Blind (2016), and Bekkers et al. (2017). Bekkers, Bongard, and Nuvolari (2011)
study a sample of approximately 10 000 patents, approximately 750 of which were
disclosed as essential to ETSI’s W-CDMA standard. They determine that both patent
quality (measured in terms of forward citations) and the patent holder’s involvement
in the standardization activity are strong determinants that a patent will be disclosed
as essential to the SSO.12
In a somewhat contrary vein, Bekkers and West (2009) analyze SEPs disclosed in
ETSI’s GSM and UMTS standardization programs. They detect a decline in the number
of citations of UMTS SEPs as compared with the earlier GSM SEPs, despite the sub-
stantial increase in patents disclosed in the UMTS program. They posit that this decline
in citations (and, in theory, value) may be due to a proliferation of patents claiming only
incremental or minor technical advances.
5. Litigation
Another measure of the potential value of SEPs is the frequency with which such patents
are litigated. In theory, once a technology becomes standardized and is thus harder to
design around, infringement of patents covering that technology should also increase.
Validating this hypothesis, Simcoe (2007a) finds that SEPs are ten times more likely than
comparable non-SEPs to be litigated; Simcoe, Graham, and Feldman (2009) find that the
rate at which SEPs are litigated is 9.4 percent versus 1.7 percent for comparable non-SEPs;
Bekkers et al. (2014, Fig. 34) find that SEPs are approximately 4.4 times as likely as com-
parable non-SEPs to be litigated; and Bekkers et al. (2017) find that SEPs are two to three
times more likely than non-SEPs to be asserted in litigation. Pohlmann and Blind (2016),
using a large sample of SEPs and comparable non-SEPs, find that the average number
of litigated SEPs in the sample was 1.93 percent, compared to 0.45 percent in the control
group. Pohlmann and Blind (2016) also tabulate the number of declared SEPs subject to
litigation by owner, finding that the majority of litigated SEPs are held by a handful of
large players (e.g., Qualcomm (20 678 SEPs and 888 litigated SEPs), Interdigital (12 522
SEPs and 978 litigated SEPs), Nokia (13 393 SEPs and 557 litigated SEPs), Panasonic
10
The number of times a patent is cited as prior art in subsequent patent applications has
become a widely used measure of the value of a patent in economic analyses (Trajtenberg, 1990;
Albert et al., 1991; Harhoff et al., 1999).
11
This effect has been referred to as the “value of the standard” (Siebrasse and Cotter, 2017;
Sidak, 2016).
12
See section V.A.2 regarding differing definitions of essentiality and over-declaration of
patents as essential to certain standards.

(6326 SEPs and 572 litigated SEPs), Samsung (10 618 SEPs and 502 litigated SEPs), and
Ericsson (9396 SEPs and 467 litigated SEPs).13
Gupta and Snyder (2014) offer a contrasting view in the context of the “smart phone
wars”, in which they argue that litigation is driven primarily by non-SEPs covering
implementation or design features of particular devices, and that fewer than one-third of
patents involved in smart phone litigation are SEPs.
Bekkers et al. (2017) find, not surprisingly, that the probability of litigation is lower for
patents subject to royalty-free licensing commitments than FRAND licensing commit-
ments. Nevertheless, as described by Contreras (2016b), outsiders to the standards process
can, and do, assert patents that may prove to read on elements of royalty-free standards.14
Moreover, such patents may be asserted against technologies that are not compliant with
the standard, and therefore do not enjoy the benefits of the licensing commitments made
with respect to the standard.
Some SEPs appear to be litigated more than others. Contreras (2017) finds that among
seven widely adopted standards, the greatest number of U.S. litigation assertions (35
percent) occurred with respect to SEPs covering IEEE’s 802.11 standards. Lemley and
Simcoe (2018) studied the SEPs asserted by non-practicing entities (NPEs), and found
that 85 percent of NPE SEP assertions related to IEEE standards.
In addition to the incidence of SEP litigation, researchers have studied the litigation
results of plaintiffs asserting SEPs. As with most litigation, most SEP cases appear to
settle. Contreras (2017) and Lemley and Simcoe (2018) both found that SEP cases settled
in the range of 70 percent of the time. As for cases litigated to a decision on the merits,
Contreras (2017) found that plaintiffs asserting SEPs are far more successful when they
are producing entities (winning nearly 53 percent of such actions) as opposed to NPEs
(winning only 20 percent). Lemley and Simcoe (2018) found no statistically significant
difference in overall win rates between assertions of SEPs and non-SEPs, a result that
tends to refute the idea that SEPs are inherently stronger (and thus more valuable) than
non-SEPs.
6. Firm performance

Pohlmann, Neuhäusler, and Blind (2015) and Mallinson (2015) approach the question of
SEP value from the standpoint of firm financial performance. Pohlmann, Neuhäusler, and
Blind analyze the return on asset (ROA) performance of 134 firms involved in standards
development over an eight-year period. They find a curvilinear (inverse U) relationship
between a firm’s disclosure of SEPs and ROA, indicating that SEPs contribute positively
to a firm’s ROA up to a point, but that excessive SEP disclosure (over-disclosure) is cor-
related with diminished performance. They also find that the marginal impact of SEPs
is stronger when disclosed to informal SSOs than formal, recognized SSOs. Mallinson
collects data on R&D spending and licensing revenue from major holders of SEPs
13
Interestingly, several large Asian firms appearing in Pohlmann and Blind’s study hold large
numbers of SEPs, but have asserted none of them in litigation: NTT Docomo (4216 SEPs), NEC
(2299 SEPs), ZTE (1640 SEPs), Datang (458 SEPs), and NTT (454 SEPs). Even more surprising
is Texas Instruments, one of the earliest firms to seek to monetize its patent portfolio, which holds
487 SEPs, with none litigated within the parameters of the study.
14
Discussing, specifically, Rembrandt v. Samsung (2015).

c overing the 4G LTE standard, as well as other statistics regarding the smart phone
market, to argue that patent licensing fees in this market are modest by comparison to
overall innovation-based gains.
Pohlmann and Blind (2016) combine a variety of the above factors—including market
coverage, technical relevance, citation count and litigation—to measure the value of
individual firms’ SEP portfolios using the proprietary IPLytics platform.
D. Acquisition and Timing of Declaration of Standards-Essential Patents
Given the potential value of SEPs and the market benefits that they confer on their
owners, firms have significant incentives to maximize the number of SEPs that they
hold.15 Several authors have hypothesized that firms participating in SSOs may thus
manipulate the patent prosecution process in order to obtain patents that are likely to be
SEPs (Hunt, Simojoki, and Takalo, 2007). Hovenkamp (2008) cites the allegations made
in the Rambus cases16 to argue that liberal continuation and division practices in patent
prosecution enable opportunistic applicants to circumvent SSO disclosure policies and
obtain patents directed to specific technical features under discussion at an SSO. To test a
similar hypothesis, Berger, Blind, and Thumm (2012) analyze the prosecution history of
approximately 300 SEPs disclosed at ETSI. Compared to similar non-SEPs, they find that
the SEPs have substantially more claims and longer pendency times. They argue that these
results suggest that patent applicants participating in SSOs strategically shape their claims
to cover standards under development and drag out the prosecution process in order to
ensure that their claims cover the latest possible version of the standard. Similar delays
were observed by Caviggioli et al. (2015) in their study of LTE SEPs.
Kang and Bekkers (2015) studied the behavior of 939 individual participants from
53 different firms who attended 77 3GPP meetings over a 12-year period during which
the W-CDMA and LTE wireless telecommunications standards were developed. They
observed a phenomenon that they term “just-in-time-patenting”: SSO participants apply
for patents of “low technical merit”17 in large quantities immediately before an SSO
meeting, then send participants to the meeting to negotiate the inclusion of the patented
technology into the standard. They further observe that this tactic is concentrated among
vertically integrated firms, typically those which champion the incumbent standard.18
15
This sub-section focuses on patenting behavior by firms that are involved in the SSO stand-
ardization process. Wen et al. (2015) offer evidence supporting the hypothesis that standardization
in a technical area tends to reduce defensive (strategic) patenting by firms that are not engaged in the
standardization process, presumably due to a lower need for patent assets as countermeasures against
infringement claims by SSO participants that are bound to license SEPs to such non-participants.
16
According to the complaint filed by the FTC, Rambus allegedly shaped its patent claims to
cover technical discussions being held contemporaneously at JEDEC. See section V.A below.
17
Forward citation rates of SEPs filed by SSO participants in the week before and during an
SSO meeting were significantly lower than those of a control group.
18
These observations were validated by the authors’ direct observations as well as reports from
participants at SSO meetings, including one former manager who explained at a public conference
“how he would send staff to a standardization meeting, and right after the meeting, in the hotel
room, they would brainstorm how to combine elements mentioned by other participants, and then
immediately prepare patent applications on these.”

The foregoing studies suggest that patent holders involved in SSOs may opportunisti-
cally seek to maximize the number of SEPs that they obtain and declare. Layne-Farrar
(2011), however, challenges this conclusion based on an analysis of approximately 1200
U.S. patents declared as essential to the ETSI UMTS standard both before (ex ante) and
after (ex post) finalization of the standard. She finds that firms did not increase their rates
of patenting after this technology had been standardized. When viewed together with the
higher citation counts found for SEPs, Layne-Farrar hypothesizes that ex post patenting
may be attributable to some combination of innovation and opportunism, but in propor-
tions that are not well understood. Moreover, Bekkers et al. (2014) found in a study of
13 major SDOs that patents are disclosed, on average, seven years after their application
date and eight years after their priority date (Fig. 9). These timeframes would not suggest
rapid patenting of SDO technical developments for strategic advantage.
E. Effect of Patents on Standardization Activity
A few studies have attempted to assess the effect of SEPs on standardization processes
within SSOs using a range of methods, including case studies, interviews, and empirical
data analysis. Egyedi (2001a, 2001b) studied Java standardization at ECMA International
during the 1990s and cites several instances in which fear of intellectual property claims
(both patent and copyright) stifled cooperation within the SSO, eventually leading to
the collapse of the standardization process. Contreras (2008, 2016a) discusses particular
instances at IETF and the World Wide Web Consortium (W3C) in which threats of patent
disputes were the impetus for policy amendments at these two organizations. Contreras
(2013a) also surveyed members of VMEbus International Trade Association (VITA),
a small SSO with relatively few declared SEPs, and found that policy amendments that
gave members greater visibility into the maximum royalties that would be charged by SEP
holders (a so-called ex ante licensing disclosure policy) were perceived to improve a variety
of standardization processes and outcomes.
Egyedi (2016) interviewed participants in ITU’s H.265 standardization process and
found that technical design choices were made for a range of practical and business
reasons, often having little to do with the technical merit or inventiveness of a particular
solution. It is not clear, however, to what degree patents were declared essential to the
H.265 standard or what impact, if any, those patents may have had on these decisions.
Baron et al. (2013) analyzed 3500 standards released between 1998–2008 and found
that higher concentrations of SEPs caused standards to be less likely to be replaced, but
increased the likelihood that they would be upgraded through new version releases. They
attribute this effect to frictions and vested interests in existing patented technologies (i.e.,
a lock-in effect based on SEP ownership).
The foregoing studies suggest that the existence of patents covering technologies being
standardized has an impact, generally negative or neutral at best, on the standardization
processes studied. It is important to note, however, that none of these studies focused
specifically on the most patent-intensive standards developed at ETSI, ITU and IEEE;
nor was the value of patented technologies to such standards assessed. Further research
would be beneficial regarding the interplay of patents and standardization processes.

F. Standards as Prior Art
One of the challenges that faces patent offices examining applications for patents covering
standardized technologies is the unavailability of SSO documents that may shed light on
the originality and inventorship of such technologies. Thus, while internal SSO docu-
ments can serve as valuable prior art in the examination of such patent applications, these
documents are often inaccessible to patent offices or, even if accessible, are not indexed
or searchable in a useful manner.
To address these concerns, the European Patent Office (EPO) has entered into coopera-
tive arrangements with several leading SSOs including ETSI, ITU, and IEEE. EC (2014),
NRC (2013). Under these arrangements, the SSOs have agreed to share a variety of docu-
mentation with the EPO for use in patent examinations. As reported by Pohlmann and
Blind (2016), in 2015 the EPO cited over 19 000 standards-related documents as prior art.
Other standards bodies may have other arrangements. For example the U.S. Patent and
Trademark Office and the EPO are both members of the DVB Project. Their membership
gives them access to all DVB technical documents. In addition, new collections of SSO
documents aggregated and organized by academic researchers may be of use in patent
office examinations (Baron and Spulber, 2018).
G. Potential Market Effects: Hold-up, Hold-out, and Stacking
There is significant debate across the industry and academy regarding the impact of
patents on the development and distribution of standardized products. Régibeau et al.
(2016) classified these potential issues and gauged their relative importance to a range
of stakeholders using a consultation exercise involving 40 respondents and a series of
36 interviews. It found that injunctions, hold-out, hold-up, royalty stacking, and over-
declaration of SEPs were perceived as the “main problems” caused by SEPs. Each of these
issues is discussed in greater detail below.
1. Hold-up
There is a large and varied theoretical literature concerning the potential effects that
patents covering standards may have on product development and markets. One of the
principal areas of debate concerns the potential for SEP owners to “hold-up” the market
by demanding excessive royalty rates after a standard has been widely adopted and
manufacturers which have made investments in the standardized technology have become
“locked-in” (Shapiro, 2001; Swanson and Baumol, 2006; Farrell et al., 2007; Lemley and
Shapiro, 2007; Lichtman, 2010; Chien and Lemley, 2012; Chien, 2014). In addition to
raising costs for potential competitors, it has been theorized that patent hold-up can
increase consumer prices and hinder innovation. The potential for standards-based patent
hold-up has been echoed by antitrust and competition authorities in the U.S., Europe, and
elsewhere (DOJ-FTC, 2007; FTC, 2011; FTC, 2012; EC, 2014), as well as by courts adju-
dicating standards-related cases (Microsoft, 2013; Ericsson, 2014). Ernst, Lee, and Kwak
(2014), taking an international perspective, argue that strategic patenting of standardized
technologies can stifle economic development, particularly in less-developed economies.
Siebrasse (2019) offers a survey of the theoretical literature concerning hold-up for
standardized products.

Nevertheless, systematic evidence of standards-based patent hold-up in the market has

not been collected, and reports of hold-up by individual firms are criticized as unreliable
or too specific to support general conclusions. As a result, several commentators have
questioned whether the threat of patent hold-up should be of concern, citing, among other
things, a lack of empirical evidence of systematic hold-up in information and communica-
tions technologies (ICT) industries (Geradin and Rato, 2007; Spulber, 2008; Sidak, 2008;
Gupta, 2013; Kieff and Layne-Farrar, 2013; Layne-Farrar, 2014b; Sidak, 2015a, 2015b). In
a recent study, Galetovic, Haber, and Levine (2015) measure market characteristics that have
traditionally been associated with industries subject to hold-up (high quality-adjusted prices
and low rates of innovation) in markets characterized by significant patenting and standards
(SEP-reliant industries). They find that SEP-reliant industries such as digital communica-
tions exhibit significantly faster price reductions and higher levels of innovation than held-
up industries such as electrical power, tending, they assert, to refute the hold-up hypothesis.
Epstein, Kieff, and Spulber (2012) are particularly critical of governmental agencies that
have taken action in response to perceived threats of hold-up, and view these agencies as
themselves introducing various market inefficiencies. Contreras (2019c) questions the need
for empirical evidence of systemic market hold-up as a basis for policy reform in this area.
Commentators have also drawn attention to the potential for opportunistic behavior
by standards implementers (licensees), which has been referred to as “reverse hold-up” or
“hold-out” (Geradin, 2010; Kieff and Layne-Farrar, 2013; Camesasca et al., 2013; Chien,
2014; Cotter, 2014; Sidak, 2015a; Siebrasse, 2019). In hold-out situations, a potential licen-
see may refuse to pay a “reasonable” royalty rate to use an SEP in a standards-compliant
product. If the SEP holder itself is bound to charge no more than a reasonable royalty,
and if it is unable to seek injunctive relief to prevent the implementer’s infringement, the
SEP holder has little recourse but to sue the implementer for patent infringement and
recover, at most, the reasonable royalty that it would have received in the first place.19
Thus, under this theory, standards implementers would have a significant incentive to
hold out for as long as possible. The Federal Trade Commission (FTC, 2011) identifies
hold-out situations as instances in which it may be reasonable for SEP holders to seek
injunctive relief to prevent ongoing infringement by recalcitrant implementers (Motorola
and Google, 2013). Nevertheless, as for hold-up, there is little empirical evidence relating
to the prevalence of hold-out in standard-setting environments.20
2. Royalty stacking

As described by the U.S. Court of Appeals for the Federal Circuit:
Royalty stacking can arise when a standard implicates numerous patents, perhaps hundreds, if
not thousands. If companies are forced to pay royalties to all [patent] holders, the royalties will
“stack” on top of each other and may become excessive in the aggregate (Ericsson, 2014).
19
Contreras and Gilbert (2015) argue that FRAND royalties should be calculated in essentially
the same manner as reasonable royalty damages. Nevertheless, the availability of enhanced damages
for willful infringement could deter hold-out behavior in some circumstances (Contreras, et al., 2019).
20
In one early interview-based study, Blind and Iversen (2004) found that of large companies
involved in standardization, 40 percent responded that their licensing conditions had not been
accepted, and 35 percent had experienced infringements of their patents.

Lemley and Shapiro (2007) have argued that stacking will result in higher prices for con-
sumers because (1) patent holders and product manufacturers will each seek to maximize
their margins (a phenomenon, which is not unique to patent licensing, known as “double
marginalization”), and (2) holders of complementary patents will increase their royalties
to a level that, in the aggregate, will depress sales of end products and, in turn, reduce their
individual profits (the well-known problem of “Cournot complements”). They also argue
that stacking is likely to exacerbate and amplify patent hold-up behavior by introducing
multiple patent holders with potential hold-up power.
Interestingly, though the large number of patents covering various standardized
technologies has been amply documented (see section II.B above), relatively little public
data is available regarding aggregate patent royalty rates for standardized products. In the
mid-2000s, a number of SEP holders voluntarily disclosed their standard licensing rates
for wireless telecommunications standards (Contreras, 2015c, Table 4), and there was at
least one attempt by a group of network operators to establish maximum aggregate roy-
alty rates for certain wireless standards through the Next-Generation Mobile Networks
(NGMN) consortium (Contreras, 2013a, pp. 178–79). More recently, evidence regarding
SEP royalty rates has emerged in litigation over SEP licensing terms (see section V.B.1
below). For example, in Microsoft (2013), the court, in analyzing the patent holder’s pro-
posed royalty rates, found “significant stacking concerns.” Specifically, it observed that:
There are at least 92 entities that own 802.11 [standard-essential patents]. If each of these 92 enti-
ties sought royalties similar to [the patent holder’s] request of 1.15% to 1.73% of the end-product
price, the aggregate royalty to implement the 802.11 Standard, which is only one feature of the
Xbox product, would exceed the total product price.
On this basis, the court determined that the royalty sought by the patent holder was
unreasonable, because “if everyone wanted the same deal, it would quickly make the
end-product price untenable commercially.” Microsoft (2013).21 In Innovatio (2013b), the
court was also required to calculate the “reasonable” royalty for patents covering different
aspects of the 802.11 standard. In doing so, it considered “the total royalties an imple-
menter would have to pay to practice the standard” and “whether the overall royalty of all
standard-essential patents would prohibit widespread adoption of the standard.” Bartlett
and Contreras (2017) discuss the methods used to determine FRAND royalties in five
different reported cases involving the 802.11 standard, suggesting not only that royalty
stacking for this standard may be an issue, but also that judicial royalty determinations,
as they are currently undertaken, are both inconsistent and unpredictable.
Outside of judicial proceedings, however, data regarding patent royalty rates is typically
protected by confidentiality agreements, making the collection and analysis of such data
difficult. Given these constraints, Contreras et al. (2016) have proposed mechanisms for
making more royalty data available to researchers and policy makers.
Notwithstanding the difficulty of obtaining FRAND royalty data, there is a growing
body of empirical literature seeking to quantify the aggregate patent royalty burden for
various standardized technologies. Geradin, Layne-Farrar, and Padilla (2008) review
earlier (non-standards based) empirical studies of royalty stacking in industries such as
21
But Layne-Farrar and Wong-Ervin (2014) use numerical examples to critique this approach.

semiconductors, software, and biomedicine, finding that stacking, while present, did not
seriously impair those industries. They also assess the potential for royalty stacking on
ETSI’s 3G WCDMA and IEEE’s 802.11 Wi-Fi standards, generally disputing estimates
of stacking offered by Lemley and Shapiro (2007).
Stasik (2010) analyzes the aggregate royalty burden on ETSI’s 4G LTE standard, as to
which several major patent holders have voluntarily announced royalty rates or ranges.
Based on early announcements by nine major LTE SEP holders, Stasik finds an aggregate
royalty burden of 14.8 percent of the end product price. Armstrong, Mueller, and Syrett
(2014) quantify the royalty “stack” for smart phone devices by identifying and aggregating
known and estimated royalty rates for the different functional subsystems of such devices.
They find that, absent cross-licensing and other royalty-reducing measures, the aggregate
patent royalty for a hypothetical $400 smart phone would be $120, or 30 percent of the
end product price.22
Other authors, however, contest the evidence of royalty stacking in standardized
technologies. In addition to disagreement over the theoretical claims made by Lemley
and Shapiro (2007), they observe that despite the large number of patents covering
standards for technologies such as 3G and Wi-Fi, these technologies have flourished in
the marketplace, suggesting that in practice royalty stacking may not present a significant
market risk (Geradin, Layne-Farrar, and Padilla, 2008; Sidak, 2008 and 2016b; Gupta,
2013; Mallinson, 2015; Galetovic et al., 2018). Mallinson (2015), and Sidak (2016b) and
Galetovic et al. (2018) estimate aggregate royalty burdens on mobile telecommunications
devices in the 4–5 percent range.
Galetovic and Gupta (2016) look to secondary market effects to draw conclusions
about whether royalty stacking has occurred in the mobile wireless industry. They observe
that from 1994–2013, the number of firms holding SEPs covering ETSI’s 2G, 3G, and
4G standards increased from two to 130. At the same time, they observe that the number
of devices implementing these standards sold annually increased by approximately 20
percent, average selling price dropped by up to 24.8 percent annually on an adjusted
basis, and the number of device manufacturers grew from one to 43. These indicia, they
contend, suggest that the market is not subject to royalty stacking.
In an effort to link claims of harm arising from royalty stacking to actual evidence of
harm, the court in Ericsson (2014) required that claims alleging damage due to royalty
stacking be supported by actual evidence of such harm to the claimant (Sidak, 2016). The
court’s reasoning is questioned by Contreras (2015d), who argues that royalty stacking
with respect to a standard may have an impact on the reasonableness of any particular
SEP holder’s individual royalty rate, irrespective of the magnitude of royalties paid by
the implementer at the time of an infringement suit by the SEP holder. Requiring the
implementer to show specific harm from royalty stacking could thus result in a race to
the courthouse by SEP holders, each seeking to levy a royalty before the aggregate paid
by the implementer becomes unreasonably high.
22
Mallinson (2015) challenges this result, estimating an aggregate smart phone royalty burden
of approximately 5 percent, based on estimated industry-wide annual U.S. smart phone SEP
licensing revenue of $19 billion. Layne-Farrar (2014b) also disputes the analysis undertaken by
Armstrong, Mueller, and Syrett (2014) on several counts.

H. Patent Pools and Standards
Several important standards, particularly in the consumer electronics industry, were

developed by small groups of firms that make their SEPs available through patent pools.
Well-known patent pools exist for popular families of standards such as MPEG, CD,
and DVD. Patent pools have also been formed for standards that were developed in
larger SSOs, often containing only a subset of the known SEPs covering those standards.
Examples include the Via Licensing and Sisvel pools for IEEE’s 802.11 standard and
MPEG-LA’s pool for ITU’s H.264 standard (Microsoft, 2013; Bekkers and Updegrove,
2012; Bekkers et al., 2014). The DVB Forum offers a unique example of a developer
of voluntary consensus standards, all members of which participate in a patent pool
(Eltzroth, 2008).
Patent pools offer numerous efficiencies in licensing and are generally characterized
by license terms (including royalties) that are published and uniform across all licensees,
though care must be taken to avoid anti-competitive collective activity in the formation
and operation of such pools (Shapiro and Varian, 1999; Shapiro, 2001; Contreras, 2013b;
Lundqvist, 2014). Despite the potential benefits offered by pools, relatively few patent
pools have been formed around technical interoperability standards. Biddle, White, and
Woods (2010) find that of 251 standards implemented in a typical laptop computer, only
3 percent were subject to patent pools, with the remainder subject to FRAND or royalty-
free licensing commitments. Pohlmann (2016), analyzing more than 200 000 individual
SEP declarations, finds that only 9 percent of declared SEPs are pooled. Bekkers et al.
(2014) analyze the number of patent pools launched following release of a standard.
They find that more than 40 patent pools were formed following the release of technical
standards, but the majority of these pools failed to attract significant membership or
coverage of the relevant technology. Fewer than one-third of the firms declaring SEPs
with respect to these standards were members of the relevant pools. There are several pos-
sible explanations for the relative scarcity of patent pools in the field, including significant
upfront costs associated with evaluating pooled patents for essentiality23 (Contreras,
2018). There is a large empirical and theoretical literature devoted to patent pools, both
standards- and non-standards related.
Though not a patent pool per se, another novel structure that sought to offer aggregated
SEP licenses was Intellectual Property Exchange International (IPXI), a trading exchange
that offered unitized license contracts covering 194 patents declared essential to IEEE’s
802.11n standard. For various financial and business reasons, IPXI ceased operations in
2015 (Contreras, 2016c).
23
Unlike SEPs subject to licensing commitments by the patent holder, current interpretations
of antitrust law require that patents contributed to a pool be found to be essential to the standard
by an objective evaluator (DOJ-FTC, 2000; DOJ-FTC, 2007). Régibeau et al. (2016) report that
the estimated cost of a third-party patent essentiality assessment is approximately €9000 (p. 50),
and that imposing such a cost on ETSI’s 2G/3G/4G standards would result in an aggregate cost of
approximately €427.5 million (p. 59).

I. Patent Assertion Entities and Standards-Essential Patents
A significant amount of recent empirical literature seeks to measure the quantity and
effect of litigation activity by so-called patent assertion entities (PAEs) (FTC, 2016; JRC,
2016). Most of this work is centered in the U.S. and Europe. Concurrently, there have
been numerous proposals to amend legislation and regulations to address perceived issues
arising from PAE litigation.
Several prominent reported cases involve assertion of SEPs by PAEs, including N-Data
(2008), Innovatio (2013) and Unwired Planet (2017), among others. However, there has
been comparatively little empirical work studying the accumulation and assertion of SEPs
by PAEs.
Contreras (2016b, 2017) measured assertion of SEPs covering seven widely adopted
standards in U.S. district courts from 2000–15. He found that 77 percent of these
assertions were initiated by NPEs, most of which were PAEs, and that both NPEs and
practicing entities asserted significant numbers of SEPs unencumbered by FRAND
or other licensing obligations. Contreras et al. (2017) have extended this study to cover
SEP assertions in Germany and the U.K., finding that NPEs (particularly Sisvel) play a
significant role in SEP assertion in these jurisdictions as well. Lemley and Simcoe (2018),
in a study of SEP litigation, found that 73 percent of SEP assertions were made by NPEs,
though NPEs only asserted 37 percent of unique SEPs (indicating that NPEs made
multiple assertions of a limited number of SEPs).
In 2016, the European Commission’s Joint Research Centre published a study of
PAE behavior in Europe (JRC, 2016), which found through 18 in-depth interviews that
PAEs in Europe are acquiring and asserting significant numbers of SEPs, particularly in
the telecommunications sector. Likewise, Pohlmann and Blind (2016) find that several
NPEs (IP Asset Trust, Cluster Technology, Sisvel, Unwired Planet, Innovative Sonic) are
among the largest acquirors of SEPs. Interestingly, these results were not observed by the
U.S. FTC in a study released in the same year (FTC, 2016). The FTC, based on surveys
distributed to 22 selected PAEs whose identities were not disclosed, found that fewer
than 1 percent of the patents held by such PAEs were SEPs. This finding led the FTC
to conclude that its survey sample did not include any PAEs that focused on monetizing
SEPs (FTC, 2016, pp. 136–37).
J. Patents and Non-ICT Standards
The vast majority of literature relating to patents and technology standardization has
focused on the ICT sector. As reported, however, by Blind et al. (2011), NRC (2013),
EC (2014), and Baron and Spulber (2018), significant standardization activities occur
outside of ICT. At least two significant actions by the U.S. Federal Trade Commission
have involved the potential assertion of patents in non-ICT fields (Unocal (gasoline
additives) and Bosch (automotive electronics)). Nevertheless, on the whole patents have
played a relatively modest role in SSOs and standardization in non-ICT fields. Blind et al.
(2011) find that among seven large formal SSOs studied, 98 percent of patent disclosures
applied to ICT standards. These results are confirmed by Baron and Pohlmann (2018)
and Pohlmann and Blind (2016).
Several recent case studies of standards in non-ICT technology fields show a similarly

small, but potentially growing, role for patents: Torrance and Kahl (2014) (synthetic
biology), Kumar and Rai (2007) (synthetic biology), Contreras (2013c) (bioinformatics),
Contreras and McManis (2013) (sustainable building materials), Jillavenkatesa, Evans, and
Wixon (2012) (nanotechnology), and Contreras (2012) (smart grid). Several of these authors
observe, however, that patents could play an increasingly prominent role in standardization
in non-ICT industries as these industries mature and spur larger product markets, and
emphasize the desirability of open standards frameworks in these industries (Contreras,
2013c, 2012; Kumar and Rai, 2007). In contrast, Blind and Iversen (2004) map standards
intensity and patent intensity for nine different technology-focused industries including
ICT, aeronautics, optoelectronics, nanotechnology, pharmaceuticals, and biotechnology
(Fig. 4.4-1). They conclude that issues relating to patents and standards are the most heated
in industries with the highest levels of standardization and patent intensity, namely ICT,
and that issues are less pressing at lower-intensity industries such as nanotechnology.
IV. PRIVATE ORDERING AND SSO PATENT POLICIES
Over the past two decades, SSOs have responded to the increasing number of patents cov-
ering standardized technologies and the perceived threats of patent hold-up and stacking
by adopting a series of policy measures intended to address these concerns.24 EC (2014)
identifies seven distinct goals of SSO patent policies: (1) enabling informed decisions
about technology inclusion, alternatives and design-around, (2) ensuring that licenses
for SEPs are available, (3) preventing patent hold-up, (4) preventing patent “ambush”
and blocking, (5) preventing excessive cumulative royalties (stacking), (6) preventing
discrimination among implementers of a standard, and (7) ensuring transparency about
SEPs. Bekkers and Updegrove (2012) find, however, that few SSOs explicitly state the
goals or intended purposes of their patent policies.
SSO patent policies today fall into two general categories—disclosure policies and
licensing policies—and often include elements of both. Disclosure policies typically
require participants in the standards development process to disclose patents they hold
that are believed to be essential to the implementation of the standard. Licensing policies
typically require that participants grant implementers licenses under their SEPs on terms
that are “reasonable and nondiscriminatory” (RAND) or FRAND.25
A. Cataloging SSO Patent Policies
Numerous efforts have been made to catalog and classify the provisions of SSO patent
policies. Lemley (2002) reviewed the policies of 43 SSOs across a range of industries
and classified them based on features including whether they included a disclosure
requirement, whether participants were required to search their patent portfolios to
24
SSO policies are generally assumed to be binding on SSO participants through a series
of legal mechanisms including contract, estoppel, and antitrust/competition law (Lemley, 2002;
Contreras, 2015a).
25
The terms RAND and FRAND are, for all practical purposes, synonymous.

identify SEPs, and whether they included a licensing requirement (RAND or royalty-
free). Chiao, Lerner, and Tirole (2007) assess the policies of 59 SSOs and code for
features including form of membership (corporate versus individual), decision-making
rule (majority or consensus), and SSO age and scope (single or multi-purpose). Bekkers
and Updegrove (2012) offer an in-depth survey of the policies of 12 SSOs representing
a cross-section of organizational models, geographic region, and technology focus. They
describe in detail the many variants that SSOs have adopted regarding the mechanics
of patent disclosure (timing, knowledge, level of detail, definition of essentiality, updat-
ing) and licensing commitments (FRAND versus royalty-free, beneficiaries, duration,
field of use, geographic scope, transfer with underlying patents, suspension of licenses,
requirements that licensees license-back their own patents (reciprocity), and the patent
holder’s ability to opt-out of granting licenses under certain circumstances). Many of
these variants are also discussed in detail in ABA (2007), which was compiled by a
committee of experts involved in standardization activities. Blind et al. (2011) analyze
the policies of 22 SSOs of varying sizes and geographic scope (including one SSO
based in China) in a similar fashion. Baron and Spulber (2018) code the policies of 36
different SSOs in the ICT sector for inclusion of various key terms pertaining to patent
disclosure and licensing, as well as other procedural aspects of SSO operation includ-
ing voting mechanics, openness, and balance of interests. Régibeau et al. (2016, Ch.3)
maps different SSO policy provisions (FRAND licensing, ex ante disclosure of royalty
rates, etc.) to underlying economic issues (e.g., hold-up, hold-out, patent ambush, etc.)
and assesses the impact of such provisions on the identified issues. Baron et al. (2019)
conducted a detailed analysis of the governance mechanisms of 18 SSOs, together with
a survey of SSO stakeholders and other inputs, to present the first comprehensive study
of SSO governance.
B. SSO Policy Features: FRAND Licensing Terms
The above studies present a wealth of information regarding the policy features of SSO.
Among the most notable results is the prevalence of SSO policies that require licensing of
SEPs on terms that are at least FRAND. Of 36 SSO patent policies reviewed by Lemley
(2002), 29 contained FRAND commitments; and of 251 laptop standards identified by
Biddle, White, and Woods (2010), 75 percent were subject to FRAND commitments.
In their recent study of 36 SSO policies, Baron and Spulber (2018) find nine SSOs that
require FRAND licensing and 23 that permit the licensor to choose from a menu of
licensing options, with FRAND licensing being the least restrictive. Pohlmann and Blind
(2016) find, based on analysis of more than 200 000 SEP disclosures across a range of
SSOs, that 68 percent of such disclosures contain FRAND licensing commitments.
Though less common than SSO policies permitting SEP holders to charge royalties
at FRAND rates, some SSOs require their participants to license patents on reasonable
terms that are royalty-free. This phenomenon is discussed in detail in section VI.C, below.
Much has been written regarding the meaning of FRAND, with respect to the level
of royalties that qualify as “fair” and “reasonable”, what degree of similar treatment of
licensees is required to comply with the “non-discrimination” prong of FRAND, as well
as the reasonableness of other terms included in license agreements (e.g., reciprocity,
grant-backs, transfer of patents, confidentiality). A full discussion of these licensing

terms is beyond the scope of this chapter. Where litigation over FRAND terms has arisen,
specific issues are discussed in greater detail in section V.B below. Discussions of the many
diverse terms found in FRAND licensing agreements can be found in Baron et al. (2019,
7.1.3); Pentheroudakis and Baron (2017), NRC (2013), Bekkers and Updegrove (2012),
ABA (2007), and Lemley (2002).
C. SSO Policy Evolution and Amendment
SSO patent policies, far from being static documents, are amended and adapted with some
regularity. Tsai and Wright (2015) find that many SSOs amend their patent policies as fre-
quently as once per year or more. Most of these amendments, however, are not significant.
Major amendments to SSO patent policies have often been prompted by prominent indus-
try litigation or enforcement actions. Layne-Farrar (2014a) identifies substantial patent
policy amendments at ten major SSOs and assesses whether they are proactive, reactive or
non-responsive to four general trends in antitrust enforcement: patent ambush,26 excessive
royalty rates, transfer of patents subject to SSO licensing commitments, and attempts to
obtain injunctions on patents subject to licensing commitments. She finds that while the
first two concerns (patent ambush and royalty rates) have largely been addressed by the
SSOs studied, the last two concerns (patent transfer and injunctions), which have emerged
more recently, have not yet been addressed fully. Tsai and Wright (2015) study SSO policy
amendments pertaining to licensing rules and disclosure at 11 SSOs. They rate the amend-
ments based on their reduction of ambiguity in policy language (significant reduction,
moderate reduction, no reduction, increase in ambiguity) and find a gradual reduction
in policy ambiguity across the board. On the basis of these findings, both Layne-Farrar
and Tsai and Wright urge enforcement agencies to moderate their enforcement actions in
order to give SSOs time to amend their policies to address concerns.
In their review of 36 SSO patent policies, Baron and Spulber (2018) observe a general
strengthening of SSO licensing requirements over time, with four SSOs moving to royalty-
free or non-assertion requirements after permitting royalties to be charged on SEPs, and
two moving to a mandatory licensing requirement from no licensing obligation at all.
They observe no significant modification to disclosure requirements over the period
studied. Contreras and Housley (2008), however, discuss a clarifying amendment to the
IETF patent disclosure policy prompted by an alleged failure of a participant to disclose
a patent covering an optional portion of a draft IETF standard.
A few recent SSO patent policy amendments have generated significant controversy.
The first of these involved VITA, a medium-sized SSO focusing on avionics and defense
electronics. In 2006, VITA developed a draft patent policy amendment requiring that its
participants disclose not only patents essential to the implementation of VITA standards,
but also the maximum royalty rates they would charge for those patents (a so-called “ex
ante” licensing disclosure policy—see section VI.A below). VITA submitted its proposed
policy amendment to the U.S. Department of Justice, which issued a favorable business
review letter approving the policy as proposed (DOJ, 2006). Nevertheless, one of VITA’s
founding members objected to the change and withdrew from the organization as a result
26
See section V.A.1 below.

(Contreras, 2013a). A dispute ensued in which opponents led an unsuccessful campaign

to have VITA’s ANSI accreditation revoked.27 Within the next several months, both IEEE
and ETSI also adopted ex ante licensing disclosure policies. However, due to internal
opposition, these two SSOs made ex ante disclosures of royalty rates optional rather than
mandatory (Contreras, 2013a; Tapia, 2010). The DOJ approved IEEE’s policy amend-
ment in 2007 (DOJ, 2007).
In 2015, IEEE adopted another set of major policy revisions. These included various
clarifications regarding the meaning of the licensing commitments made to IEEE, limit-
ing the ability of participants to seek injunctive relief against willing licensees, requiring
commitments by transferees of committed patents, and permitting the arbitration of dis-
putes over licensing terms (IEEE, 2015). The DOJ approved these amendments as having
“the potential to benefit competition and consumers by facilitating licensing negotiations,
mitigating hold up and royalty stacking, and promoting competition among technologies
for inclusion in standards” (DOJ, 2015). Nevertheless, there was strong opposition to
the amendments, from both industry and commentators, much of which focused on the
DOJ’s approval of the IEEE amendments, rather than the content of the amendments
themselves (Lindsay and Karachalios, 2015; Sidak, 2015b).
The organizational and institutional mechanisms for making SSO policy changes have
been examined in depth by Baron et al. (2019, 5.2.3).
D. SSO Patent Policy Influences on Firm Membership Decisions
By and large, SSOs are voluntary membership organizations. Thus, members are free to
leave, or decline to join, SSOs that have adopted policies that they view to be sufficiently
adverse to their interests to outweigh the benefits of membership. Several authors have
attempted to correlate SSO patent policy terms with firm decisions to participate in col-
laborative standardization activities and to model competition for membership among
SSOs by means of differentiated policy documents.
Based on a survey of 149 European firms across industries, Blind and Thumm (2004)
find that firms with greater rates of patenting activity are less likely to join collective
standardization efforts due, in part, to the disclosure and licensing requirements imposed
by SSOs. These results, however, are not necessarily borne out by more focused studies of
firms in the ICT sector, in which the value of participating in SSOs is perceived to be high
(DeLacey et al., 2006). Lerner and Tirole (2006) develop a model, supported by interviews
with SSO participants, in which differing SSO rules regarding the disclosure and licensing
of patents may impact an organization’s choice of which SSO to join. That is, they view
SSOs as competing for members on the basis of their patent and other policies. Chiao,
Lerner, and Tirole (2007) empirically test some of the predictions made by Lerner and
Tirole (2006) using a sample of 59 SSO policies. Among other things, they find that SSOs
that are oriented toward a small group of sponsor firms are less likely to demand policy-
based concessions from members.
27
Despite opposition by a vocal minority, both the approval vote at VITA (35 to two in favor
of the amendments, 12 abstaining) and survey data compiled by Contreras (2013a) indicate that a
substantial majority of VITA members supported the amendments.

Two recent case studies have examined the effects of specific SSO patent policy changes
on SSO membership. Contreras (2013a) tested predictions that VITA’s adoption of a
mandatory ex ante licensing disclosure policy (see section IV.C) would drive members
from the organization. He finds that, other than the single member that most strenuously
objected to the policy change, no members departed the SSO as a result of the change, and
overall membership increased substantially in the years following the change. Stoll (2014)
examines the effects of a policy shift at OASIS, an SSO focused on software standards,
from requiring FRAND licensing to allowing individual working groups to require
royalty-free licensing of SEPs. He finds that, following the OASIS policy amendment,
overall membership levels dropped and member composition shifted, with the number
of software producers decreasing and the number of research organizations and systems
integrators increasing. These studies may be informative for future research, particularly
as the impact of IEEE’s 2015 policy amendments is observed.
V. SSO PATENT POLICY DISPUTES
As discussed in section III above, many SSOs have adopted policies that require their
participants to disclose patents that are essential to the SSO’s standards, and/or to license
those patents to others on terms that are either royalty-free or subject to FRAND royalties.
Disputes and enforcement actions concerning SSO patent policies have generally arisen in
two waves: the first, roughly from the mid-1990s to the mid-2000s, dealing with the scope
and contours of the disclosure obligation; and the second, roughly from the mid-2000s to
the present, dealing with the meaning of the FRAND licensing commitment.
A. Disclosure Disputes
1. Intentional non-disclosure—patent ambush

Four principal U.S. cases—Dell (1998), Rambus (2003 and 2008), Unocal (2005), and
Qualcomm (2008)—characterize the disputes regarding SSO disclosure policies that
emerged from the mid-1990s to the mid-2000s.28 Of these, Dell, Rambus (2008) and
Unocal were enforcement actions brought by the FTC against patent holders for alleg-
edly anti-competitive conduct in violating SSO disclosure requirements; and Rambus
(2003) and Qualcomm were private actions in which the patent holder’s alleged violation
of SSO disclosure rules was raised either as a defense to patent infringement or as an
affirmative private antitrust claim. In each of these cases, an SSO’s rules were alleged to
require disclosure of SEPs held by its participants, one of the SSO’s participants failed to
disclose one or more patents covering the SSO’s standards, and the participant thereafter
sought to exploit those patents by licensing or enforcing them against implementers of
the standard. This pattern of conduct has been referred to as “patent ambush” or, as one
commentator colorfully puts it, “snake in the grass” (Merges and Kuhn, 2009).
28
Other cases that are discussed less frequently in the literature include Wang v. Mitsubishi,
860 F. Supp. 1448 (C.D. Cal. 1993) and Stambler v. Diebold, 1988 WL 95479 (E.D.N.Y. 1988). For
a brief summary, see Contreras (2011).

One of the principal issues raised in each case was the degree to which the relevant
SSO’s policy actually required disclosure of the patents in question. In Dell, the Video
Electronics Standards Association (VESA) required individual meeting attendees to
certify that all known SEPs were disclosed. Dell’s representative to VESA signed the
required certification, but did not disclose any patents. When Dell later sought to collect
royalties under a patent covering the VL-bus standard, the FTC brought an action accus-
ing Dell of engaging in unfair methods of competition under Section 5 of the FTC Act.
The FTC action resulted in the entry of a 1996 consent decree permanently enjoining Dell
from enforcing its VL-bus patents against any third party. The Dell case is controversial,
as there was no allegation that Dell’s representative to VESA knew of Dell’s patent when
the VL-bus standard was adopted. In response, SDO policies today often specify that
searches of corporate patent portfolios are not required to comply with SSO disclosure
requirements, and that disclosure be limited to the “knowledge” of the individuals partici-
pating in standards development at the SSO (ABA, 2007; Bekkers and Updegrove, 2012).
The Rambus and Qualcomm disputes also centered on the language of SSO disclosure
policies. In Rambus (2003), the Federal Circuit sharply criticized the patent policy of the
Joint Electron Device Engineering Council (JEDEC), an SSO in which Rambus, Inc.
participated in the early 1990s. Though Rambus was accused of concealing patents that
it filed specifically to cover nascent JEDEC standards, the court declined to find that
Rambus had committed fraud because JEDEC’s patent policy did not clearly impose a
disclosure requirement. The court criticized the policy as suffering from “a staggering lack
of defining details” that left SSO participants with only “vaguely defined expectations
as to what they believe the policy requires”. The court concluded that, while Rambus’s
attempt surreptitiously to patent JEDEC standards might “impeach Rambus’s business
ethics, the record does not contain substantial evidence that Rambus breached its duty
under the . . . policy.” Rambus (2003).29 The holding in the Rambus (2003) case served as
a wake-up call to the SSO community and prompted many SSOs to clarify their patent
policies in an effort to avoid some of the perceived weaknesses of the JEDEC policy
(Tsilas, 2004; ABA, 2007; Layne-Farrar, 2014a).
Qualcomm (2008) involved a similar interpretive dispute over whether an SSO policy
that “encouraged” participants to disclose their essential patents actually required such
disclosures. Basing its decision on the practices and expectations of SSO participants,
other language concerning “best efforts” to disclose, and the rules of related SSOs, the
court held that the SSO policy imposed an affirmative obligation to disclose, and that
Qualcomm had breached that obligation (p. 1019).
29
In a subsequent action, the FTC found Rambus liable, among other things, for attempted
monopolization in violation of Section 2 of the Sherman Act and deceptive conduct under Section
5 of the FTC Act (Rambus, 2008). The FTC’s decision, however, was reversed by the Court of
Appeals for the District of Columbia, which held that Rambus’s attempt to increase prices fol-
lowing adoption of a standard did not amount to anti-competitive conduct unless such conduct
also resulted in adoption of the standard, which was not shown. This decision has been criticized,
both on the basis of its antitrust analysis and as a matter of public policy. Citations to literature
discussing the case are collected in Contreras (2011).

2. Essentiality and disclosure

A key element of all SSO policies that require either disclosure or licensing of SEPs is
the manner in which patent claims are classified as “essential” to the relevant standard.
The determination whether a particular patent is indeed essential to a particular
standard is typically left to the patent holder. Bekkers and Updegrove (2012). Due to
resource constraints, SSOs rarely review or validate their participants’ determinations
of essentiality.30
Compounding the problem of unverified self-reporting is the range of definitions of
essentiality used by different SSOs. Bekkers and Updegrove (2012) identify 13 different
features of SSO essentiality definitions which varied considerably over the ten SSOs they
studied.31 One major divide among SSOs is whether they define an “essential” patent
claim as covering a technology that must, as a technical mater, be included in a product
implementing the standard (technical essentiality) or whether that patented technology,
though not strictly required as a technical matter, is the only commercially feasible way
that the standard can be implemented (i.e., considering cost, efficiency, reliability, manu-
facturability, etc.) (commercial essentiality) (ABA, 2007; Bekkers and Updegrove, 2012;
Contreras, 2013b, 2017; Kesan and Hayes, 2014).
Given the incentives that firms have to obtain and disclose patents that are essential
to standards, penalties for failing to disclose potentially essential patents (see section
V.A.1),32 and the lack of verification of claims of essentiality by SSOs, some authors sug-
gest that the many patents disclosed as essential to standards are not actually necessary for
the implementation of those standards. This phenomenon is known as “over-disclosure”
or “over-declaration”. Goodman and Myers (2005) and Fairfield (2007, 2010) find
that only 27, 28 and 50 percent of patent families declared “essential” to ETSI’s GSM,
WCDMA, and LTE standards, respectively, are actually essential to implementation
of those standards.33 Cyber Creative (2013) independently evaluated a sample of 2129
SEPs declared essential to the LTE standard (representing 36 percent of the 5919 total
declared SEPs). It found that 56 percent of the sampled SEPs were “truly” essential to
the standard, while 29 percent were partially essential and 15 percent were not essential
at all. The investigators then calculated the ratio of essential to non-essential declared
SEPs for each major contributor to the standard, finding that several firms—including
Apple, Alcatel-Lucent, Freescale, and Nortel—had essentiality ratios of 40 percent or
30
This situation is different than that observed in patent pools, in which significant upfront
costs are incurred to verify the “essentiality” of all patents proposed to be included in the pool in
order to comply with relevant antitrust requirements (Contreras, 2018; Lundqvist, 2014). Estimated
costs of such essentiality analyses are presented in Régibeau et al. (2016) (see note 23, above).
31
Some of these variations involved the degree to which copyrights and other intellectual
property, in addition to patents, could be considered essential to a particular standard.
32
Cases such as Dell, Rambus and Qualcomm (see section V.A.1 above) gave a strong message
to industry that failing to disclose SEPs could be viewed by antitrust enforcement authorities as
deceptive and anti-competitive conduct. There appears to be no similar legal disincentive to over-
disclosing patents to an SSO, though it is theoretically possible that intentional over-disclosure
could support a claim of fraud or deception, especially if the patent holder sought to charge
royalties on patents that were not essential to the standard (Contreras, 2013b).
33
The methodology of the first such study has been critiqued in an unpublished working
paper by Martin and DeMeyer (2006).

less; while ZTE, CATT, NTT Docomo, and InnovativeSonic had ratios above 80 percent.
The average essentiality ratio observed was 56.6 percent, or 53.8 percent when only issued
patents were considered.
Interviews conducted by Blind et al. (2011) also point to widespread over-disclosure
of patents at SSOs. As noted above, Régibeau et al. (2016) interviews identified over-
disclosure as a significant area of concern for stakeholders in standardization.
Over-disclosure is not an issue, of course, when firms make so-called “blanket” licens-
ing statements to the effect that they will license all of their patents that cover a particular
standard on specified terms (royalty-free or FRAND). According to Bekkers et al. (2014,
p. 37), of 14 SDOs studied, nine allow blanket disclosures and five prohibit them. In a
study of 13 SDOs, Bekkers et al. (2017) found that disclosure of weaker patents occurs
more frequently in SDOs, such as ETSI, that prohibit blanket licensing statements. At
IETF, blanket statements are permitted only if the SEP holder agrees to license its SEPs
on royalty-free terms. Baron and Pohlmann (2018, Table 1) catalog the number of blanket
disclosures made at each of 18 SDOs studied. Bekkers et al. (2014) conclude that “the
larger a firm’s patent portfolio, the less likely it is to make blanket disclosures, all other
things being equal” (p. 45).
Challenges to claims of essentiality can be raised in litigation if and when those patents
are asserted by their owners. One such challenge occurred in Innovatio (2013), in which
the court found that 168 disputed patent claims were “essential” to the 802.11 standard,
reasoning, with respect to many of such claims, there was “no commercially feasible alter-
native to using the . . . claims to implement the standard”.34 Eltzroth (2008) also discusses
claims brought before the European Commission challenging a declaration submitted by
Sun Microsystems to ETSI. The challenger asserted, among other things, that the patent
declared was not essential.
To begin to address over-disclosure, Régibeau et al. (2016) have proposed a number of
possible approaches, including increasing the cost of declaring SEPs, limiting the number
of patents that any SSO participant may declare as essential, and instituting random
essentiality testing of patents declared as essential.35 Contreras (2013b) has proposed
an aggregate royalty system in which other holders of SEPs may challenge each other’s
essentiality determinations, with penalties assessed against over-declaration. The Japan
Patent Office (JPO) has recently initiated a “Hantei” program to issue advisory opinions
regarding the essentiality of Japanese patents that have been declared essential to certain
industry standards (JPO, 2018). The results of this program have not yet been subject to
study.
B. FRAND Licensing Disputes
As noted in section VI.B above, many SSOs require participants to grant patent licenses
to standards implementers on terms that are FRAND. But despite their prevalence, few,
34
IEEE has a “commercial essentiality” requirement (IEEE, 2015; Contreras and Layne-Farrar,
2017).
35
It is not clear whether such testing would be conducted by the SSO, the SEP holder or a
governmental regulator.

if any, SSOs define FRAND with sufficient detail to fully specify the scope of the obliga-
tion.36 Baron et al. (2019, 7.1.2) surveyed SSO stakeholders regarding the perceptions of
“patent-centric” and “product-centric” firms involved in standardization:
Both types of firms on average were more likely to agree than disagree with the statements
that “SDO policies requiring FRAND commitments have proven generally successful” and
“FRAND ensures an adequate balance between implementers and IPR holders”. Patent-Centric
respondents generally evidenced higher levels of satisfaction (average scores of 4.64 and 4.43 [on
a 5-point scale], respectively) than Product-Centric respondents (3.59 and 3.50, respectively).
Conversely, Product-Centric respondents gave an average score of 3.64 to the statement “The
terms ‘fair’ and ‘reasonable’ are too vague and open to too many conflicting interpretations”,
compared to an average score of 2.71 from Patent-Centric respondents.
Not surprisingly, disagreements have arisen regarding the scope and details of SSO
participants’ FRAND licensing obligations. A survey of worldwide FRAND-related
litigation can be found in Pentheroudakis and Baron (2017).
1. Royalty rates

Much FRAND litigation centers on whether an SEP holder has breached its FRAND
licensing commitment by charging (or demanding) royalties that are too high to be con-
sidered “reasonable” under the relevant SSO’s policy. Such claims may be raised either in
an affirmative breach of contract action by an implementer seeking a license under those
SEPs (Microsoft, 2013; Apple, 2012) or as an affirmative defense to an infringement action
brought by the SEP holder (Ericsson, 2014).
The analysis of “reasonable” royalty rates for purposes of FRAND commitments has
been substantially informed by a long line of cases and extensive analysis of reasonable
royalty damages in patent infringement cases. This analysis, at least in the United States,
generally takes as its starting point the 15-factor analytical framework introduced in
Georgia-Pacific (1970), as modified to accommodate perceived unique characteristics
of FRAND commitments. Microsoft (2013), Innovatio (2013), Ericsson (2014). Courts
outside the United States, most notably the UK High Court (Patents) in Unwired Planet
(2017), have not adopted the Georgia-Pacific framework for the calculation of FRAND
royalties, and authors adopting a comparativist perspective have questioned whether the
“reasonable royalty” damages calculation framework is even apt for the calculation of
FRAND royalties committed under a contractual regime rather than a compensatory
damages regime (Contreras et al., 2019). A detailed analysis of the cases addressing
FRAND royalties and their royalty calculation methodologies can be found in Siebrasse
and Cotter (2017b), Pentheroudakis and Baron (2017), Contreras and Gilbert (2015),
Cotter (2014), Layne-Farrar and Wong (2014), NRC (2013), and Sidak (2013).
A significant body of theoretical work also exists regarding FRAND royalty rates and
the pricing of SEP licenses. Notable contributions to this literature include Swanson and
Baumol (2005), Geradin, Layne-Farrar, and Padilla (2007), Farrell et al. (2007), Lemley
and Shapiro (2007), Layne-Farrar, Padilla, and Schmalensee (2007), Elhauge (2008),
Lichtman (2010), Sidak (2013, 2016a), Cotter (2014), Contreras and Gilbert (2015),
36
For implications of this lack of specificity on the contract law analysis of FRAND commit-
ments, see Lemley (2002) and Contreras (2015a).

Régibeau et al. (2016), Bartlett and Contreras (2017), and Siebrasse and Cotter (2017a).
A summary of many of these theoretical debates can be found in Siebrasse (2019).
The principal empirical data relating to FRAND royalty rates has emerged as a result
of evidence presented in litigated cases. The court in Microsoft (2013), in particular,
conducted a detailed analysis of numerous “comparable” licensing arrangements to
determine FRAND royalty rates and ranges for two important standards (ITU’s H.264
audio-video encoding standard and IEEE’s 802.11 Wi-Fi standard). These comparable
licenses included both bilateral license agreements and licenses granted by patent pools.
Likewise, the court in Innovatio (2013) considered various comparable license agreements
to benchmark royalty rates in making its FRAND determination. Less data on industry
royalty rates emerges from cases in which FRAND rates are determined by a jury rather
than the bench,37 as were the determinations in Ericsson (2013) and Realtek (2013).
The use of comparable license agreements to determine patent royalty rates in litigation
has been the subject of some controversy, with some commentators supporting the use
of comparable licenses as the best available evidence of the royalty rates that the parties
would have agreed in a hypothetical negotiation (Sidak, 2016; Geradin and Layne-Farrar,
2011). Others, however, express concern that license agreements between different parties
are unreliable, non-transparent, and seldom comparable enough to be useful in determin-
ing royalties that would have been agreed by the parties in litigation (Masur, 2015).
In addition to litigation data, a few studies of FRAND royalty data have been
conducted in standards-heavy industries. Stasik (2010) compiles royalty rates for patents
essential to the ETSI 4G LTE standard, Armstrong, Mueller, and Syrett (2014) report
known and estimated patent royalty rates for the various components of a smart
phone; and Mallinson (2015), refuting Armstrong, Mueller, and Syrett, compiles data
regarding publicly announced licensing rates for various standards included in mobile
wireless devices. Blind et al. (2011), rather than relying on published royalty rates, conduct
structured interviews to develop profiles of typical licensing rates for different telecom-
munications standards. These interviews confirm the variability of licensing royalty rates
from firm to firm and technology to technology, as well as factors, such as cross-licensing,
that tend to affect rates.
2. Royalty base and apportionment

Closely related to the question of royalty rate in FRAND discussions is that of royalty
“base”: the amount to which the royalty rate is applied. Royalty base is a critical variable in
patent licensing discussions, as it directly impacts the revenue received by the licensor, and
it has recently become the subject of U.S. litigation over the proper measure of “reasonable
royalty” damages in patent infringement suits. The crux of the problem is that hundreds or
even thousands of patents often cover a single technology product. Though each such patent
typically claims only a small subset of the overall product’s functionality, in commercial
license agreements royalties are often measured on the basis of the end product’s sale price.
Under longstanding U.S. precedent, “reasonable royalty” patent damages must be
“based on the incremental value that the patented invention adds to the end product”
37
Contreras (2015b) discusses considerations around judicial versus jury determinations of
FRAND royalties.

(Ericsson, 2014). This incremental value approach requires the court to determine the
portion of the overall product value that is contributed by the patented feature, in view of
all the other features of the product (id.). This analysis is often referred to as “apportion-
ment” (Sidak 2016; Siebrasse and Cotter, 2017b). As an aid to juries seeking to apportion
damages in this manner, courts may instruct them to apply the determined royalty to the
“smallest saleable patent practicing unit” [SSPPU] of a product”: Ericsson (2014).38 In
contrast, royalties should be payable on the entire market value of the end product (the
“entire market value rule”) when the value of the entire product is attributable to the
patented feature (id.).
There has been considerable debate over the propriety and application of the SSPPU
principle to cases involving standardized products. In its 2015 policy amendments, IEEE
stated that the determination of reasonable royalty rates should include consideration of
the contribution made by the patented feature to the value of the SSPPU (IEEE, 2015). In
its approval of the amendments, the Department of Justice noted that such considerations
support the appropriate valuation of technologies subject to a FRAND commitment
(DOJ, 2015). Nevertheless, numerous commentators have criticized the application of
the SSPPU principle, particularly in cases involving standardized products. Petit (2016),
Teece and Sherry (2016), and Sidak (2014) argue that the SSPPU rule runs contrary to
commercial licensing practice and needlessly increases transaction costs and uncertainty
in licensing transactions. Régibeau et al. (2016) argue that different royalty bases may
be appropriate in different market situations, making governmental rules relating to the
vertical level of licensing “misguided”.
From an empirical standpoint, Putnam and Williams (2016) sampled patents declared
by Ericsson as essential to ETSI 2G, 3G, and 4G standards, and found, among other
things, that the majority of the sampled patent claims read on more than one system or
component, calling into question the use of the SSPPU methodology with respect to such
patents.
3. Non-discrimination
While much of the debate concerning FRAND licensing has centered on the “reasonable-
ness” of royalty rates, increasing attention has focused on the “non-discrimination” prong
of the FRAND commitment. The issue has arisen largely in connection with some SEP
holders’ desire to deny licenses to “upstream” component vendors in order to collect
royalties from more lucrative “downstream” vendors.
In general, a patent holder may choose to license its patent, and charge a royalty, to any
producer in the supply chain for a product, or to the product’s end users. However, due to
the doctrine of patent exhaustion,39 the patent holder can only collect a royalty once per
patented product. As soon as the patented technology is sold by the patent holder or its
authorized licensee, the patent is “exhausted” and no further royalties can be collected.
For example, the holder of a patent covering an aspect of a wireless communications
standard could license either the manufacturer of the wireless chipset that embodies
38
The SSPPU doctrine originated in Cornell Univ. v. Hewlett-Packard Co., 698 F.Supp.2d 279
(N.D.N.Y. 2009).
39
Quanta Computer, Inc. v. LG Elecs., Inc., 553 U.S. 617 (2008).

that standard, the manufacturer of a smart phone that incorporates that chipset, or the
consumer who uses the smart phone’s wireless communications capability. Thus, the
patent holder’s hypothetical 2 percent royalty on the sale price of the $600 smart phone
would yield the patent holder $12, while the same 2 percent royalty applied to the $50
wireless chipset inside the smart phone would yield the patent holder only $1. Yet if the
patent holder licenses the chipset vendor, the patent is “exhausted” and no royalty can be
charged to the smart phone manufacturer that incorporates the chipset into its phone, or
the consumer who uses the phone.40
This is the classic supply chain issue faced by patent holders. To address it and
maximize royalty revenue, patent holders often seek to license the entity that is furthest
“downstream,” or selling products at the highest price, while declining to grant licenses
to suppliers of low-priced intermediate components. While this practice is within a
patent holder’s rights in the ordinary course of business, it is not clear that the practice
is permitted when patents are subject to a FRAND commitment. That is, if a FRAND
commitment implies that a patent holder must grant or at least offer licenses to all that
request them, then refusal to grant a license to a low-value component vendor could con-
stitute impermissible “discrimination” (Carlton and Shampine, 2013; Contreras, 2015b).
The court in Microsoft (2013) adopted a similar view, and IEEE recently clarified that its
patent policy requires the grant of licenses to all applicants (IEEE, 2015).
Patent holders that prefer to select which implementers to license have taken the position
that, absent express SSO policy language to the contrary, they are permitted to choose
their licensees under their FRAND commitments (Ericsson, 2013). They argue, among
other things, that the purpose of the FRAND “non-discrimination” requirement is to
ensure that all licenses that are granted have comparable, if not identical, terms. While the
question whether FRAND commitments generally imply the licensing of all applicants
remains open, most commentators agree that some variation among FRAND licensing
terms will be tolerated by the “non-discrimination” requirement, and that all FRAND
licenses need not be identical (Layne-Farrar, 2010; Gilbert, 2011; Carlton and Shampine,
2013; Contreras, 2015b). Gilbert (2011) proposes that greater transparency of FRAND
licensing terms may ensure greater compliance with the non-discrimination requirement—
a requirement that has, in some cases, been alleged to have been breached (Qualcomm,
2008). These issues are addressed at greater length in Contreras and Layne-Farrar (2018).
4. Injunctions
Much of the global controversy concerning FRAND licensing relates to an SEP holder’s
ability to seek injunctive relief to prevent the ongoing infringement of an SEP after a
FRAND commitment has been made. The question arises because an SEP holder that
has made a FRAND commitment has agreed, in theory, to grant licenses under its SEPs
to implementers of relevant standards. Seeking an injunction to prevent such an imple-
menter from practicing the SEP would therefore be contrary to the SEP holder’s FRAND
obligation. An argument can thus be made that an SEP holder subject to a FRAND
40
Of course, from an economic standpoint, the particular point in the supply chain at which a
royalty is charged is irrelevant, as the royalty rate can be adjusted to yield an equivalent charge no
matter what the royalty base is (Teece, Grindley, and Sherry, 2013).

commitment should not be permitted, as a matter of remedies law, to seek an injunction

against an implementer of a standard, and that if it does seek an injunction, doing so may
violate the FRAND commitment and antitrust/competition laws.
A sizeable body of theoretical literature and advocacy has been produced to address this
question and is summarized in NRC (2013), Cotter (2014), Sidak (2015a), Sidak (2018),
and Larouche and Zingales (2017). Some of the principal legal and theoretical arguments
against permitting such injunctions are outlined in Lemley and Shapiro (2007), Lemley
(2007), Miller (2007), Shapiro (2010), Michel (2011), and Chien and Lemley (2012); while
arguments against the limitation of such injunctions can be found in Sidak (2008, 2015a),
Epstein, Kieff, and Spulber (2012), Camesasca (2013), and Layne-Farrar (2014).
The issuance of injunctions on FRAND-committed patents has now been considered
by the U.S. federal courts (Apple, 2014), the International Trade Commission (Samsung
ITC, 2013), the FTC (Motorola and Google, 2013), the European Commission (Samsung
EC, 2014), the European Court of Justice (Huawei, 2015), and the courts and enforcement
agencies of China, Japan, and Korea (see Contreras et al., 2019). While there is still some
disagreement over the precise standards for review, a consensus appears to be forming
among courts and agencies that making a FRAND commitment with respect to an SEP
is generally inconsistent with seeking an injunction to prevent ongoing infringement of
that patent by an implementer of the standard, and that such injunctions should, as a
general matter, be issued only when the infringer is unwilling to negotiate for a FRAND
license, refuses to pay a FRAND royalty, or is beyond the jurisdictional reach of the
relevant court or agency.
Despite the surge of recent empirical work investigating the award of injunctive relief
in patent cases generally, relatively little of this work has focused on SEPs. This may be
because, as some commentators have pointed out, few if any injunctions have been issued
with respect to FRAND-encumbered SEPs (Layne-Farrar, 2014b).
Most of the debate regarding the appropriateness of injunctive relief in the face of
FRAND commitments has focused on the interpretation and application of legal rules
and procedure, in the absence of express guidance from SSO policies. A significant
amount of commentary has emerged in Europe in the wake of Huawei (2015), but the
case and its aftermath are still too new to draw any firm conclusions. For a discussion
of the European case law following Huawei, see Pentheroudakis and Baron (2017) and
Larouche and Zingales (2017).
On the policy side, IEEE was the first major SSO to amend its patent policy (see section
IV.C above) to prohibit its participants from seeking injunctive relief against any imple-
menter of an IEEE standard unless the implementer refused to pay the FRAND royalty
rate determined by a court (IEEE, 2015). Such a rule will now, presumably, guide courts
and agencies regarding requests for injunctive relief made by patent holders with respect
to IEEE standards. It remains to be seen how many other SSOs, if any, follow this lead.
VI. OTHER SSO POLICY FEATURES AND PROPOSALS
In addition to patent disclosure and FRAND licensing (see section V), several other issues
emerging from SSO patent policies have recently received attention from commentators
and agencies.

A. Ex Ante Licensing Disclosure
Many SSO disclosure policies require that patent holders disclose to the SSO patents
that are essential to the implementation of the SSO’s standards. Though such policies
often permit patent holders to grant licenses on FRAND terms, very few require the
patent holder to disclose in advance (ex ante) the royalty rates that it intends to charge.
Rather, these and other key contractual terms are left to bilateral negotiations between the
patent holder and individual implementers of the standard, and are usually protected by
confidentiality restrictions that prevent this information from being shared with others.
In the early 2000s, some commentators began to caution that royalty rates left
unspecified during standardization could lead to hold-up situations characterized by
excessive royalty demands following industry lock-in to a particular standard (Ohana,
Hansen, and Shah, 2003; Skitol, 2005; Updegrove, 2006; Lemley, 2007). As a result,
proposals were made at several large SSOs—including IEEE and ETSI, as well as the
smaller VITA—to require the disclosure of royalty information prior to approval of a
standard (Contreras, 2013a). Due to internal opposition and the specter of liability for
encouraging anti-competitive price collusion (Sidak, 2009), IEEE and ETSI amended
their patent policies to permit, but not require, the disclosure of royalty and other
licensing information (see section IV.C). VITA, in contrast, proceeded with its policy
amendments requiring mandatory ex ante disclosure of maximum patent licensing rates.
Both VITA and IEEE requested, and received, business review letters from the DOJ
approving their proposed policy amendments (DOJ, 2006, 2007). There was at least one
coordinated attempt by a group of European network operators to reveal maximum
aggregate royalty rates for certain wireless standards through the NGMN consortium
(Contreras, 2013a, pp. 178–79).
As noted in section IV.C above, at least one large VITA member withdrew from the
organization as a result of the adoption of the ex ante policy. Some commentators
predicted that the adoption of ex ante disclosure policies would have the effect of driv-
ing other members away from SSOs adopting such policies, in addition to reducing the
efficiency and quality of the standards development process at such SSOs (Skitol, 2005;
Tapia, 2010; Herman, 2010). Contreras (2013a) empirically tested these predictions at
VITA, IEEE, and IETF, a large SSO that permits voluntary ex ante disclosures, finding
no evidence of the predicted process deterioration. In interviews with VITA participants,
there was moderate to strong support for the ex ante policy. Blind et al. (2011) and
Régibeau et al. (2016) both interviewed SSO participants regarding ex ante disclosure
policies, receiving a range of responses both positive and negative.
Potential antitrust liability has often been raised when ex ante disclosure policies are
discussed. Specifically, requiring a patent holder to disclose its licensing terms ex ante
could enable potential licensees (implementers of a standard) to collectively exert anti-
competitive pressure on the patent holder to reduce its royalties toward zero, resulting in
the devaluation of patents covering the standard.41 Sidak (2008). This type of improper
buyer cartel is avoided when patent holders are permitted to negotiate license terms with
41
Contreras (2013a) finds no evidence that VITA’s mandatory ex ante disclosure policy caused
patent holders to reduce their requested royalty rates.

implementers on a bilateral basis, constrained only by FRAND guidelines. Given these

arguments, Skitol (2005), Lemley and Shapiro (2007), and Contreras (2013b) propose that
antitrust authorities should more clearly authorize limited degrees of collective royalty
negotiation in the context of SSOs.
To date, few if any additional SSOs have adopted ex ante disclosure policies, likely
due to a combination of inertia, antitrust concerns, and internal opposition from patent
holders (Contreras, 2013a). Given these market realities, Lerner and Tirole (2015) propose
that SSOs be mandated by law to require patent holders to make ex ante maximum royalty
commitments (what they call structured price commitments).42
B. Aggregate Royalty Caps
As discussed in section III.G.2, the principal risk associated with royalty stacking is that
individual patent holders, each acting to maximize its own profit, will in the aggregate
charge a cumulative royalty that is above efficient levels. To address this risk, proposals
have been made to cap the aggregate royalties chargeable with respect to a given standard
or product. Some firms, in order to promote the adoption of particular standards, have
voluntarily committed to such caps. Bekkers and West (2009) and Contreras (2015c)
describe such commitments made by firms in the wireless telecommunications sector,
and Layne-Farrar (2014c) offers a case study of commitments relating to ETSI’s LTE
standard. Tapia (2010) and Bekkers and West (2009) also discuss efforts made to impose
aggregate royalty caps at ETSI for 3G and 4G telecommunications standards, but these
efforts were ultimately unsuccessful as major patent holders were unwilling to participate.
In addition to these industry efforts, commentators have proposed more coordinated
approaches to containing royalties at the SSO level. Lemley (2007) proposes a “step-
down” royalty structure in which each successive patent holder seeking to charge royalties
on a particular standard would be entitled to seek an increasingly diminished royalty.
Contreras (2013b) proposes collective agreements among SSO members on aggregate
royalty caps for particular standards, with proceeds divided among SEP holders and
meaningful penalties for over-disclosure. Régibeau et al. (2016) and Lerner and Tirole
(2015) likewise make proposals regarding agreement on aggregate royalty caps for particu-
lar standards. In contrast, Herman (2010) and Blind et al. (2011) list arguments that have
been made against aggregate royalty caps, including their potentially disproportionate
impact on patent holders. Régibeau et al. (2016, Sec. 5.1.2) summarize various royalty cap
approaches as well as stakeholder reactions to these proposals.
C. Royalty-free Licensing
Much recent litigation concerning standards and patents revolves around appropriate
levels for, and methods of calculating, FRAND royalty rates. These challenges, as well as
42
Ganglmair, Froeb, and Werden (2012) propose that optimal hold-up avoidance and innova-
tion could be achieved if patent holders and implementers entered into formal “option-to-license”
contracts before the implementer makes investments in the standardized technology. While this
solution is theoretically attractive, significant transaction costs are involved in licensing negotia-
tions, making the widespread use of such pre-implementation contracts unlikely.

the risk of patent hold-up by SSO participants, are alleviated if patent holders covenant
not to assert their standards-essential patents, or commit to license them on a royalty-free
basis. Such royalty-free licensing requirements might be commercially advantageous
in some settings, such as when standards developers value widespread interoperability
more highly than patent-based revenue generation (Blind et al., 2011; Updegrove, 2012;
Contreras, 2016a), have complementary assets that can benefit from adoption of the
standard (Teece and Sherry, 2003), or wish to promote the adoption of a new technology
platform or infrastructure (Contreras, 2015c).
The obvious trade-off of royalty-free requirements is that patent holders seeking a
financial return on their patent portfolios may not wish to participate in such SSOs,
thereby depriving such SSOs of skilled developers and technology inputs, and expos-
ing implementers of SSO standards to infringement of these patents without even the
cold comfort of a FRAND licensing commitment (Herman, 2010; Bekkers and West,
2009).43 Moreover, in some technology areas, it may not be feasible to develop techni-
cally adequate standards without the inclusion of at least some royalty-bearing patented
technology. In this vein, Choi and Jang (2014) discuss the challenges of developing a
royalty-free codec. Chiao, Lerner, and Tirole (2007) observe that SSOs with royalty-free
licensing requirements tend to have fewer patent disclosure requirements, as the effort
of identifying and disclosing patents may be considered less critical under a royalty-free
regime. Bekkers et al. (2017) reason that the decision regarding royalty-free licensing is
made during standards development, noting that “royalty-free licensing commitments
occur when a patent-holder faces ex ante competition from a non-infringing alternative
for inclusion the standard, and the benefits of having its own technology included in the
standard outweigh the costs of foregone royalties” (p. 2).
Greenbaum (2016) raises questions regarding the availability of legal remedies when
royalties are not charged under a license agreement.
Despite these potential drawbacks, royalty-free SSO licensing policies are not uncom-
mon. Biddle, White, and Woods (2010) find that of 251 standards embodied in a typical
laptop computer, 22 percent were available on a royalty-free basis. Of 36 SSOs coded,
Baron and Spulber (2018) identify five that require royalty-free licensing or patent
non-assertion, at least at the working group level. They also find that four SSOs that
previously permitted royalty-bearing FRAND licensing have moved to royalty-free or
patent non-assertion policies.
Significant SSOs that require non-assertion or royalty-free licensing of SEPs include
W3C, the Bluetooth Special Interest Group, the HDMI forum, and the USB Forum.
Other groups, such as software standards developer OASIS, permit technical committees
to determine, upon formation, whether they will require FRAND or royalty-free licensing
commitments from their participants. Bekkers and Updegrove (2012) find that of 83
active OASIS committees, all had selected a royalty-free licensing approach, and none had
selected a royalty-bearing FRAND licensing approach.
43
Rysman and Simcoe (2011) propose a model in which SEPs would be licensed at FRAND
rates for an initial period, after which licenses would become royalty-free. This Non-Assertion
After Specified Time pricing model could enable innovators to recoup development costs, while
thereafter promoting widespread adoption of a standard and eliminating disputes regarding
FRAND royalty rates.

The above discussion relates to de jure royalty-free policies, which are imposed by SSOs
on their participants. In addition to de jure royalty-free standards, it is also possible that
standards developed at SSOs permitting royalty-bearing FRAND licensing may, as a
practical matter, be free from royalty demands. This could be because a standard is simply
not covered by patents held by SSO participants, thus avoiding the need for any licenses
at all, or because the holders of patents covering the standard have elected not to assert
those patents or seek licenses from implementers. The second scenario arises frequently in
the case of IETF, which requires neither FRAND nor royalty-free licensing, but expressly
prefers standards that are available on a royalty-free basis (Contreras, 2016a). Contreras
(2013a) finds that between 2007 and 2010, approximately 59 percent of all patent disclo-
sures made at IETF were accompanied by voluntary commitments not to assert patents
or to license them on a royalty-free basis. Bekkers et al. (2017) also find that voluntary
royalty-free commitments are more likely at IETF than at other SDOs studied.
Another instance of royalty-free standards usage is associated with the so-called “sleep-
ing dog” phenomenon: some patent holders simply do not wish to expend the substantial
time and resources necessary to pursue a patent licensing program (Contreras, 2013a). It is
currently not known how many SEPs declared at different SSOs are held by sleeping dogs,
but they, together with voluntary royalty-free commitments made at SSO such as IETF,
may result in a meaningful body of “de facto” royalty-free standards in the marketplace.
D. Alternative Dispute Resolution
Given the increase in litigation concerning standardization and SSO policies, several
commentators have suggested the use of alternate dispute resolution (ADR) mechanisms
to streamline the resolution of disputes relating to SEPs (Kühn, Scott, Morton, and
Shelanski, 2013). The FTC and European Commission have also recognized arbitration
as a suitable method for resolving SEP-related disputes (Motorola and Google, 2013;
Samsung EC, 2014).
As a matter of implementation, Lemley and Shapiro (2013) propose that disputes
regarding FRAND royalty rates be settled by binding “final offer” or “baseball” arbitra-
tion. In such proceedings, each party provides the arbitrator with a sealed “final offer,”
of which the arbitrator must choose only one, without modification. This approach is
supported by Régibeau et al. (2016, p.80), who offer the alternative of “night baseball”, in
which the arbitrators are not informed of the parties’ offers, but must make an independ-
ent assessment of the royalty level, after which the royalty is set at the party’s offer that
is closest to the arbitrator’s assessment. Larouche, Padilla, and Taffet (2014) challenge
baseball arbitration as unnecessary and likely to undermine the standardization process.
Contreras and Newman (2014) develop a framework for conducting arbitration concern-
ing standards and SEPs. Among other issues, they raise concerns regarding the general
confidentiality of arbitral awards.
A few SSOs have adopted ADR mechanisms in their rules and policies. The DVB Forum
has had such a policy in place since 1995 (Eltzroth, 2008). Contreras and Newman (2014)
identify and describe four longstanding SSO ADR policies. Most recently, IEEE amended
its patent policy to permit, but not require, arbitration of SEP-related disputes (IEEE, 2015).
In addition to SSOs, several international arbitration bodies have begun to modify their
practices and policies to accommodate proceedings concerning SEPs and standardization.

The most ambitious of these has been the World Intellectual Property Organization, which
has developed a bespoke procedure specifically addressed to SEP disputes (Greenbaum,
2015). Contreras (2019b) has proposed the establishment of a global, non-governmental
rate-setting tribunal that would determine top-down aggregate royalty rates for different
standards as well as the allocation of royalties paid among different SEP holders.
E. Transfer of Commitments
When an SEP holder makes a commitment to license SEPs to manufacturers of standard-

ized products, it is not always clear whether that commitment applies only to the SEP
holder making the commitment, or whether it binds subsequent holders of the SEP (see,
generally, Block, 2018). As observed by Régibeau et al. (2016), “a FRAND commitment
is not worth much if an SEP can be sold without transferring the commitment” (p. 71).
Yet the legal basis for the transfer of such commitments is not entirely clear.
The FTC has taken the position that a party acquiring an SEP with the knowledge of
a prior FRAND commitment must abide by that commitment, and that a failure to do
so constitutes an unfair method of competition in violation of Section 5 of the FTC Act
(N-Data, 2008). This matter was settled, however, before any judicial ruling on the ques-
tion. The issue arose again in 2011, when bankrupt Nortel Networks, a major contributor
to several SSOs, proposed the sale of its remaining assets, including approximately 4000
patents, on a “free and clear” basis (Nortel, 2011). Several product vendors, together with
IEEE, argued that Nortel’s “free and clear” sale could invalidate patent licensing commit-
ments that Nortel had previously made to SSOs. Ultimately, the purchaser of the patents,
a consortium including several large product vendors, agreed to abide by Nortel’s prior
licensing commitments and the issue was not adjudicated.
Because courts have not yet definitively ruled on the binding nature of SEP licensing
commitments on subsequent SEP holders, the state of the law is unsettled in this regard.
As a result, an increasing number of SSOs have required in their internal policies that
participants that transfer SEPs as to which licensing commitments have been made must
ensure that those commitments are binding on successive owners of the SEPs. Bekkers
and Updegrove (2012) catalog SSOs that impose such transfer requirements, and NRC
(2013) discusses the variety of SSO policy provisions that can be employed in this regard.
IEEE’s 2015 policy amendments are an example of such provisions. Most commentators
who have considered the matter support the implementation of voluntary policy mecha-
nisms to ensure the binding nature of SEP licensing commitments following a transfer
of the SEPs (Kühn et al., 2013; NRC, 2013; Kesan and Hayes, 2014; Contreras, 2015c;
Régibeau et al., 2016).
In some cases, SSO participants have transferred SEPs to PAEs for the purpose of
monetization and assertion (this practice is sometimes referred to as “privateering”)
(Lundqvist, 2014); Golden, 2013). JRC (2016) found that approximately 80 percent of
patents asserted by PAEs were obtained from operating companies. In one recent case, a
product manufacturer has alleged that a SEP holder conspired with a number of PAEs in
violation of its FRAND commitments and U.S. antitrust laws to subdivide a portfolio of
SEPs in order to collect excessive licensing fees (Apple, 2016). These issues will bear close
scrutiny as such cases progress.

VII.
NON-PATENT INTELLECTUAL PROPERTY AND
STANDARDS
Though the vast majority of literature concerning standards and intellectual property has
focused on patents, a few significant copyright and trademark issues have recently gained
the attention of scholars, regulators and the public. Only a few scholars have attempted
to address multiple forms of intellectual property protection for standards. De Carvalho
(2015) offers a holistic view of patents, copyrights, and trademarks on standards within
an international competition law framework, while Contreras and McManis (2013) assess
the interplay of patents, copyrights, and trademarks covering materials sustainability
standards.
A. Copyright
Because most technical standards take the form of written documents, it is generally
understood that they are considered copyrightable works of authorship and protected by
copyright law.44 While various individuals and firms make textual and other contributions
to standards, the relevant SSO often claims ownership of the collective work embodied
in a published standard. While several SSOs make their standards freely available to the
public via the Internet, others exercise tight control over their published standards, often
charging for access and prohibiting copying and distribution.
1. Incorporation by reference

The control exercised by some SSOs over standards documents has become controversial
with respect to standards that have been incorporated by governmental agencies into
legislation and regulation. Such incorporation by reference (IBR) frequently occurs in the
case of local building, safety, and electrical codes, which often reference and incorporate
standards developed within SSOs. Mendelson (2015) estimates that there are more than
9000 of these IBR standards within various federal regulatory codes and agency rules.
Bremer (2015) collects data regarding the pricing of various IBR standards.
A circuit split currently exists in the U.S. courts regarding the copyright status of stand-
ards that have been incorporated into law (Practice Management, 1998; Veeck, 2002).
Nevertheless, in 2012 a petition was filed with the Office of the Federal Register seeking
a rule that technical standards referenced in federal regulations be made freely available
via the Internet (Strauss, 2013). In 2012, a public interest group began to reproduce IBR
standards and make them freely available online, leading to copyright infringement suits
by several SDOs (Strauss, 2013; Bremer, 2015; Sheffner, 2019). In an effort to mediate
this dispute, ANSI has established a controlled-access (read-only) online portal for IBR
standards (ANSI, 2014; Bremer, 2015). Other proposals have been made, including host-
ing of IBR standards by federally controlled websites (Mendelson, 2015).
44
Samuelson (2007) and Samuelson and Hashimoto (2019) challenge this assumption argu-
ing, among other things, that standards documents are functional and should not have the benefit
of copyright protection.

2. Software in standards

Some technical standards include software code that must be implemented in products in
order for them to conform to the standard. Other standards include reference or exem-
plary software code that illustrates the means by which the standard may be implemented
in software (Contreras and Updegrove, 2016). Both of these uses of software within
standards raise issues relating to the ownership and licensing of copyrights. Conflicts over
copyrights and software in standards arose as early as the 1990s over Sun Microsystems’s
efforts to standardize the Java programming language at ECMA and JCT-1 (Egyedi,
2001b; Lemley and McGowan, 1999).
These issues are further complicated when such software is intended for use in open
source code products, which often impose their own demanding licensing conditions
(Vetter, 2007; Updegrove, 2009). Lundell et al. (2015) studied the effect of ISO’s use condi-
tions, as well as filed patent disclosures, on the implementation of three ISO standards
(PNG, JPEG 2000, and TIFF/EP) in open source software, and their potential adoption
under national regulations calling for open standards. They found that while PNG (which
was also recognized by W3C) could be considered an “open” standard, JPEG 2000 and
TIFF/EP would present problems regarding open implementation and were not compat-
ible with typical open source code licenses.
Given the increasing importance of open source software to the global technology
infrastructure, further research in this area is needed.
B. Trademarks and Certification Marks
The names and designations of standards that are widely adopted—such as Bluetooth,
Blu-ray and USB—can acquire substantial market value. SSOs often retain ownership
of these trademarks and license their use in connection with products conforming to the
standard. In some cases, a standard or series of standards may be associated with a mark
not originated by the SSO, such as the well-known Wi-Fi® designation for IEEE’s 802.11
series of wireless networking standards, which is owned and licensed by an independent
organization.45
A large area of activity surrounds the testing and certification of products for
conformity to different standards. Barnett (2012) discusses certification markets and
the sometimes imperfect role played by intermediary firms in certifying third-party
products and services. Often, when a product is certified as compliant with a standard,
its manufacturer is permitted to display a designated logo or certification mark on that
product. Well-known examples include the Underwriters Laboratories’ “UL” certification
for electrical products and the Green Building Council’s “LEED” certification for new
buildings. Contreras (2019a) notes discrepancies in the use of certification marks and
trademarks by SSOs when seeking to protect the names of standards that they develop.
The proliferation of certification marks on certain categories of products, especially
“green” or “eco-friendly” products, has been cataloged and critiqued by Chon (2009) and
Contreras and McManis (2013). Fischer and Lyon (2014) compare environmental certifica-
45
The Wi-Fi® trademark is controlled by Wi-Fi Alliance, a non-profit organization formed in
1999 and which is independent of IEEE.

tion labels developed by non-governmental organizations versus industry bodies. However,

little empirical literature exists regarding the impact of product certification and certifica-
tion marks on technology products and markets, an area that is ripe for further research.
VIII. CONCLUSIONS
Important theoretical work relating to standards has been done in the areas of SSO
dynamics, firm behavior, market effects of patents, and royalty pricing. This work has
been supplemented by a significant body of research and empirical data on the acquisition
and disclosure of patents within SSOs, particularly in the ICT sector. Several important
catalogs and analyses of SSO patent policies now exist, together with rich databases of
SSO membership and policy data.
Despite this large body of literature, there are numerous areas at the confluence of intel-
lectual property and standardization that warrant further investigation. These include the
influence and internal organization of consortia and other informal standards groups; the
prevalence and market impact of de jure and de facto royalty-free standards; the effect of
patents on standardization in growing fields outside of ICT, including clean technology,
medical devices, and automotive infrastructure; the interaction of technology standards
with open source software; the impact of product certification and certification marks
on technology products and markets; and the institutional, legal, and policy landscape of
standardization outside of North America and Europe, particularly in China and other
Asian economies.
In addition, more public data is needed regarding patent licensing and royalty rates for
standardized technologies. The data that currently exists is gleaned largely from public
sources such as litigation records, government licenses and public securities filings. This
data, however, represents only the tip of the iceberg. The largest and most meaningful
accumulation of data concerning patent licensing is locked within the files of private
firms, subject to strict confidentiality restrictions, and beyond the reach of researchers,
policy makers, enforcement agencies, and courts. Greater public access to this data has
the potential to lower licensing transaction costs, reduce the number of disputes regarding
FRAND royalty rates, improve the accuracy of judicial damages determinations, inform
agency enforcement decisions, and improve policy making. As such, it is in the interest of
all participants in the standardization ecosystem to contribute to the growing public data
resources in this important area of economic activity.
REFERENCES
Cases
Allied Tube & Conduit Corp. v. Indian Head Inc., 486 U.S. 492 (1988).
Apple, Inc. v. Motorola, Inc., 2012 U.S. Dist. LEXIS 157525 (W.D. Wis., Nov. 2, 2012).
Apple, Inc. v. Motorola, Inc., 757 F.3d 1286 (Fed. Cir. 2014).
Apple Inc. v. Acacia Research Corp., No. 16-CV-7266 (N.D. Cal. Dec. 20, 2016).
In re. Certain Electronic Devices, Including Wireless Communication Devices, Portable Music and Data Processing
Devices, and Tablet Computers, USITC Inv. No. 337-TA-794 (Jul. 5, 2013) (Samsung ITC).

Decision of the European Commission under Article 102 of the Treaty on the Functioning of the European
Union, Case AT.39939 – Samsung – Enforcement of UMTS Standard Essential Patents (Apr. 29, 2014)
(Samsung EC).
In re. Dell Computer Corp., 121 F.T.C. 616 (1996).
Ericsson v. D-Link, Case No. 6:10-CV-473, 2013 U.S. Dist. LEXIS 110585 (E.D. Tex. Aug. 6, 2013).
Ericsson Inc. v. D-Link Sys., 773 F.3d 1201 (Fed. Cir. 2014).
Georgia-Pacific Corp. v. U.S. Plywood Corp., 318 F. Supp. 1116, 1120 (S.D.N.Y. 1970), modified and aff’d, 446
F.2d 295 (2d Cir. 1971), cert. denied, 404 U.S. 870 (1971).
Huawei v. ZTE, European Court of Justice (2015).
Hydrolevel Corp. v. American Society of Mechanical Engineers, 456 U.S. 556 (1982).
In re Innovatio IP Ventures, LLC, Memorandum and Order, 956 F.Supp.2d 925 (N.D.Ill. 2013).
Microsoft Corp. v. Motorola, Inc., Findings of Fact and Conclusions of Law, 2013 U.S. Dist. LEXIS 60233
(W.D. Wash., Apr. 25, 2013).
In re. Motorola Mobility LLC & Google Inc., FTC Docket No. C-4410 (Jul. 23, 2013) (decision and order).
In re Nortel Networks Inc., No. 09-10138(KG), 2011 WL 4831218 at *6 (Bankr. D. Del., Jul. 11, 2011).
Practice Management Information Corp. v. American Medical Ass’n., 121 F.3d 516, 517 (9th Cir. 1997), opinion
amended by 133 F.3d 1140 (9th Cir. 1998).
Qualcomm Inc. v. Broadcom Corp., 548 F.3d 1004 (Fed. Cir. 2008).
Rambus, Inc. v. Infineon Techs AG, 318 F.3d 1081, 1102 (Fed. Cir. 2003), cert. denied, 124 S. Ct. 227.
In re. Rambus, Inc., No. 9302, 2006 WL 2330117 (F.T.C, Aug. 2, 2006), rev’d, 522 F.3d 456 (D.C. Cir. 2008).
Realtek Semiconductor Corp. v. LSI Corp., Order re: Joint Submission Pursuant to Supplemental Case
Management Order, No. 5:12-cv-03451 RMW (N.D. Cal. Sept. 26, 2013).
Rembrandt Wireless Techs., LP v. Samsung Elecs. Co., Case No. 2:13-CV-213-JRG (E.D. Tex., Feb. 16, 2015).
In re. Robert Bosch GmbH (Bosch), F.T.C. File No. 121-0081, at 3 (F.T.C. Nov. 26, 2012) (Statement of the
Federal Trade Commission).
In re. Union Oil Co. of Cal. (Unocal), FTC Docket No. 9305, 2005 FTC LEXIS 116 (Jul. 27, 2005).
Unwired Planet Intl. Ltd. v. Huawei Techs. Co. Ltd. [2017] EWHC 711 (Pat) (Apr. 5, 2017), aff’d [2018] EWCA
Civ. 2344 (Oct. 23, 2018).
Veeck v. Southern Building Code Congress Intl., 293 F.3d 791 (5th Cir. 2002) (en banc), cert. denied, 537 U.S.
1043 (2002).
Regulatory and Agency Materials
European Commission – Directorate-General for Enterprise and Indus. (EC). 2014. Patents and Standards: A
Modern Framework for IPR-Based Standardization.
European Commission Joint Research Centre (JRC). 2016. JRC Science for Policy Report: Patent Assertion Entities
in Europe, available at http://publications.jrc.ec.europa.eu/repository/bitstream/JRC103321/jrc103321%20onli
ne%20version.pdf.
Japan Patent Office (JPO)—Trial and Appeal Department. 2018. Manual of “Hantei” (Advisory Opinion) for
Essentiality Check (Mar. 2018).
U.S. Dept. of Justice (DOJ). 2006. Business Review Letter to VMEbus International Trade Association (VITA)
(Oct. 30, 2006).
U.S. Dept. of Justice (DOJ). 2007. Business Review Letter from Thomas O. Barnett, Assistant Attorney
General, U.S. Dept. of Justice to Michael A. Lindsay, Partner, Dorsey & Whitney LLP 4 (Apr. 30, 2007).
U.S. Dept. of Justice (DOJ). 2015. Business Review Letter from Hon. Renata B. Hesse, Acting Assistant
Attorney Gen., U.S. Dep’t of Justice, to Michael A. Lindsay, Esq., Dorsey & Whitney, L.L.P. (Feb. 2,
2015).
U.S. Dept. of Justice (DOJ) and U.S. Fed. Trade Comm. (FTC). 2000. Guidelines for Collaborations Among
Competitors.
U.S. Dept. of Justice (DOJ) and U.S. Fed. Trade Comm. (FTC). 2007. Antitrust Enforcement and Intellectual
Property Rights: Promoting Innovation and Competition.
U.S. Federal Trade Commission (FTC). 2016. Patent Assertion Entity Activity.
U.S. Federal Trade Commission (FTC). 2011. The Evolving IP Marketplace: Aligning Patent Notice and
Remedies with Competition.
U.S. Federal Trade Commission (FTC). 2012. Third Party United States Federal Trade Commission’s
Statement on The Public Interest, In re. Certain Gaming and Entertainment Consoles, Related Software, and
Components Thereof, ITC Investigation No. 337-TA-752 (Jun. 6, 2012).
U.S. Office of Management and Budget (OMB). 1998. Circular A-119—Federal Participation in the Development
and Use of Voluntary Consensus Standards and in Conformity Assessment Activities.

SSO Materials
American National Standards Institute (ANSI). 2014. Written Testimony of the American National Standards
Institute before the U.S. House of Representatives Committee on the Judiciary, Subcommittee on Courts,
Intellectual Property, and the Internet, Hearing on “The Scope of Copyright Protection,” Jan. 14, 2014.
American National Standards Institute (ANSI). 2015. “ANSI Standards Activities, Domestic programs overview”,
www.ansi.org/standards_activities/domestic_programs/overview.aspx?menuid=3.
American National Standards Institute (ANSI). 2016. ANSI Essential Requirements: Due process requirements
for American National Standards.
ASTM International (ASTM). 2015. “About ASTM”, www.astm.org.
IEEE Standards Assn. (IEEE). 2015. IEEE-SA Standards Board Bylaws, http://standards.ieee.org/develop/polici
es/bylaws/sect6-7.html.
International Organization for Standardisation (ISO). 2015. “About ISO”, www.iso.org/iso/about.htm.
Books and Articles
Albert, Michael B. et al. 1991. “Direct Validation of Citation Counts as Indicators of Industrially Important
Patents,” 20 Research Policy 251–76.
American Bar Association (ABA). Committee on Technical Standardization, Section of Science &
Technology Law. 2007. Jorge L. Contreras, ed., Standards Development Patent Policy Manual. Chicago:
ABA Publishing.
Armstrong, Ann, Joseph J. Mueller, and Timothy D. Syrett. 2014. “The Smartphone Royalty Stack: Surveying
Royalty Demands for the Components Within Modern Smartphones” (Working Paper, May 29), available at
http://papers.ssrn.com/sol3/papers.cfm?abstract_id=2443848.
Bar, Talia and Aija Leiponen. 2014. “Committee Composition and Networking in Standard Setting: The Case
of Wireless Telecommunications,” 23 Journal of Economic Management Strategy 1–23.
Barnett, Jonathan M. 2012. “Intermediaries Revisited: Is Efficient Certification Consistent with Profit
Maximization?,” 37 Journal of Corporation Law 475–528.
Baron, Justus and Daniel F. Spulber. 2018. “The Searle Center Database of Technology Standards and Standard
Setting Organizations,” 27 Journal of Economics & Management Strategy 462–503.
Baron, Justus and Kirti Gupta. 2018. “Unpacking 3GPP Standards,” 27 Journal of Economics & Management
Strategy 433–61.
Baron, Justus and Tim Pohlmann. 2013. “Who Cooperates in Standards Consortia—Rivals or Complementors?,”
9 Journal of Competition Law & Economics 905–29.
Baron, Justus and Tim Pohlmann. 2018. “Mapping Standards to Patents Using Databases of Declared
Standard-Essential Patents and Systems of Technological Classification,” 27 Journal of Economics &
Baron, Justus, Jorge Contreras, Martin Husovec, and Pierre Larouche. 2019. Making the Rules: The Governance
of Standard Development Organizations and their Policies on Intellectual Property Rights. JRC Science for
Policy Report (March).
Baron, Justus, Knut Blind, and Tim Pohlmann. 2013. “Essential Patents and Standard Dynamics” (Working
Paper, Mar. 15), available at www.law.northwestern.edu/research-faculty/searlecenter/innovationeconomics/
documents/Essential_Patents_and_Standard_Dynamics_2013.pdf.
Bartlett, Jason R. and Jorge L. Contreras. 2017. “Rationalizing FRAND Royalties: Can Interpleader Save the
Internet of Things,” 36 Review of Litigation 285–334.
Bekkers, Rudi and Andrew Updegrove. 2012. “A Study of IPR Policies and Practices of a Representative Group
of Standard Setting Organizations Worldwide,” Presented at National Academies of Science Symposium on
Management of IP in Standards-Setting Processes, Session 4, available at http://sites.nationalacademies.org/
xpedio/groups/pgasite/documents/webpage/pga_072197.pdf.
Bekkers, Rudi and Isabelle Liotard. 1999. “The Tense Relation Between Mobile Telecommunications Standards
and IPR,” 3 European Intellectual Property Review 110–26.
Bekkers, Rudi and Jan Smits. 1997. English trans. 1999. Mobile Telecommunications: Standards, Regulation, and
Applications. Deventer, Netherlands: Kluwer BedrijfsInformatie. London: Artech House.
Bekkers, Rudi and Joel West. 2009. “The Limits to IPR Standardization Policies as Evidenced by Strategic
Patenting in UMTS,” 33 Telecommunications Policy 80–97.
Bekkers, Rudi, Bart Verspagen, and Jan Smits. 2002. “Intellectual Property Rights and Standardization: The
Case of GSM,” 26 Telecommunications Policy 171–88.
Bekkers, Rudi, Christian Catalini, Arianna Martinelli, Cesare Righi, and Timothy Simcoe. 2017. “Disclosure
Rules and Declared Essential Patents,” NBER Working Paper No. 23627.

Bekkers, Rudi, Geert Duysters, and Bart Verspagen. 2002. “Intellectual Property Rights, Strategic Technology
Agreements and Market Structure: The Case of GSM,” 31 Research Policy 1141–61.
Bekkers, Rudi, Justus Baron, Arianna Martinelli, Yann Ménière, Önder Nomaler, and Tim Pohlmann. 2014.
“Selected Quantitative Studies of Patents in Standards,” Hitotsubashi Univ. (Tokyo), Inst. of Innovation
Research, PIE/CIS Working Paper 626 (June).
Bekkers, Rudi, René Bongard, and Alessandro Nuvolari. 2011. “An Empirical Study on the Determinants of
Essential Patent Claims in Compatibility Standards,” 40 Research Policy 1001–15.
Berger, Florian, Knut Blind, and Nikolaus Thumm. 2012. “Filing Behavior Regarding Essential Patents in
Industry Standards,” 41 Research Policy 216–25.
Besen, Stanley M. 1990. “The European Telecommunications Standards Institute: A Preliminary Analysis,” 14
Telecommunications Policy 521–30.
Biddle, C. Bradford. 2018. “No Standard for Standards: Understanding the ICT Standards-Development
Ecosystem,” in Contreras, ed., Chapter 2.
Biddle, Brad, Andrew White, and Sean Woods. 2010. “How Many Standards in a Laptop? (And Other Empirical
Questions),” Int’l Telecomm. Union Sec. Telecomm. Standardization Kaleidoscope Acad. Conf. Proc. 123.
Biddle, Brad et al. 2012. “The Expanding Role and Importance of Standards in the Information
and Communications Technology Industry,” 52 Jurimetrics 177–208.
Biddle, C. Bradford, Jorge L. Contreras, Brian J. Love, and Norman V. Siebrasse, eds. 2019. Patent Remedies and
Complex Products: Toward a Global Consensus. New York: Cambridge Univ. Press (forthcoming).
Blind, Knut. 2011. “An Economic Analysis of Standards Competition: The Example of the ISO ODF and
OOXML Standards,” 35 Telecommunications Policy 373–81.
Blind, Knut and Eric Iversen. 2004. “The Interrelationship Between IPR and Standardisation: Patterns and
Policies,” 9th EURAS Workshop on Standardization, May 13–14, Paris. Available at http://eprints.utas.edu.
au/1282/.
Blind, Knut and Nikolaus Thumm. 2004. “Interrelation Between Patenting and Standardization Strategies:
Empirical Evidence and Policy Implications,” 33 Research Policy 1583–98.
Blind, Knut et al. 2011. Study on the Interplay between Standards and Intellectual Property Rights (IPRs), Final
Report, Tender No. ENTR/09/015 available at www.iplytics.com/download/docs/studies/ipr_study_final_re
port_en.pdf.
Block, Marc Sandy. 2018. “Transfers of Standards Essential Patents,” in Contreras, ed., Chapter 14.
Bremer, Emily. 2015. “On the Cost of Private Standards in Public Law,” 63 University of Kansas Law Review
279–344.
Breshnahan, Timothy F. and Pai-Ling Yin. 2007. “Standard Setting in Markets: The Browser War,” in Shane
Greenstein and Victor Stango, eds., Standards and Public Policy 18. New York: Cambridge Univ. Press.
Camesasca, Peter et al. 2013. “Injunctions for Standards-Essential Patents: Justice is not Blind,” 9 Journal of
Competition Law & Economics 285–311.
Cargill, Carl and Sherrie Bolin. 2007. “Standardization: A Failing Paradigm,” in Shane Greenstein and Victor
Stango, eds., Standards and Public Policy 296. New York: Cambridge Univ. Press.
Carlton, Dennis W. and Allan L. Shampine. 2013. “An Economic Interpretation of FRAND,” 9 Journal of
Caviggioli, Federico et al. 2015. “Patenting Strategies and Characteristics of Declared Inventions in the Long
Term Evolution Standard,” 46 R&D Management 664–76.
Chiao, Benjamin, Josh Lerner, and Jean Tirole. 2007. “The Rules of Standard-Setting Organizations: An
Empirical Analysis,” 38 RAND Journal of Economics 905–30.
Chien, Colleen V. 2014. “Holding Up and Holding Out,” 21 Michigan Telecommunications & Technology Law
Journal 1–42.
Chien, Colleen V. and Mark A. Lemley. 2012. “Patent Holdup, the ITC and the Public Interest,” 98 Cornell
Law Review 1–45.
Choi, Kiho and Euee S. Jang. 2014. “Royalty-Free Video Coding Standards in MPEG,” IEEE Signal Processing
Magazine 145–55.
Chon, Margaret. 2009. “Marks of Rectitude,” 77 Fordham Law Review 2311–51.
Contreras, Jorge L. 2011. “Equity, Antitrust and the Reemergence of the Patent Unenforceability Remedy,”
Oct. Antitrust Source 1–8.
Contreras, Jorge L. 2012. “Standards, Patents and the National Smart Grid,” 32 Pace Law Review 641–75.
Contreras, Jorge L. 2013a. “Technical Standards and Ex Ante Disclosure: Results and Analysis of an Empirical
Study,” 53 Jurimetrics 163–211.
Contreras, Jorge L. 2013b. “Fixing FRAND: A Pseudo-Pool Approach to Standards-Based Patent Licensing,”
79 Antitrust Law Journal 47–97.
Contreras, Jorge L. 2013c. “Technical Standards and Bioinformatics,” in Jorge L. Contreras and A. James
Cuticchia, eds., Bioinformatics Law: Legal Issues for Computational Biology in the Post-Genome Era. Chicago:
ABA Publishing.

Contreras, Jorge L. 2014. “Divergent Patterns of Engagement in Internet Standardization: Japan, Korea and
China,” 38 Telecommunications Policy 914–32.
Contreras, Jorge L. 2015a. “A Market Reliance Theory for FRAND Commitments and Other Patent Pledges,”
Utah Law Review 479–558.
Contreras, Jorge L. 2015b. “A Brief History of FRAND: Analyzing Current Debates in Standard Setting and
Antitrust Through a Historical Lens,” 80 Antitrust Law Journal 39–120.
Contreras, Jorge L. 2015c. “Patent Pledges,” 47 Arizona State Law Journal 543–608.
Contreras, Jorge L. 2015d. “Standards, Royalty Stacking, and Collective Action,” CPI Antitrust Chronicle, March.
Contreras, Jorge L. 2016a. “A Tale of Two Layers: Patents, Standardization and the Internet,” 93 Denver Law
Review 833–95.
Contreras, Jorge L. 2016b. “When a Stranger Calls: Standards Outsiders and Unencumbered Patents,” 12
Journal of Competition Law & Economics 507–40.
Contreras, Jorge L. 2016c. “FRAND Market Failure: IPXI’s Standards-Essential Patent License Exchange,” 15
Chicago-Kent Journal of Intellectual Property 419–555.
Contreras, Jorge L. 2017. “Assertion of Standards-Essential Patents by Non-Practicing Entities,” in D. Daniel
Sokol, ed., Patent Assertion Entities and Competition Policy, Ch. 4, 50–71. New York: Cambridge Univ. Press.
Contreras, Jorge L., ed. 2018. The Cambridge Handbook of Technical Standardization Law: Competition,
Antitrust, and Patents. New York: Cambridge Univ. Press.
Contreras, Jorge L. 2018. “Essentiality and Standards-Essential Patents,” in Contreras, ed., Chapter 13.
Contreras, Jorge L., ed. 2019. The Cambridge Handbook of Technical Standardization Law: Intersections of
Public and Private Law. New York: Cambridge Univ. Press (forthcoming).
Contreras, Jorge L. 2019a. “Trademarks, Certification Marks and Technical Standards,” in Contreras, ed.,
Chapter 12 (forthcoming).
Contreras, Jorge L. 2019b. “Global Rate-Setting: A Solution for Standards-Essential Patents?,” 94 Washington
Contreras, Jorge L. 2019c. “Much Ado About Hold-up,” University of Illinois Law Review (forthcoming).
Contreras, Jorge L. and Andrew Updegrove. 2016. “A Primer on Intellectual Property Policies of Standards
Bodies,” in Kai Jacobs, ed., Effective Standardization Management in Corporate Settings. Hershey, PA: IGI
Global.
Contreras, Jorge L. and Anne Layne-Farrar. 2018. “Non-Discrimination and FRAND Commitments,” in
Contreras, ed., Chapter 12.
Contreras, Jorge L. and Charles R. McManis. 2013. “Intellectual Property Landscape of Material Sustainability
Standards,” 14 Columbia Science & Technology Law Review 485–558.
Contreras, Jorge L. and David L. Newman. 2014. “Developing a Framework for Arbitrating Standards-
Essential Patent Disputes,” 2014 Journal of Dispute Resolution 23–50.
Contreras, Jorge L. and Richard J. Gilbert. 2015. “A Unified Framework for RAND and other Reasonable
Royalties,” 30 Berkeley Technology Law Journal 1447–504.
Contreras, Jorge and Russ Housley. 2008. “IP Squared: Internet Standards and Intellectual Property,” Nov./
Dec. IEEE Internet Computing 80–86.
Contreras, Jorge L., et al. 2019. “The Effect of FRAND Commitments on Patent Remedies”, in Biddle, et al.,
eds., Chapter 5.
Contreras, Jorge L., Colleen Chien, Thomas Cotter, and Brad Biddle. 2016. Study Proposal—Commercial
Patent Licensing Data, https://papers.ssrn.com/sol3/papers.cfm?abstract_id=2755706.
Contreras, Jorge L., Fabian Gaessler, Christian Helmers, and Brian J. Love. 2017. “Litigation of Standards-
Essential Patents in Europe: A Comparative Analysis,” 32 Berkeley Technology Law Journal 1457–87.
Cotter, Thomas F. 2014. “The Comparative Law and Economics of Standard-Essential Patents and FRAND
Royalties,” 22 Texas Intellectual Property Law Journal 311–63.
Cowhey, P.F., J.D. Aronson, and J.E. Richards. 2006. “The Peculiar Evolution of 3G Wireless Networks:
Institutional Logic, Politics and Property Rights,” in J. Zysman and A. Newman, eds., How Revolutionary
Was the Digital Revolution? National Responses, Market Transitions, and Global Technology. Stanford, CA:
Cyber Creative Inst. Co. Ltd. 2013. “Evaluation of LTE Essential Patents Declared to ETSI, Ver. 3.0,” www.
cybersoken.com/file/lte03SummaryEN.pdf.
De Carvalho, Nuno Pires. 2015. “Technical Standards, Intellectual Property, and Competition—A Holistic
View”, 47 Washington University Journal of Law & Policy 61–130.
DeLacey, Brian, et al. 2006. “Strategic Behavior in Standard-Setting Organizations,” Harvard NOM Working
Paper No. 903214, available at http://papers.ssrn.com/sol3/papers.cfm?abstract_id=903214.
Delcamp, Henry and Aija Leiponen. 2014. “Innovating Standards Through Informal Consortia: The Case of
Wireless Telecommunications,” 36 International Journal of Industrial Organization 36–47.
DeNardis, Laura. 2009. “Open Standards and Global Politics,” 13 International Journal of Communications
Law & Policy 168–219.

Egyedi, Tineke M. 2001a. “Beyond Consortia, Beyond Standardisation? New Case Material and Policy
Threads—Final Report for the European Commission,” 1 Advanced Topics in Information Technology
Standards and Standardization Research 91–110.
Egyedi, Tineke M. 2001b. “IPR Paralysis in Standardization: Is Regulatory Symmetry Desirable?,” Apr. IEEE
Communications Magazine 108–14.
Egyedi, Tineke M. 2016. “On Technology Choice in Standardization: Implications for Patent Valuation,” 13th
Intl. Conference—Standardization, Prototypes and Quality: A Means of Balkan Countries’ Collaboration,
Nov. 3–4.
Elhauge, Einer. 2008. “Do Patent Holdup and Royalty Stacking Lead to Systematically Excessive Royalties?,” 4
Eltzroth, Carter. 2008. “IPR Policy of the DVB Project: Negative Disclosure, FR&ND Arbitration Unless Pool
Rules OK, Parts 1 and 2,” 6(2) and 7(2) Journal of IT Standards & Standardization Research 21 and 1.
Epstein, Richard A., F. Scott Kieff, and Daniel F. Spulber. 2012. “The FTC, IP, and SSOs: Government
Hold-Up Replacing Private Coordination,” 8 Journal of Competition Law & Economics 1–46.
Ernst, Dieter. 2012. “America’s Voluntary Standards System—A Best Practice Model for Innovation Policy?,”
East-West Center Working Paper No. 128.
Ernst, Dieter, Heejin Lee, and Jooyoung Kwak. 2014. “Standards, Innovation, and Latecomer Economic
Development: Conceptual Issues and Policy Challenges,” 38 Telecommunications Policy 853–62.
Fairfield Resources, Inc. 2007. “Analysis of Patents Declared as Essential to GSM as of June 6, 2007,” Working
Paper Dec. 31.
Fairfield Resources, Inc. 2010. “Review of Patents Declared as Essential to LTE and SAE (4G Wireless
Standards) Through June 30, 2009,” Working Paper Jan. 6.
Farrell, Joseph and Carl Shapiro. 1992. “Standard Setting in High Definition Television,” 1992 Brookings
Papers: Microeconomics 1–93.
Farrell, Joseph and G. Saloner. 1988. “Coordination Through Committees and Markets,” 19 RAND Journal of
Economics 235–52.
Farrell, Joseph and Timothy Simcoe. 2012. “Choosing the Rules for Consensus Standardization,” 43 RAND
Farrell, Joseph, John Hayes, Carl Shapiro, and Theresa Sullivan. 2007. “Standard Setting, Patents and
Hold-Up,” 74 Antitrust Law Journal 603–70.
Fischer, Carolyn and Thomas P. Lyon. 2014. “Competing Environmental Labels,” 23 Journal of Economics &
Froomkin, A. Michael. 2003. “Habermas@Discourse.net: Toward a Critical Theory of Cyberspace,” 116
Galetovic, Alexander and Kirti Gupta. 2016. “Royalty Stacking and Standard Essential Patents: Theory and
Evidence from the World Mobile Wireless Industry” (Working Paper, June), www.law.northwestern.edu/resea
rch-faculty/searlecenter/events/innovation/documents/Galetovic_Royalty_stacking_060416_GG.pdf.
Galetovic, Alexander, Stephen Haber, and Ross Levine. 2015. “An Empirical Examination of Patent Hold-Up,”
11 Journal of Competition Law & Economics 549–78.
Galetovic, Alexander, Stephen Haber, and Lew Zaretski. 2018. “An Estimate of the Average Cumulative Royalty
Yield in the World Mobile Phone Industry: Theory, Measurement and Results,” 42 Telecommunications Policy
263–76.
Gamalielsson, Jonas and Björn Lundell. 2016. “On Involvement in Open Standards: How Do Organisations
Contribute to W3C Standards through Editorship,” in K. Crowston, I. Hammouda, B. Lundell, G. Robles,
J. Gamalielsson, and J. Lindman, eds., Open Source Systems: Integrating Communities, IFIP Advances in
Information and Communication Technology, Vol 472. Springer.
Gandal, Neil and Pierre Régibeau. 2015. “Standard-Setting Organisations: Current Policy Issues and Empirical
Evidence”, in Panagiotis Delimatsis, ed., The Law, Economics and Politics of Standardization. Cambridge,
UK: Cambridge Univ. Press.
Gandal, Neil, Nataly Gantman, and David Genesove. 2007. “Intellectual Property and Standardization
Committee Participation in the US Modem Industry,” in Shane Greenstein and Victor Stango, eds., Standards
and Public Policy. New York: Cambridge Univ. Press.
Ganglmair, Bernhard, Luke M. Froeb, and Gregory J. Werden. 2012. “Patent Hold-Up and Antitrust: How a
Well-Intentioned Rule Could Retard Innovation,” 60 Journal of Industrial Economics 249–73.
Geradin, Damien. 2010. “Reverse Hold-ups: The (Often Ignored) Risks Faced by Innovators in Standardized
Areas,” Working Paper.
Geradin, Damien and Anne Layne-Farrar. 2011. “Patent Value Apportionment Rules for Complex, Multi-
Patent Products,” 27 Santa Clara Computer & High Technology Law Journal 763–93.
Geradin, Damien and Miguel Rato. 2007. “Can Standard-Setting Lead to Exploitative Abuse? A Dissonant
View on Patent Hold-Up, Royalty-Stacking and the Meaning of FRAND,” 3 European Competition Journal
101–61.

Geradin, Damien, Anne Layne-Farrar, and A. Jorge Padilla. 2007. “The Ex Ante Auction Model for the Control
of Market Power in Standard Setting Organizations,” Working Paper, http://papers.ssrn.com/sol3/papers.
Geradin, Damien, Anne Layne-Farrar, and A. Jorge Padilla. 2008. “The Complements Problem Within
Standard Setting: Assessing the Evidence of Royalty Stacking,” 14 Boston University Journal of Science &
Technology Law 144–76.
Ghosh, Rishab. 2011. “An Economic Basis for Open Standards,” in Laura DeNardis, ed., Opening Standards:
The Global Politics of Interoperability 75–96. Cambridge, MA: MIT Press.
Gilbert, Richard J. 2011. “Deal or No Deal? Licensing Negotiations in Standard-Setting Organizations,” 77
Antitrust Law Journal 855.
Golden, John M. 2013. “Patent Privateers: Private Enforcement’s Historical Survivors,” 26 Harvard Journal of
Law & Technology 545–695.
Goodman, David J. and Robert A. Myers. 2005. “3G Cellular Standards and Patents,” Proceedings of IEEE
2005 Intl. Conference on Wireless Networks, Communications and Mobile Computing.
Greenbaum, Eli. 2015. “Forgetting FRAND. The WIPO Model Submission Agreements,” Jun. les Nouvelles 81–88.
Greenbaum, Eli. 2016. “Puzzles of the Zero-Rate Royalty,” 27 Fordham Intellectual Property, Media &
Greenstein, Shane and Marc Rysman. 2007. “Coordination Costs and Standard Setting: Lessons From
56K Modems,” in Shane Greenstein and Victor Stango, eds., Standards and Public Policy 123. New York:
Cambridge Univ. Press.
Gupta, Kirti. 2013. “The Patent Policy Debate in the High-Tech World,” 9 Journal of Competition Law &
Economics 827–58.
Gupta, Kirti and Mark Snyder. 2014. “Smart Phone Litigation and Standard Essential Patents,” Hoover
Institution Working Paper 14006.
Harhoff, Dietmar, Francis Narin, F.M. Scherer, and Katrin Vopel. 1999. “Citation Frequency and the Value of
Patented Inventions,” 81 Review of Economics and Statistics 511–15.
Herman, Michele K. 2010. “How the Deal is Done, Part 1,” Sept.–Oct. Landslide 35–41.
Hunt, Robert M., Samuli Simojoki, and Tuomas Takalo. 2007. “Intellectual Property Rights and Standard
Setting in Financial Services: The Case of the Single European Payments Area,” Federal Reserve Bank of
Philadelphia Working Paper 07-20, available at http://papers.ssrn.com/sol3/papers.cfm?abstract_id=1018483.
Jillavenkatesa, Ajit, H. Evans, and Henry Wixon. 2012. “Patents and Intellectual Property Management in
Nanotechnology Standardization: A NIST Perspective”, presented at National Academies of Science Symposium
on Management of IP in Standards-Setting Processes, Session 4, available at http://sites.nationalacademies.org/
cs/groups/pgasite/documents/webpage/pga_072332.pdf.
Kang, Byeongwoo and Rudi Bekkers. 2015. “Just-in-time Patents and the Development of Standards,” 44
Katz, M.L. and Carl Shapiro. 1985. “Network Externalities, Competition, and Compatibility,” 75 American
Kesan, Jay P. 2019. “Open Standards,” in Contreras, ed., Chapter 9 (forthcoming).
Kesan, Jay P. and Carol M. Hayes. 2014. “FRAND’s Forever: Standards, Patent Transfers, and Licensing
Commitments,” 89 Indiana Law Journal 231–314.
Kieff, F. Scott and Anne Layne-Farrar. 2013. “Incentive Effects from Different Approaches to Holdup
Mitigation Surrounding Patent Remedies and Standard-Setting Organizations,” 9 Journal of Competition
Law & Economics 1091–123.
Krechmer, Ken. 2011. “Open Standards: Definition and Policy,” in Laura DeNardis, ed., Opening Standards:
The Global Politics of Interoperability 219. Cambridge, MA: MIT Press.
Kühn, Kai-Uwe, Fiona Scott Morton, and Howard Shelanski. 2013. “Standard Setting Organizations Can Help
Solve the Standard Essential Patents Licensing Problem,” 3 CPI Antitrust Chronicle 1.
Kumar, Sapna and Arti Rai. 2007. “Synthetic Biology: The Intellectual Property Puzzle,” 85 Texas Law Review
1745–68.
Langlois, Richard N. 2006. “Competition Through Institutional Form: The Case of Cluster Tool Standards,”
in Shane Greenstein and Victor Stango, eds., Standards and Public Policy 2 60–82. New York: Cambridge
Univ. Press.
Larouche, Pierre and Nicolo Zingales. 2018. “Injunctive Relief in the EU—Intellectual Property and
Competition Law at the Remedies Stage”, in Contreras, ed., Chapter 25.
Larouche, Pierre, Jorge Padilla, and Richard S. Taffet, “Settling FRAND Disputes: Is Mandatory Arbitration
a Reasonable and Nondiscriminatory Alternative?,” 10 Journal of Competition Law & Economics 581–610.
Layne-Farrar, Anne. 2009. “Preventing Patent Holdup: An Economic Assessment of Ex Ante Licensing
Negotiations in Standard Setting,” 37 AIPLA Quarterly Journal 445–78.
Layne-Farrar, Anne. 2010a. “Is Ex Ante the Norm? An Empirical Look at IPR Disclosure Timing within
Standard Setting,” Working Paper.

Layne-Farrar, Anne. 2010b. “ Non-Discriminatory Pricing: Is Standard Setting Different?,” 6 Journal of

Layne-Farrar, Anne. 2011. “Innovative or Indefensible? An Empirical Assessment of Patenting within Standard
Setting,” 9 IT Standards & Standardization Research 1–18.
Layne-Farrar, Anne. 2014a. “Proactive or Reactive? An Empirical Assessment of IPR Policy Revisions in the
Wake of Antitrust Actions,” 59 Antitrust Bulletin 373–429.
Layne-Farrar, Anne. 2014b. “Patent Holdup and Royalty Stacking Theory and Evidence: Where Do We Stand
After 15 Years of History?” (Submitted for 122nd Meeting of the OECD Competition Committee, Dec. 17–18).
Layne-Farrar, Anne. 2014c. “Moving Past The SEP RAND Obsession: Some Thoughts On The Economic
Implications Of Unilateral Commitments And The Complexities Of Patent Licensing,” 21 George Mason
Layne-Farrar, Anne and Koren Wong-Ervin. 2014. “Methodologies for Calculating FRAND Damages (Parts
1–3),” Oct. Law360 8–10.
Layne-Farrar, Anne, A. Jorge Padilla, and Richard Schmalensee. 2007. “Pricing Patents for Licensing in
Standard-Setting Organizations: Making Sense of FRAND Commitments,” 74 Antitrust Law Journal
671–706.
Lehr, William. 1995. “Compatibility Standards and Interoperability: Lessons From the Internet,” in Brian
Kahin and Janet Abbate, eds., Standards Policy for Information Infrastructure 121. Cambridge, MA: MIT
Press.
Leiponen, Aija Elina. 2008. “Competing Through Cooperation: The Organization of Standard Setting in
Wireless Telecommunications,” 54 Management Science 1904–19.
Lemley, Mark A. 2002. “Intellectual Property Rights and Standard-Setting Organizations,” 90 California Law
Review 1889–980.
Lemley, Mark A. 2007. “Ten Things to Do About Patent Holdup of Standards (and One Not to),” 48 Boston
Lemley, Mark A. and Carl Shapiro. 2007. “Patent Holdup and Royalty Stacking,” 85 Texas Law Review
1991–2050.
Lemley, Mark A. and Carl Shapiro. 2013. “A Simple Approach to Setting Reasonable Royalties for Standard-
Essential Patents,” 28 Berkeley Technology Law Journal 1135–583.
Lemley, Mark A. and David McGowan. 1999. “Could Java Change Everything? The Competitive Propriety of
a Proprietary Standard (Parts 1 and 2),” 43 Antitrust Bulletin 715–74.
Lemley, Mark A. and Timothy Simcoe. 2018. “How Essential are Standards-Essential Patents?” Stanford Public
Law Working Paper (March), https://papers.ssrn.com/sol3/papers.cfm?abstract_id=3128420.
Lerner, Josh and Jean Tirole. 2006. “A Model of Forum Shopping,” 96(4) American Economic Review 1091–1113.
Lerner, Josh and Jean Tirole. 2015. “Standard-Essential Patents,” 123 Journal of Political Economics 547–86.
Lichtman, Doug. 2010. “Understanding the RAND Commitment,” 47 Houston Law Review 1023–1421.
Lindsay, Michael A. and Konstantinos Karachalios. 2015. “Updating a Patent Policy: The IEEE Experience,”
CPI Antirust Chronicle.
Lundell, Björn, Jonas Gamalielsson, and Andrew Katz. 2015. “On Implementation of Open Standards in
Software: To What Extent Can ISO Standards be Implemented in Open Source Software?,” 13 International
Journal of Standardization Research 47–73.
Lundqvist, Björn. 2014. Standardization Under EU Competition Rules and US Antitrust Laws. Cheltenham,
UK: Edward Elgar.
Malamud, Carl. 2014. Testimony at Hearing on the Scope of Copyright Protection, House Judiciary Committee,
U.S. House of Representatives, Jan. 14.
Mallinson, Keith. 2015. “Smartphone Revolution: Technology Patenting and Licensing Fosters Innovation,
Market Entry, and Exceptional Growth,” Apr. IEEE Consumer Electronics Magazine 60–66.
Martin, Donald L. and Carl De Meyer. 2006. “Patent Counting, A Misleading Index of Patent Value: A
Critique of Goodman & Meyers and its Uses,” Working Paper, available at http://papers.ssrn.com/sol3/
Masur, Jonathan. 2015. “The Use and Abuse of Patent Licenses,” 110 Northwestern Law Review 115–58.
Mattli, Walter and Tim Büthe. 2003. “Setting International Standards: Technological Rationality or Primacy
of Power?,” 56 World Politics 1–42.
Mendelson, Nina A. 2015. “Taking Public Access to the Law Seriously: The Problem of Private Control Over
the Availability of Federal Standards,” 45 Environmental Law Reporter 10776–82.
Merges, Robert P. and Jeffrey M. Kuhn. 2009. “An Estoppel Doctrine for Patented Standards,” 97 California
Law Review 1–50.
Michel, Suzanne. 2011. “Bargaining for RAND Royalties in the Shadow of Patent Remedies Law,” 77 Antitrust
Miller, Joseph Scott. 2007. “Standard Setting, Patents, and Access Lock-In: RAND Licensing and the Theory
of the Firm,” 40 Indiana Law Review 351–96.

Murphy, Craig N. and Joanne Yates. 2009. ISO, the International Organization for Standardization: Global
Governance through Voluntary Consensus. Abingdon, UK and New York, NY: Routledge.
Natl. Research Council (NRC). 2013. Keith Maskus and Stephen A. Merrill, eds., Intellectual Property
Challenges For Standard-Setting in the Global Economy. Washington, DC: National Academies Press.
Neil, Suzanne, Lee McKnight, and Joseph Bailey. 1995. “The Government’s Role in the HDTV Standards
Process: Model or Aberration?,” in Brian Kahin and Janet Abbate, eds., Standards Policy for Information
Infrastructure 276. Cambridge, MA: MIT Press.
Nickerson, Jeffrey V. and Michael zur Muehlen. 2006. “The Ecology of Standards Processes: Insights from
Internet Standard Making,” 30 MIS Quarterly 467–88.
Ohana, Gil, Marc Hansen, and Omar Shah. 2003. “Disclosure and Negotiation of Licensing Terms Prior to
Adoption of Industry Standards: Preventing Another Patent Ambush?,” 12 European Competition Law
Review 644–56.
Pentheroudakis, Chryssoula and Justus A. Baron. 2017. Licensing Terms of Standard Essential Patents: A
Comprehensive Analysis of Cases. JRC Science for Policy Report EUR 28302, http://publications.jrc.ec.europa.
eu/repository/bitstream/JRC104068/jrc104068%20online.pdf.
Petit, Nicolas. 2016. “The Smallest Salable Patent-Practicing Unit (“SSPPU”) Experiment, General
Purpose Technologies and the Coase Theorem,” Working Paper, https://papers.ssrn.com/sol3/Papers.cfm?
abstract_id=2734245.
Pohlmann, Tim and Knut Blind. 2016. Landscaping Study on Standard-Essential Patents (SEPs), Report
Commissioned by European Commission DG GROW Unit F.5, available at http://ec.europa.eu/growth/tools-
databases/newsroom/cf/itemdetail.cfm?item_id=8981.
Pohlmann, Tim, Peter Neuhäusler, and Knut Blind. 2015. “Standard Essential Patents to Boost Financial
Returns,” 46 R&D Management 612–30.
Putnam, Jonathan D. and Tim A. Williams. 2016. “The Smallest Salable Patent-Practicing Unit (SSPPU): Theory
and Evidence,” Working Paper, available at https://papers.ssrn.com/sol3/papers.cfm?abstract_id=2835617.
Régibeau, Pierre, Raphaël De Coninck, and Hans Zenger. 2016. “Transparency, Predictability, and Efficiency
of SSO-based Standardization and SEP Licensing: A Report for the European Commission”.
Rens, Andrew. 2011. “Open Document Standard for Government: The South African Experience,” in Laura
DeNardis, ed., Opening Standards: The Global Politics of Interoperability 63. Cambridge, MA: MIT Press.
Russell, Andrew L. 2011. “Constructing Legitimacy: The W3C’s Patent Policy,” in Laura DeNardis, ed.,
Opening Standards: The Global Politics of Interoperability 159. Cambridge, MA: MIT Press.
Russell, Andrew L. 2014. Open Standards and the Digital Age: History, Ideology, and Networks. New York, NY:
Rysman, Marc and Timothy Simcoe. 2008. “Patents and the Performance of Voluntary Standard-Setting
Organizations,” 54 Management Science 1920–34.
Rysman, Marc and Timothy Simcoe. 2011. “A NAASTy Alternative to RAND Pricing Commitments,” 35
Telecommunications Policy 1010–17.
Samuelson, Pamela. 2007. “Questioning Copyrights in Standards,” 48 Boston College Law Review 193–224.
Samuelson, Pamela and Kathryn Hashimoto. 2019. “Questioning Copyright in Standards”, in Contreras, ed.,
Chapter 5 (forthcoming).
Shah, Rajiv C. and Jay P. Kesan. 2007. “Open Standards and the Role of Politics,” Proceedings of the 8th Annual
International Digital Government Research Conference.
Shah, Rajiv C. and Jay P. Kesan. 2012. “Lost in Translation: Interoperability Issues for Open Standards,” 8 I/S
Journal of Law & Policy 113–41.
Shapiro, Carl. 2001. “Navigating the Patent Thicket: Cross Licenses, Patent Pools, and Standard Setting,” in
Adam Jaffe, Josh Lerner, and Scott Stern, eds., Innovation Policy and the Economy: Vol. 1 119. Cambridge,
MA: MIT Press.
Shapiro, Carl. 2010. “Injunctions, Hold-Up, and Patent Royalties,” 12 American Law & Economics Review 280.
Shapiro, Carl and Hal R. Varian. 1999. Information Rules: A Strategic Guide to the Network Economy. Boston,
Sheffner, Daniel J. 2019. “Integrating Technical Standards into Federal Regulations: Incorporation by
Reference,” in Contreras, ed., (forthcoming).
Shurmer, Mark and Gary Lea. 1995. “Telecommunications Standardization and Intellectual Property Rights:
A Fundamental Dilemma?,” in Brian Kahin and Janet Abbate, eds., Standards Policy for Information
Infrastructure 378. Cambridge, MA: MIT Press.
Sidak, J. Gregory. 2008. “Holdup, Royalty Stacking, and the Presumption of Injunctive Relief for Patent
Infringement: A Reply to Lemley and Shapiro,” 92 Minnesota Law Review 714–48.
Sidak, J. Gregory. 2009. “Patent Holdup and Oligopsonistic Collusion in Standard-Setting Organizations,” 5
Sidak, J. Gregory. 2013. “The Meaning of FRAND, Part I: Royalties,” 9 Journal of Competition Law &

Sidak, J. Gregory. 2014. “The Proper Royalty Base for Patent Damages,” 10 Journal of Competition Law &
Sidak, J. Gregory. 2015a. “The Meaning of FRAND, Part II: Injunctions,” 11 Journal of Competition Law &
Economics 201–69.
Sidak, J. Gregory. 2015b. “The Antitrust Division’s Devaluation of Standards-Essential Patents,” 104
Georgetown Law Journal Online 48–73.
Sidak, J. Gregory. 2016a. “Apportionment, FRAND Royalties, and Comparable Licenses After Ericsson v.
D-Link,” Illinois Law Review 1810–70.
Sidak, J. Gregory. 2016b. “What Aggregate Royalty to Manufacturers of Mobile Phones Pay to License
Standard-Essential Patents?,” 1 Criterion Journal of Innovation 701–19.
Sidak, J. Gregory. 2018. “Injunctive Relief and the FRAND Commitment in the United States,” in Contreras,
ed., Chapter 24.
Siebrasse, Norman. 2019. “Holdup, Holdout and Royalty Stacking: A Review of the Literature,” in Biddle, et
al., eds., (forthcoming).
Siebrasse, Norman and Thomas F. Cotter. 2017a. “The Value of the Standard,” 101 Minnesota Law Review
1159–246.
Siebrasse, Norman and Thomas F. Cotter. 2017b. “Judicially Determined FRAND Royalties,” in Contreras,
ed. (2018), Chapter 23.
Simcoe, Timothy. 2007a. “Explaining the Increase in Intellectual Property Disclosure,” in Sherri Bolin, ed., The
Standards Edge. Bolin Group.
Simcoe, Timothy. 2007b. “Delay and de jure Standardization: Exploring the Slowdown in Internet Standards
Development,” in Shane Greenstein and Victor Stango, eds., Standards and Public Policy 260. New York, NY:
Simcoe, Timothy. 2012. “Standard Setting Committees: Consensus Governance for Shared Technology
Platforms,” 102 American Economics Review 305–36.
Simcoe, Timothy S., Stuart J. H. Graham, and Maryann P. Feldman. 2009. “Competing on Standards?
Entrepreneurship, Intellectual Property, and Platform Technologies,” 18 Journal of Economics and Management
Strategy 775–816.
Skitol, Robert. 2005. “Concerted Buying Power: Its Potential for Addressing the Patent Holdup Problem in
Standard Setting,” 72 Antitrust Law Journal 727–44.
Spring, Michael B. and Martin B. H. Weiss. 1995. “Financing the Standards Development Process,” in Brian
Kahin and Janet Abbate, eds., Standards Policy for Information Infrastructure 289. Cambridge, MA: MIT Press.
Stasik, Erik. 2010. “Royalty Rates and Licensing Strategies for Essential Patents on LTE (4G) Telecommunication
Standards,” Sept. les Nouvelles 114–19.
Steinfield, Charles W. et al. 2007. “Promoting E-business Through Vertical IS Standards: Lessons From the US
Home Mortgage Industry,” in Shane Greenstein and Victor Stango, eds., Standards and Public Policy 160.
New York, NY: Cambridge Univ. Press.
Stoll, Thimo. 2014. “Are You Still In?—The Impact of Licensing Requirements on the Composition of Standards
Setting Organizations,” Max Planck Institute for Innovation & Computer Research Paper No. 14-18.
Strauss, Peter. 2013. “Private Standards Organizations and Public Law,” 22 William & Mary Bill of Rights Law
Journal 497–562.
Swanson, Daniel G. and William J. Baumol. 2005. “Reasonable and Nondiscriminatory (RAND) Royalties,
Standards Selection, and Control of Market Power,” 73 Antitrust Law Journal 1–58.
Tapia, Claudia. 2010. Industrial Property Rights, Technical Standards and Licensing Practices (FRAND) in the
Telecommunications Industry. Köln, DE: Carl Heymanns Verlag.
Teece, David J. and Edward F. Sherry. 2003. “Standards Setting and Antitrust,” 87 Minnesota Law Review
1913–2173.
Teece, David J. and Edward F. Sherry. 2016. “On the ‘Smallest Salable Patent Practicing Unit’ Doctrine: An
Economic and Public Policy Analysis,” Tusher Center for the Management of Intellectual Capital, Working
Paper Series No. 11.
Teece, David J., Peter C. Grindley, and Edward F. Sherry. 2013. “SDO IPR Policies in Dynamic Industries,”
submitted to National Academies of Science in connection with its Oct. 2012 Symposium on Intellectual
Property in Standard-Setting.
Torrance, Andrew W. and Linda J. Kahl. 2014. “Bringing Standards to Life: Synthetic Biology Standards and
Intellectual Property,” 30 Santa Clara High Technology Law Journal 199–555.
Trajtenberg, Manuel. 1990. “A Penny for Your Quotes: Patent Citations and the Value of Innovations,” 21
Tsai, Joanna and Joshua D. Wright. 2015. “Standard Setting, Intellectual Property Rights, and the Role of
Antitrust in Regulating Incomplete Contracts,” 80 Antitrust Law Journal 157–83.
Tsilas, Nicos L. 2004. “Toward Greater Clarity and Consistency in Patent Disclosure Policies in a Post-Rambus
World,” 17 Harvard Journal of Law & Technology 475–619.

Updegrove, Andrew. 1995. “Consortia and the Role of the Government in Standard Setting”, in Brian Kahin
and Janet Abbate, eds., Standards Policy for Information Infrastructure 321. Cambridge, MA: MIT Press.
Updegrove, Andrew. 2003. “Survey: Major Standards Players Tell How They Evaluate Standard Setting
Organizations,” 2 Consortium Standards Bull.
Updegrove, Andrew. 2006. “Ex Ante Disclosure: Risks, Rewards, Process and Alternatives,” 5 Consortium
Standards Bull. 1.
Updegrove, Andrew. 2009. “A Concise Introduction to Free and Open Software,” 8 Standards Today.
Updegrove, Andrew. 2012. “Openness and Legitimacy in Standards Development,” in Shane Coughlan, ed.,
First OpenForum Academy Conference Proceedings 5. OpenForum Europe Ltd.
Updegrove, Andrew. 2015. “Standard Setting Organizations and Standards List,” www.consortiuminfo.org/
links/#.VarhPnjDRD0.
Vetter, Greg R. 2007. “Open Source Licensing and Scattering Opportunism in Software Standards,” 48 Boston
Wen, Wen, Chris Forman, and Sirkka Jarvenpaa. 2015. “Standards, Intellectual Property Rights, and Strategic
Patenting: Evidence from IETF,” Georgia Tech Scheller College of Business Research Paper No. 35, available
at https://papers.ssrn.com/sol3/papers.cfm?abstract_id=2709645.
West, Joel. 2007. “The Economic Realities of Open Standards: Black, White, and Many Shades of Gray,”
in Shane Greenstein and Victor Stango, eds., Standards and Public Policy 87. New York, NY: Cambridge
University Press.
Wright, F.D. 2008. “The IEEE Standards Association and its Ecosystem,” IEEE Communications Magazine.

10. Empirical studies of patent pools
Michael Mattioli*
Contents
I. Introduction
II. Assembling the Historical Record
III. Placing Competition Studies in Context
IV. Assessing the Impact of Patent Pools on Competition and Innovation
V. Assessing the Impact of Patent Pools on Patent Litigation
VI. Assessing the Impact of Patent Pools on Transaction Costs
VII. Patent Pools as Forums for Valuation
VIII. Conclusion
References
I. INTRODUCTION
This chapter provides an overview of the legal and economic impact of patent pools,
with a special focus on the methods scholars have used to study these institutions empiri-
cally. Patent pools are privately governed institutions that license complementary patent
rights under unified agreements.1 Licensees of the aggregated rights typically include
members of the pool (i.e., licensors), and technology manufacturers, service providers,
or researchers who are not members (sometimes referred to herein as “pure licensees”).
A primary purpose of this organizational form is to reduce transaction costs—sources
of economic friction that theorists believe can cause patented technologies to go underused
and undeveloped.2 To appreciate how patent pools do this, it is helpful to understand the
sources of such costs: as the technological sophistication of products and processes has
increased since the patent system’s earliest days, so too has the number of related patent
rights and, consequently, the number of patent holders that might lay claim to parts of
a product or technology. As a result, manufacturers that wish to assemble products from
diverse technological components must also assemble the necessary patent licenses.
This presents several potential transaction costs. First, the manufacturer must research
the set of patents that read upon its planned product or hire a lawyer to conduct such a
search. Then, it must identify the owners of those patents—not a trivial challenge, espe-
cially if patents have changed owners in the course of a corporate merger, acquisition, or
* Professor of Law, Indiana University Maurer School of Law, Indiana University.

1
Here, the term “complementary” has a special meaning, laid out later in this discussion.
2
Another important purpose, discussed later in this chapter, is to prevent or settle patent
disputes.
236

Empirical studies of patent pools 237
outright sale. Finally, the manufacturer must attempt to secure a license from each patent
holder. This final step presents a risk that has been recognized by mathematicians and
economists for centuries: double-marginalization. In simplified terms, the problem is that
the sum of prices charged by multiple monopolists acting independently is expected usu-
ally to be far higher than the price the same monopolists would rationally charge if they
coordinated (Hovenkamp and Hovenkamp, 2015; Lemley and Shapiro, 2013; Cournot,
1838). Relatedly, any patent holder that becomes aware that it holds a necessary piece
of the puzzle may demand supra-competitive royalties from licensees—the well-known
“holdout” or “holdup” problem. Economic theorists believe that, in the minds of some
would-be licensees, these costs and risks may often just be too great to bear. As a result,
at least some manufacturers and service providers (perhaps many) will elect not to sell
the products or services they might wish to. All other things being equal, this result is
thought to be undesirable not only to would-be licensees but also to patent holders and
consumers.3
Patent pools solve these problems by promising licensees all (or effectively all) of the
complementary patent permissions they need through a single transaction—a “one-stop-
shop,” in effect. By agreeing beforehand to license their related patents through a single
contract, the patent holders compress what would ordinarily be a series of transactions
to a single event, dramatically reducing transaction costs (Merges and Mattioli, 2017).
By offering the same license to multiple licensees, the pool permits competition in down-
stream markets that draw upon the underlying technology (Hovenkamp and Hovenkamp,
2015). Pooling also makes it difficult for a single licensor who has committed to the pool
to strategically hold out for excessive royalties.
Patent pools impose some costs, of course. The patent holders must identify each other
at the outset, cooperate to establish the pool, agree on a plan for collecting royalties from
licensees, and develop a plan for dividing the spoils. (As discussed below, they may under
some circumstances impose other costs that are less clear ex ante, such as competitive
harms.) It speaks to the perceived power of this cooperative form that the patent holders
that form pools expect such costs to be outweighed by the benefits of cooperation.
Although they are not formed frequently (Strandburg, 2011), patent pools have been
integral to the competitive production and commercialization of technologies central to
the U.S. economy since the nineteenth century—from sewing machines to cheap and reli-
able steel, to automobiles and aircraft (Vaughan, 1956). Since the 1990s, pools have taken
on special importance in the consumer electronics industry, where they have emerged to
facilitate the use of technological standards covering digital media and communications
(Merges, 1999; Lerner and Tirole, 2004; Contreras, 2019). There have also been notable
3
A robust discussion of the potential harms of sub-optimal patent licensing is beyond the scope
of this volume, but a short summary is helpful to consider: First, under the conditions described,
patent holders would likely receive royalties from fewer licensees and perhaps no licensees at all. If
some number of manufacturers had the wherewithal to obtain the necessary licenses, that number
might be too low to allow for healthy competition in the relevant market for products or services.
This would disadvantage consumers, who would have fewer choices and would likely need to pay
higher prices. Those consumers unable to pay these higher costs would constitute a deadweight
loss. Meanwhile, some number of manufacturers and licensees that would have wished to offer the
relevant products or services would by necessity devote their resources to less preferred enterprises.

recent attempts to pool patents related to medicines and disease treatments (Mattioli,
2012). Some such pools are organized around assembling “upstream” patent rights that
researchers may need to make “downstream” discoveries and advances in the fields of
biotechnology and drug development (Eisenberg and Heller, 1999; Mattioli, 2012).
In addition to facilitating the production of well-defined technologies and services,
some patent pools may encourage innovation through internal rules that allow them
to incorporate new patents over time (e.g., most contemporary standards-based pools
are designed to accommodate new patents). By holding out the promise of royalties to
would-be inventors that might be able to join in the future, pools can encourage research
and development in the narrow technology areas that they cover. (If the pool operates
in a setting where independent licensing is difficult absent coordination, this incentive
might be greater than the incentive that a patent would typically offer in the same setting.)
Limited evidence supports this theory: the formation of patent pools has been correlated
with technological advances in some industries (Shapiro, 2000; Flamm, 2012). Other
evidence suggests that patent acquisition behavior is more aggressive in the absence of
coordinated licensing, however (Ziedonis, 2004).
It is helpful to consider the nuanced relationships that can exist between patents in a
pool, the technologies those patents may cover, and in turn, the products and services
that could be based upon those technologies. Traditionally, pools often combined patents
covering distinct technologies that were valuable when used together, such as related
mechanisms in sewing machines (Mossoff, 2011), or interrelated steps in a process for
making steel (Mattioli, 2012). Technologies that work well together in this way are
typically called complements; technologies that are valuable only when used together are
typically called “pure” complements. Patents within a pool can also be vertically related,
such as when the practice of one patented invention necessarily requires infringing an
earlier patent—a blocking relationship.
These distinctions are relevant to the central concern that patent pools raise: the risk
of competitive harm. Despite their potential to facilitate the use of complementary
patents, pools have long concerned scholars and regulators for this reason. Readers
familiar with antitrust law and policy will find this unsurprising: antitrust regulators are
usually concerned when competitors cooperate. Patent pools typically include members
that compete with one another in product and service markets. (However, whether such
cooperation dampens competition in downstream markets is a complicated question,
as scholarship discussed later in this chapter explains.) The most common competitive
concern is that a single patent pool might include substitutive, rather than complementary
patents. This concern presents a special challenge because it is often difficult to character-
ize patent relationships as solely complementary or solely substitutive. Blocking claims in
two patents pooled together may be complementary, for instance, while the same patents
are substitutive in other respects (Hovenkamp and Hovenkamp, 2015). Brenda Simon
and Ted Sichelman (2017) note a secondary concern: “cross-licensing agreements among
incumbents in a market [including patent pools],” they note, “can afford them market
power that may exclude entrants from the market.” Antitrust regulators are tasked with
evaluating the competitive threat posed by specific pools.
Alongside these traditional competition-related concerns, a related vein of scholar-
ship suggests, somewhat counterintuitively, that patent pooling may reduce long-term
innovation. These concerns have been provoked by the use of so-called “grant-backs”—

contractual provisions that require patent pool members to license back to the pool any
future technologies they develop related to the subject matter of the pool. By removing
a patent holder’s ability to obtain an injunction, the theory goes, such provisions reduce
the incentive that patents represent. As this chapter explains, these concerns are largely
qualitative, relying primarily upon theory bolstered by selected anecdotes. As a result,
some scholars have recently begun to search for empirical evidence that sheds light upon
how patent pools affect competition and innovation.
In the late 1990s, a leading legal commentator observed that patent pooling theory
is “now well established, but empirical studies have yet to catch up” (Heller, 1998).
Today, nearly two decades later, this assessment is still apt: much remains unknown
about how these institutions form, operate, and influence the technological landscape.
Documentation is scarce, in part because patent pools are private institutions governed
by contracts that usually are not published. Publicly available sources of information that
scholars have relied upon include court decisions in which patent pools were charged with
antitrust violations (Vaughan, 1956; Mattioli, 2014), Congressional hearings dating to the
1930s (Merges, 1999), and business advisory letters published by the Antitrust Division
of the Department of Justice (Merges, 1999; Lerner and Tirole, 2004; Contreras, 2019).
The paucity of documentation on patent pools has not caused research into this field
to dry up, however. On the contrary, it has inspired scholars to be creative. Patent pool-
ing scholarship reflects a wide range of research methods—from traditional historical
research to the creative search for proxies for quantities that cannot be directly measured,
to the use of interviews and ethnographic research techniques (Merges and Mattioli,
2017). Among these methodologies, the ethnographic approach has gained critical
support and guidance from leading scholars. Most notably, Katherine Strandburg, Brett
Frischmann, and Michael Madison (2010) have provided a methodological framework for
examining a wide variety of information-sharing arrangements, including patent pools.
Their “Knowledge Commons” framework is adapted from and inspired by the Nobelist,
Elinor Ostrom.
This chapter summarizes the empirical methods that scholars have deployed to study
patent pools. In the interests of clarity and organization, some important works that dis-
cuss patent pools are not addressed, or are perhaps not addressed to the extent that some
readers might wish. The discussion is divided into five parts: first, sources that helped
develop the historical record of patent pools; second, methods that scholars have used
to assess the impact of patent pools on competition and innovation; third, works that
explore the relationship between patent pooling and litigation; fourth, new methods used
to estimate the transaction cost savings of patent pools; and fifth, works that examine
whether patent pools are useful forums for patent valuation. Because the present volume is
focused on analytical methods used within empirical studies, economic and legal scholar-
ship concerning pools that is not primarily empirical in nature (e.g., theoretical work) is
discussed only to help explain the motivation for, or conclusions of, empirical studies.
II. ASSEMBLING THE HISTORICAL RECORD
Because they are creatures of private ordering, patent pools present a unique challenge
for researchers: the historical record of these institutions is likely incomplete, and worse,

there is no way to know how much information is missing. Neither law nor common
industry practice strongly motivates patent pools to disclose (or for that matter, to even
preserve) information about their work, leaving open the possibility that many pools
have simply been lost to history. Scholars who have sought to assemble the historical
record have been resourceful in their methods, however, and the information they have
collected is a foundation upon which much empirical and theoretical patent pooling
scholarship rests.
Early academic investigations into patent pooling draw primarily upon court papers
originating from civil and criminal antitrust lawsuits brought against patent pools and
their members. Floyd L. Vaughan authored some of the earliest accounts of this kind.
In two volumes—The United States Patent System (1956) and The Economics of Our
Patent System (1925)—Vaughan reported on patent pool jurisprudence from the late
eighteenth century through the middle of the twentieth century. These volumes contain
accounts of the first known patent pool, an 1856 licensing arrangement among sewing
machine manufacturers, another pool formed decades later around agricultural tools, and
later pools relating to diverse technologies and industries, from raisin-seeding to motion
pictures. Vaughan’s helpful descriptive accounts included important information about
patent pool structure, governance, and composition. These accounts were based upon
descriptions of patent pools in published court decisions, government complaints, and
consent decrees. Perhaps most importantly, court dockets often contained briefs with
exhibits that included copies of patent pooling agreements. Although these sources neces-
sarily limited the scope of Vaughan’s work to patent pools involved in antitrust litigation,
his was the first attempt to assemble a useful catalog.
In a seminal law review article, Robert Merges (1999) added to the picture of early
twentieth-century patent pools by calling upon new sources. Significant among these was
a series of Congressional hearings on patent pooling held in 1935. This record contains
full copies of several patent pooling agreements dating to the 1910s, as well as illuminating
testimony of pool organizers and participants. (See “Legislative materials” for examples.)
Merges further expanded the historical record with new information about patent pools
formed in the 1990s based upon the MPEG and DVD technology standards. This infor-
mation mostly came from publicly available business advisory letters sent by the Antitrust
Division of the Department of Justice to certain pool organizers (Merges, 1999).
In 2005, Josh Lerner, Marcin Strojwas, and Jean Tirole (2005) added to the scholarly
catalog of patent pools. The scholars constructed a list of “all identifiable patent pools”
by combining the earlier work of Vaughan, Merges, and a third scholar, Steven C. Carlson
(1999). Working from this list, they obtained pooling documentation from a variety of
new sources, including Congressional hearings, exhibits appended to antitrust court
filings, and Freedom of Information Act requests directed to the Department of Justice
(Lerner et al., 2005). Some of these records were included within regional offices of the
National Archives. Their work yielded documentation on 63 pools in total.
Anne Layne-Farrar and Josh Lerner (2008) offered an even wider-angle picture of the
history of patent pooling and the determinants of licensing structure. Through various
industry sources, they obtained copies of membership agreements used by many contem-
porary standards-based patent pools (Layne-Farrar and Lerner, 2008). These included
the 1394, 3G Partnership Project, AVC, Bluetooth, DVB-T, DVD-1, DVD-2, MPEG-2,
and MPEG-4 patent pools. (By comparing these agreements to membership statistics

published by the administrators of these pools, the scholars observed that the governance
rules of patent pools appear to influence participation rates.) In a recent publication,
Justus Baron and Tim Pohlman (2015) assembled and published a rich dataset of 50
contemporary patent pools, including some that ultimately faltered and failed. This work
drew upon public announcements by pool organizers and unconventional online sources,
such as the Internet Archive and the websites of licensing administrators. Much of this
data has been made publicly available through a website maintained by the Searle Center
on Law, Regulation, and Economic Growth at the Northwestern Pritzker School of Law
(Searle Database, 2017).
Many of the scholars who have assembled the historical record of patent pools
described the institutional structures these groups adopted. In his case studies of over two
dozen patent pools, Vaughan (1956) observed a “diversity of organizational forms” and a
similar heterogeneity of goals and membership—qualities that seem to eschew generaliza-
tions about patent pooling. In their review of standards-based patent pools, Layne-Farrar
and Lerner (2008) examined governance rules and institutional structures. Mattioli (2014)
conducted a broad study of patent pools by drawing upon original contracts attached as
litigation exhibits stored in regional National Archives facilities. This study explained that
patent pools generally take two forms: contract-based patent pools and corporate patent
pools. Contract-based patent pools are built upon an agreement or a set of agreements
by which patent holders exchange mutual promises not to sue one another for patent
infringement. Members of contract-based patent pools typically agree to allow an agent
to execute patent licenses on their behalf, to collect royalties, and to apportion these
payments to the patent holders according to an agreed-upon formula. The article points
to a set of cross-licenses between the holders of sewing machine patents in the 1850s
(“the Singer Combination”) as a paradigmatic example of a contract-based patent pool.
Corporate patent pools, by contrast, acquire patents directly from their members through
assignments. In exchange for transferring their patents, members receive corporate shares
that convey equity in the joint enterprise, and in some cases, voting power and rights to
receive dividends.
III. PLACING COMPETITION STUDIES IN CONTEXT
Competition is a central theme in patent pooling scholarship. To appreciate the questions

that scholars have asked about pooling and competition, and the empirical methods they
have used to answer those questions, it is helpful to understand how courts and regulators
have regarded patent pools over time. The following compressed history of patent pooling
jurisprudence and regulation serves as a jumping-off point.
The first American patent pool was formed in October 1856, when the largest sewing
machine manufacturers in America—Grover, Baker, Wilson & Wilson, and I.A. Singer—
met in Albany, New York to work out a deal (Mossoff, 2011; Mattioli, 2011, 2014). For
years, these companies had been ensnared in patent litigation with one another, stemming
from the design and operation of their products. The cross-licensing solution they devel-
oped in Albany was the Singer Combination, named for its most famous member. These
licenses permitted each company to make and sell sewing machines that incorporated
technologies owned by other members. The agreements also imposed some significant

limitations: the partners promised, for instance, to limit the prices they would charge
consumers, and the numbers of machines each would manufacture (Singer Agreement,
1856). Similar restrictions would place future patent pools under the scrutiny of antitrust
authorities. In the 1850s, however, federal antitrust law was, as of yet, inchoate and no
legal action was taken against the group.
At the turn of the twentieth century, federal antitrust law had matured (most notably
through the passage of the Sherman and Clayton Acts), but patent pools similar to the
Singer Combination still enjoyed near unbridled freedom. In 1902, the Supreme Court
considered a dispute between a patent pool organized around harrows—agricultural tools
designed to break apart dirt—and one of its many licensees. E. Bement & Sons v. Nat’l
Harrow Company, 186 U.S. 70 (1902). The licensee argued that the pool had unlawfully
regulated prices and output through patent licenses. (These terms were quite similar to
those used by the Singer Combination decades earlier.) The Court disagreed, reasoning
that patent rights are, by definition, government-sanctioned monopolies, the exploitation
of which should not be hindered by antitrust: “[T]he general rule is absolute freedom in
the use or sale of rights under the patent laws of the United States,” the Court wrote. “The
very object of these laws is monopoly . . . The fact that the conditions in . . . contracts keep
up the monopoly or fix prices does not render them illegal.” The right to pool, it seemed,
was an entitlement that any group of patent holders could enjoy freely.
Just ten years later, the Supreme Court’s analysis in the decision of Standard Sanitary
Manufacturing v. United States (1912) marked an abrupt turnaround. As in National
Harrow (1902), the primary competition concern was price-fixing: three manufacturers
of enameled ware had formed a patent pool wherein they appointed a small committee to
set the prices at which licensees that manufactured products covered by the patents could
sell their wares. The Court held that the pool violated the Sherman Act, based on new
reasoning that the exclusive rights conferred by a patent should not trump the limitations
imposed by antitrust. “The added element of the patent in this case at bar cannot confer
immunity from a like condemnation,” the Court explained. “Rights conferred by patents
are indeed very definite and extensive, but they do not give any more than other rights a
universal license against positive prohibitions.”
Two decades later, the Supreme Court offered a critical insight in Standard Oil Co. v.
United States, 51 S. Ct. 421 (1931). The dispute involved a patent pool that contained
“blocking patents”—i.e., patents related to one another such that the practice of one
necessarily infringes upon the other, as is the case when a new invention improves upon an
old one. Although the patents were thus complementary with respect to blocking claims,
they disclosed and related to manufacturing methods that were competing. (The patents
related to various processes for “cracking,” a way of manufacturing gasoline.) Writing
for the majority, Justice Brandeis explained that the pooling of patents, including those
in blocking relationships, can be highly efficient: “If the available advantages [of the
patent pools] are open on reasonable terms to all manufacturers desiring to participate,
such interchange may promote rather than restrain competition,” he wrote. Brandeis
also explained that a patent pool does not violate the Sherman Act by demanding high
royalties, so long as those royalties are not tantamount to price-fixing by virtue of the
pool’s dominance over an entire industry: “[A]n agreement for cross-licensing and divi-
sion of royalties violates the Act only when used to effect a monopoly, or to fix prices,
or to impose otherwise an unreasonable restraint upon interstate commerce.” In short,

Brandeis indicated that the proper focus of an antitrust inquiry into a patent pool should
be on actual market effects.
National Harrow, 186 U.S. 70 (1902), Standard Sanitary Manufacturing, 226 U.S. 20
(1912), and Standard Oil, 51 S. Ct. 421 (1932) are just three of many Supreme Court
decisions that have analyzed patent pools, but they explain much about the competitive
concerns these private institutions can raise. The facts that gave rise to these suits show
two ways that a patent pool could theoretically impose social costs by diminishing
competition: firstly, by limiting the number of products licensees are permitted to make
or the prices they may charge (the concern in National Harrow), a patent pool can limit
competition in the market for the product; secondly, by aggregating patents that cover
competitive (i.e., substitutive) technologies as the defendants in Standard Oil did, a
pool can theoretically limit competition in the downstream market for that technology.4
(This latter concern is sometimes referred to as the “lost substitutes” concern in the
literature.) Beyond these insights, the decisions reveal an interesting evolution in judicial
thinking—from unbridled permissiveness to condemnation, to a middle ground in
which judges, following Brandeis’ lead perhaps, voice appreciation for the theoretical
social harm that patent pooling may cause, but reject categorical assumptions. In fact,
as Brandeis acknowledged in Standard Oil, patent pools can encourage competition in
some settings.
Despite this seemingly clear progression of thought, judges, regulators, and other experts
vacillated, pendulum-like, in their views in the years following Standard Oil. A 1955 report
on antitrust and patents written by a committee assembled by the U.S. Attorney General
explained, “In any given case, a determination of legality requires an examination of the
purpose of the interchange, the power possessed by the interchange when formed, and its
operating practices. Depending upon the situation, some of these factors, or even one of
them, may be decisive.” (Congressional Report, 1955). By the late 1960s, the Department
of Justice presumed that patent ownership was prima facie evidence of market power—a
belief that engendered a critical view of pooling. It appears that regulators at that time
gave scant consideration to countervailing efficiency justifications for patent pools. By
the mid-1980s, regulators’ attitudes had warmed again, however. Carpet Seaming Tape
Licensing Corp. v. Best Seam, Inc., 694 F.2d 570 (9th Cir.1982). In 1985, the Chief of
the Antitrust Division of the US Dept. of Justice published an academic article arguing
strongly against per se regulatory disapproval of pools (Andewelt, 1985).
The most significant era for patent pooling that followed was the 1990s. At that time,
many patent pools arose to facilitate the use of technological standards—that is, protocols
that embodied patented methods and processes (Contreras, 2019). Patent pools covering
audio and video encoding and telecommunications protocols have been central to the
commercialization of widely used consumer electronics products, including smartphones,
music players, DVDs, Bluetooth-enabled speakers, and Wi-Fi.5 In the late 1990s, the
4
Because many gasoline cracking processes existed outside of the patent pool, Brandeis
determined that this harm had not occurred in Standard Oil, 51 S. Ct. 421 (1931).
5
The patent pools that permit licensing of these technological typically began as direct out-
growths of the standard-setting efforts that defined them. (Contreras, 2019). According to Merges
and Mattioli (2016), a standard practice has been for contributors to a standard to issue a public
request for patents potentially relevant to the standard; then, often working with a private company

Antitrust Division of the Department of Justice published a trio of advisory letters

endorsing patent pools organized to facilitate the licensing of technological standards,
including MPEG-2 (see “Administrative materials”). This generally permissive attitude
carries on at the time of this writing.
IV. ASSESSING THE IMPACT OF PATENT POOLS ON

COMPETITION AND INNOVATION
Because competition and innovation are difficult to measure, scholars of patent pools
have searched for useful proxies. These have included both qualitative and quantitative
information on the relationships between patents in pools, characteristics of patent pool-
ing agreements, market and industry data, and financial data reported in court decisions
of patent infringement cases. Very recently, charts and graphics generated with the aid of
“Big Data” have provided new and useful windows into the relationship between pool-
ing, competition, and innovation. Ultimately, however, most scholarship addressing the
competitive effects of patent pools is based more on theory than empirical observations.
Studies in this area sometimes seem to draw contradictory conclusions—a fact that sug-
gests it is a topic of inquiry still under development.
A theory tacitly woven into Vaughan’s work is that the number of patent pools in opera-
tion at any time is, in part, a reflection of shifting regulatory attitudes. More regulation
discourages pooling, and vice versa. Fusing legal analysis and historical research, Vaughan
(1956) told the history of patent pooling through the lens of antitrust regulation, dividing
the historical record into eras of alternating regulatory permissiveness and restrictiveness.
According to Vaughan, these periods reflected the shifting political climate. Vaughan
noted that the passage of the Sherman and Clayton Acts, for example, marked the begin-
ning of a period in which patent pools were more heavily scrutinized than they had been
in earlier times. He explained that regulators during this period were concerned with a
specific kind of competitive threat: the power of patent pools to restrict output—that is,
the number of machines licensees could sell or the prices they could charge consumers.
Vaughan also reported that a patent pool covering motion picture technology also raised
concerns of illegal “tying” at the time.
Recent scholarship challenges Vaughan’s thesis somewhat. Vaughan identified three
distinct phases of regulatory attitudes toward patent pools: an initial period of permis-
siveness (1856–1909), a subsequent period of regulatory disapproval of pools that
retrained trade (1909–42), and a third period of qualifiedly renewed permissiveness
following 1942. Some patent pools flourished during the 1930s, however—a period
Vaughan describes as one of regulatory strictness (Mattioli, 2014). Likewise, two of the
largest and most influential patent pools—the Automobile Manufacturers Association
and the Manufacturers’ Aircraft Association (MAA)—formed and thrived during the era
Vaughan defines as restrictive (Mattioli, 2014).
that specializes in administering patent pools, the various patent holders will submit the patents
they believe to be essential to the standard to a neutral expert for evaluation; finally, the members
of the burgeoning patent pool will meet to negotiate terms.

Elsewhere, Vaughan’s work occasionally demonstrates the limitations of the materials he

had at his disposal (mostly court decisions). A passage describing the MAA, an important
patent pool formed just before America’s involvement in the first world war, is illustrative:
The airplane pool, like the automobile or any other, has avoided conflict of technology and
therefore litigation among the members. This was especially important at its beginning in 1917.
However, this very harmony—pooling all patents of members and giving each the right to use
the inventions of the other—took away each member’s incentive for basic inventions either
through its own research or through purchase from an outsider. In fact, such a group generally
opposes any innovation that would depreciate what they already have. Revolutionary changes
in aviation have come from outside the pool—for example, the jet engine from an inventor in
another country . . . (Vaughan, 1956).
More recent scholarship based directly upon the MAA pooling agreement (which may
have been unavailable to Vaughan) shows that the MAA may well have encouraged inno-
vation by offering pecuniary incentives to members who made significant technological
advances (Mattioli, 2014).
Steven Carlson (1999) also explored patent pools’ potentially negative impact on com-
petition and innovation. Insightful and valuable, Carlson’s critiques were rooted heavily in
theory. He began by cataloging a litany of competitive threats that patent pools may pose.
He posited, for instance, that some patent pools may, by sheltering invalid patents, act as
staging grounds for anti-competitive collusion among would-be competitors. Carlson also
explored the possibility that patent pools could be used to create proprietary standards
that advance anti-competitive goals. As he explained, “Patent pools, however, permit
holders of intellectual property to establish their technology as proprietary standards.”
Carlson (1999) cleverly supported his arguments by citing the lack of empirical evidence
to support the contrary view—in essence, flipping the burden of proof. He did so by
challenging the widely held belief that patent pools solve a real problem:
[I]t is not clear how thoroughly technology is suppressed by overlapping patent claims . . . There
certainly are high-profile instances where patent rights do block the development of important
technologies. Often, however, patented goods are sold competitively, notwithstanding the legal
entanglements created by blocking patents. The DOJ and the FTC have announced a broad
shelter for the pooling of patents that, if unconstrained, may invite a revival of the abusive patent
pools that dominated American industry throughout much of this century.
Carlson argued that it is inappropriate for regulators to accept the threat to competition
posed by patent pools, in light of the absence of any empirical evidence that pools truly
overcome transactional deadlocks. Importantly, Carlson did not argue that the absence of
evidence concerning the benefits of pools amounted to evidence of an absence; instead, he
argued that the burden of proof should fall on those that wish to justify pools, rather than
those that would regulate them strictly. This observation generally accords with a vein of
scholarship that doubts the existence of an “anti-commons” that would necessitate patent
pooling (Adelman, 2005).
Supreme Court jurisprudence also offers some helpful methodologies for estimating the
competitive harm posed by patent pools. As discussed earlier, a chief concern of antitrust
regulators is that pools permit the bundling of patent rights that are substitutes—that
is, rights that would ordinarily compete in the licensing market absent the pool. The

Supreme Court looked to market data when it analyzed this issue in a trio of decisions:
Standard Oil, 51 S. Ct. 421 (1931), Hartford-Empire, 323 U.S. 386 (1945), and Line
Materials, 333 U.S. 287 (1948). As mentioned earlier, the Court set forth in Standard Oil
the rule that a patent pool threatens competition only if the patents involved “dominate”
an industry. Looking to market concentration data showing that the licensees of the pool
constituted only 55 percent of the industry, the Court concluded that the pool did not
threaten competition. “This development . . . is clear evidence,” the majority wrote, “that
[the pool] did not concentrate in the hands of the four primary defendants the licensing
of patented processes for the production of cracked gasoline.” Standard Oil, 51 S. Ct.
421 (1931). In Line Materials, by contrast, the Court examined the claims of patents
themselves to ascertain whether a specific pool contained substitutes or complements.
In the 1990s, as patent pools formed in the consumer electronics industry, scholars
focused on a new concern: the possibility that these groups may reduce the level of innova-
tion within industries. On the surface, this conclusion may seem a bit counterintuitive.
By facilitating the licensing of complementary technologies, patent pools are widely
thought to facilitate not only commercialization but also basic research (Heller and
Eisenberg, 1999). And, as recent scholarship mentioned earlier in this chapter indicates,
some evidence suggests that pools have promoted technological advances (Shapiro, 2000;
Flamm, 2012).
How, then, might pools reduce the rate of innovation? Theorists believe the problem
has to do with special licensing provisions called grant-backs.6 As respected economist
Richard Gilbert (2004) has explained, some patent pools have required their members to
license-back to the pool after-arising patents that cover competing technologies. In such
cases, Gilbert explains, “[p]atent pools can harm consumers by reducing incentives to
innovate.” The DOJ describes the “grant-back” concern as follows:
An important factor in the Agencies’ analysis of a grant-back will be whether the licensor has
market power in a relevant technology or innovation market. If the Agencies determine that a
particular grant-back provision is likely to reduce significantly licensees’ incentives to invest in
improving the licensed technology, the Agencies will consider the extent to which the grant-back
provision has offsetting procompetitive effects, such as (1) promoting dissemination of licensees’
improvements to the licensed technology, (2) increasing the licensors’ incentives to disseminate
the licensed technology, or (3) otherwise increasing competition and output in a relevant technol-
ogy or innovation market. (U.S. Department of Justice, 2007).
6
(Goter, 2011) “A more difficult question arises when future competition (and consumer
choice) between a current technology and a nascent technology is restrained. In this case, the court
must adjudge where on the continuum between ‘certainly would have been viable’ and ‘certainly
could not have been viable’ the nascent technology lies. The party asserting misuse should be
required to demonstrate a reasonable probability that the product or technology would become
commercially viable or technically feasible in the absence of the challenged restraint”; (Iden, 2011)
“The problem is, the individual ‘injuries’ that joint ventures in technology suppression tend to cause
are, indeed, more speculative and difficult to quantify: the elimination of a firm’s chance to develop
a technology into a viable commercial alternative or the public’s chance to enjoy it. Frustratingly,
the advantages of antitrust litigation—keeping out meritless claims—also serve to bar those
plaintiffs that would be aggrieved by a harmful joint venture suppressing technology”; (Bohannan,
2011) noting a potential rival producer’s difficulty in showing a technology “would have come to
fruition and would have become commercially successful but for the IP holder’s restraint.”

As the theory goes, a grant-back provision could alter innovation incentives for any
company it applies to. A related theory holds that patent holders that enjoy the benefits
of membership in a pool—that is, steady royalties and decreased litigation costs—have
fewer financial incentives to invest in new patents (Joshi and Nerkar, 2011).
Scholars have had no easy way to verify these theories empirically. There are two rea-
sons for this: first, innovation itself is an imprecisely defined term and is thus inherently
difficult to quantify and measure; second, any causal relationship that may exist between
certain licensing behavior—for example, the use of grant-backs—and the level of innova-
tion within an industry is naturally complex. That is, even if experts agreed on a common
definition of “innovation” and also agreed on how best to measure it, as an empirical
matter, it would be difficult to control for the many factors that may determine the level of
innovation. This helps explain why, as Petra Moser (2015) has noted, “Empirical evidence
on the effects of modern pools on innovation is limited so far.”
Despite these challenges, scholars have pressed ahead by searching for proxies for
innovation. One such proxy is the level of patenting behavior within an industry. In a 2011
research paper, Amol M. Joshi and Atul Nerkar (2011) worked to answer two questions:
first, does the formation of patent pools enhance or inhibit firm-level innovation; second,
are the effects on innovation the same of licensor and licensee firms? To answer these
questions, the researchers first identified all patents included within the contemporary
patent pool that governs the DVD video standard. The researchers then identified the
top ten technology classes these patents fell into and analyzed the patenting behavior of
the most active patentee-assignees in those technology fields from the years spanning
1976 through 2006—a time period that covered the ascendance of DVDs. From this, they
identified a set of companies that were actively patenting inventions related to optical
disc technology. Within that group, they observed a drop in the level of patenting by
firms that participated in the patent pool. This observation led them to conclude that the
DVD patent pool had slowed the quantity and quality of innovation among its members.
Notably, Joshi and Nerkar drew this conclusion, in part, by assessing the quantity and
quality (as measured by patent citations) of patents granted to the firms in question
during a time period just prior to the growth of optical disc technologies.
As rigorous as this quantitative approach appears, qualitative evidence has led leading
scholars to draw opposing conclusions. In 2000, Carl Shaprio (2000) observed, “[t]he
impressive rate of innovation in the semiconductor industry in the presence of a web
of such cross licenses”—a blanket term within which he included pools—“offers direct
empirical support for the view that these cross-licenses promote rather than stifle innova-
tion.” Shapiro’s conclusion would seem to apply to patent pools that are composed of
webs of cross-licenses. Kenneth Flamm (2012) looked to online advertisements for CD
drives to determine the rate of innovation in that particular field. Specifically, Flamm
drew upon “primarily product reviews in online computer publications. Approximately
198 observations on price and characteristics for . . . models sold as retail products in U.S.
markets were harvested on the Web over the years 2000 through 2011.” Relying on these
qualitative measures, Flamm concluded that innovation increased as a result of a 1980s
patent pool for CDs.
What are we to make of these seemingly contradictory conclusions? The fact that Joshi
and Nerkar (2011) and Flamm (2012) drew different conclusions about the impact of
pooling in the same field of technology—optical discs—might reveal that the innovation

measures they looked to captured different things. Flamm studied advertisements and
product reviews and discovered increases in the rates of read and write speeds to optical
disc drives, along with decreased costs. These trends could be the result not of patentable
technological advances, but rather competition in the market for optical drives. That is,
one might expect that a set of manufacturers that all enjoy the same permissions would
find new ways to distinguish their products in terms of price performance characteristics
such as speed. Such distinctions have not always been the results of patentable inventions.
Thus, a decrease in patenting such as that measured by Nerkar and Joshi would not
be contradictory. As Flamm notes, “introductions of qualitatively new products have
declined over time.”
Lerner et al. (2005) observed an important new correlation between the composition
of patent pools and their licensing rules: they observed that patent pools composed of
complementary patents (as opposed to substitutive patents) are more likely to permit their
members to independently license their rights outside of the pool, and also are more likely
to require licensees to grant back to the pools patents of their own. This observation is
relevant to antitrust concerns about the role that such grant-back clauses may play in reduc-
ing incentives for innovation among pool contributors. In essence, Lerner provided a set of
correlations between licensing rules and pool composition that could be useful to regulators.
Petra Moser (2013) explained that looking to patenting rates is a limited method
of measuring the impact of patent pools on innovation. “Patent data,” she wrote
“may, however, fail to capture innovation that occurs outside the patent system—for
example, in countries without patent laws or in industries in which inventors rely on
alternative mechanisms to protect their intellectual property.” This insight led Moser (in
a paper co-authored with Ryan Lampe) (2010) to search for new qualitative sources of
information on innovation in the sewing machine industry during the years the sewing
machine patent pool operated. These sources included “[a]rticles on sewing machines in
nineteenth-century magazines, such as Scientific America and the Ladies’ Home Journal,”
which revealed what the authors believed were “key characteristics consumers valued,”
including “improvements in sewing machine speed” (Moser, 2013). These performance
characteristics are similar in some ways to the speeds of optical disc drives that Flamm
considered. On balance, the authors concluded that “improvements slowed soon after the
pool had been established and did not recover until it had dissolved.”
Robert Merges and Michael Mattioli (2017) presented novel methods to estimate wel-
fare losses generated both by lost substitutes and by grant-back terms. Their approach to
the lost substitutes issue was inspired by a new insight: including substitutive technologies
within a patent pool is, from a market competition perspective, the reverse of infringing
upon a patent owner’s exclusive rights. As the authors explain:
patent infringement cases are about actual duopolists who by rights should have been monopo-
lists; whereas ‘lost substitute’ analysis from patent pools is about actual monopolists (the pool
members) who should have been duopolists (competitors), because they each owned patents on
rival substitute technologies.
This insight is valuable because it shows that regulators may look to patent infringement
cases for data that permits an estimate, in dollar values, of the market impact of a patent
pool in the same industry that includes substitutes. Turning to innovation losses caused by
grant-backs, the authors turned to another new source: newly available visual “maps” that

present detailed patenting behavior of companies. These maps are the product of recent
advances in “Big Data” that draw upon natural language search capabilities to show
precisely what kinds of technologies specific companies devote research and development
efforts to. Using these new tools, the authors explain, regulators can estimate the degree
of overlap between two companies involved in a patent pool. From there, researchers
can develop a baseline estimate of the possibility that future research options will be
foreclosed by a grant-back requirement.
V. ASSESSING THE IMPACT OF PATENT POOLS ON PATENT

LITIGATION
The relationship between patent pooling and patent litigation is more complex than
one might guess. Because many patent pools form in order to settle litigation, it seems
natural to assume that the very existence of a patent pool serves as prima facie evidence
of private cooperation triumphing over costly disagreements. Justice Brandeis said it
more eloquently in Standard Oil (1983): “An interchange of patent rights and a division
of royalties according to the value attributed by the parties to their respective patent
claims is frequently necessary if technical advancement is not to be blocked by threatened
litigation.” This view does not address the possibility, however, that a patent pool may
sometimes impede socially desirable litigation, such as validity challenges to patents that
should not have been issued. As a result, researchers seeking to understand the costs
patent pools impose on society, or alternatively, the benefits they bestow, vis-à-vis litiga-
tion, have looked beyond raw quantitative data on litigation rates to qualitative sources.
It is helpful to begin by considering more carefully the optimistic view that patent pool-
ing can dampen undesirable litigation. There are two costs that easily might stem from
patent litigation: first, patent holders must fund litigation with capital that could other-
wise fuel research; second, because companies in an adversarial relationship are unlikely
to cooperate (e.g., pursuing joint ventures and the like), the dissemination of technology
and related technological information—arguably a central goal of the patent system—is
impeded. By eliminating litigation between members, a patent pool could cancel both
costs out. As Carl Shapiro (2001) has noted, “The creation of a pool may also encourage
investments in research and development by reducing litigation risks for members and
thereby increasing expected profits from research and development . . .” A 1931 note in
the Yale Law Journal similarly observes, “The patent pool is of great economic utility in
relieving this so-called ‘congestion’ of patents, the expenses of which are shifted to the
consumer.” (Yale Law, 1931).
The historical record of patent pools offers many examples that directly support this
theory through evidence of patent pools overcoming litigation. In his exploration of
the sewing machine patent pool, for instance, Adam Mossoff (2009) has written, “The
sewing machine was the result of numerous incremental and complementary incentive
contributions, which led to a morass of patent infringement litigation given overlapping
patent claims to the final commercial products.” Arti Rai (1999) has similarly noted, “The
two most historically prominent pools arose in the automobile and aircraft industries. In
both of these industries, patent pools formed only after significant litigation and refusal
to license.”

A reduction in the rate of litigation may be a sign of undesirable behavior, however.

Theorists have posited that a patent pool sheltering weak patents from validity chal-
lenges would reduce the rate of patent litigation at the expense of useful attacks.
Commenting generally on patent pooling in the course of his historical accounts,
Vaughan (1956) commented, “since the members agree to accept each other’s patents
without question, there has been practically no occasion to adjudicate their validity and
therefore determine the extent to which they monopolize ideas which rightfully belong
to the public.”
Robert Merges (1999) has provided a more hopeful perspective on this issue. In the
course of examining the MAA membership agreement, Merges noted that the pool
included an internal procedure for settling disputes and challenging patent validity.
This suggests that perhaps even if a patent pool directly reduces the number of validity
challenges in court, the pool is not necessarily sheltering low-quality patents. Many pools
may include procedures and norms, formal and informal, that resolve disputes between
members and weed away low-quality patents. On a deeper level, this finding urges away
from quickly drawing conclusions from litigation data alone.
Some theorists have stepped away from the question of whether a reduction in litigation
caused by a pool is normatively good or bad, and have instead focused on what effect a
pool’s reduction in litigation might have on the incentives of prospective members. Jay
Pil Choi (2003) has argued that reduced litigation may serve as an incentive to join a
patent pool. Lampe and Moser’s paper (2010) drew a consistent conclusion with an added
wrinkle—that some patent pools may not only reduce their members’ risks of litigation,
but also raise litigation risks for outsiders. Drawing upon litigation data from the time
of the sewing machine patent pool, they commented, “For example, the sewing machine
pool appears to have exacerbated litigation risks for outside firms, even as it reduced such
risks for members.”
VI. ASSESSING THE IMPACT OF PATENT POOLS ON

TRANSACTION COSTS
Merges and Mattioli (2017) conducted the first empirical study that attempted to measure
the transaction cost savings that patent pools provide. The study was motivated by the
fact that previous scholarship had focused primarily on the potential anti-competitive
effects of patent pools (i.e., social welfare losses) without addressing the potential upside.
From a high level, their methodology was straightforward: they sought to learn the cost
of establishing and operating a patent pool by interviewing pool administrators, and then
they subtracted this number from the transaction costs that they argued would persist in
the absence of the pool (i.e., a web of two-party licenses).
The first number encompasses many costs related to setting up and operating a patent
pool—for example, setup costs incurred by founding members of a pool, setup costs
incurred by a pool administrator, ongoing operational costs for licensees and licensors,
etc. Merges and Mattioli (2017) obtained this number through a set of semi-structured
interviews with the administrators of prominent patent pools. This research allowed them
to estimate the total costs that licensors, licensees, and administrators incur in a typical
patent pool covering a technological standard.

To obtain the second figure—that is, the costs incurred in the absence of a pool—they
relied upon a formula that takes as “inputs” the number of patent holders and the
number of licensees in a given scenario and computes from this the number of individual
cross-licenses that would be necessary in the absence of a pool. (The authors took pains
to explain why they believed that a web of two-party licensing transactions was the “next
best thing” to a patent pool, and they also acknowledged that, under some hypothetical
scenarios, other optimal arrangements could be optimal.) Merges and Mattioli estimated
the cost of each license by calling upon a conventional proxy: the average cost of patent
litigation. Finally, they subtracted the cost of setting up and operating a pool from the
licensing costs in the absence of a pool to arrive at an estimate of the net transaction
costs conserved. The results of this study indicated that patent pools provide substantial
transaction cost savings—on the order of hundreds of millions of dollars, conservatively.
In light of this, the authors argued that policymakers should begin with transaction cost
savings as a baseline for comparison when examining the desirability of a specific patent
pool. This amount can then be compared to an estimate of the anti-competitive effects of
a pool (Merges and Mattioli, 2017). In a follow-on, Mattioli explored the role of “outsid-
ers” on patent pools—that is, patent holders that strategically decline to join, or defect
from patent pools (Mattioli, 2018). The article, which draws largely upon interviews and
financial and patent data shared by industry participants, responds to a growing theory
that, by operating in the shadows of pools, such outsiders have the power to undermine
and offset the transaction costs that pools can conserve.
VII. PATENT POOLS AS FORUMS FOR VALUATION
An important theoretical question about patent pools is whether they solve transactional
deadlock more effectively than the state might be able to—for example, through a system
of compulsory licensing (Carlson, 1999; Merges, 1999; Mattioli, 2011). This question is
motivated by the same fundamental concern that has motivated other avenues of patent
pooling scholarship: whether the benefits of patent pools overall outweigh their theoreti-
cal potential for harm.
Merges (1999) concluded that many patent pools are efficient forums for valuation of
patents—and indeed may perform valuation more effectively than the state could because
they reflect the collective judgment of patent holders themselves. He explained that this is
one of the first things patent pools do after they form.
Soon after coming together, one of the first things they do is settle two issues of valuation: the
rates licensees will pay for access to the entire pool; and rules for dividing the spoils among the
pool’s members. In the case of licensing rates, the price(s) are the same for all takers, or at least
for all licensees of similar size. Through their collectively-determined prices, these institutions
operate in much the same way as a compulsory license. Essentially all comers are welcome to use
the right(s) so long as they pay the pre-established price.
This led Merges to an important insight: that the varying structures, governance rules, and
compositions of patent pools reflect how the needs of specific industries differ from the
patent rights originally apportioned by the state. In Merges’ view, patent pools can thus
be understood as “contractual substitutes for statutory entitlements.” This observation

situated patent pooling within Entitlements Theory—an insight that would prompt future
studies of how effectively patent pools value patent rights as compared to organs of the
state (Merges, 1999; Mattioli, 2014).
Drawing upon original membership agreements from a large set of patent pools,
Mattioli (2014) provided some counterexamples that supported the view that patent
pools do not always serve as ideal democratic forums for patent valuation. The article
conducted an original empirical study of patent licensing agreements that governed
collectives between the years 1856 and 2011. These agreements revealed how royalty rates
were established and how profits were shared among participants. The study revealed
that some of the largest and most notable patent licensing collectives were designed to
accommodate new patents and members. Many early patent licensing groups, however,
provided for the inclusion of new patents, but not new members. By and large, modern
collectives were more accommodating than historical groups in this regard. The predic-
tion that patent licensing collectives would fine-tune royalty distributions to match the
economic value of contributions was met only in a few cases. Many patent pools instead
adopted “rough and ready” approaches, such as pro-rata distributions based on patent
ownership, simple per-capita distributions, or mutual royalty-free licensing. The observa-
tions suggest that it is often too difficult or too costly to determine the economic value of
individual patents within a set. Turning to the theme of representation, the agreements
show that most collectives provide patent holders with an indirect form of control over
pricing decisions, either in the form of corporate voting shares, or through representation
in complex valuation procedures. There are, however, very few instances of direct voting
over decisions that influence pricing.
VIII. CONCLUSION
Despite the diverse industries and technologies that patent pooling has touched since the
1850s, it is remarkable how unchanged (and largely unresolved) the policy concerns they
have raised are. When is cooperation among patent holders socially productive? Under
what circumstances can a patent pool harm competition? Are patent pools sometimes the
only way patent holders can extract the full value of the rights they hold? Should the gov-
ernment take affirmative steps to either encourage or discourage patent pools, and if so,
what forms of intervention are most appropriate? Such questions are difficult to answer,
in part because patent pools are not required or encouraged to share useful information
about their operations with scholars. As this chapter has explained, great scholarly effort
has gone into simply assembling a list of known patent pools; immense creativity too has
been directed toward understanding the impact these groups have had on competition,
transaction costs, and innovation. As with technology standards, mountains of useful
information are no doubt locked away within the private firms that contribute to and
license from patent pools. Policymakers would do well to explore possible incentives the
government could offer to encourage the disclosure of such information. Until then,
scholars must continue to imagine and employ new and creative ways to investigate this
important and complex phenomenon.

REFERENCES
Agreement between Elias Howe, Jr., the Wheeler & Wilson Mfg. Co., the Grover & Baker Sewing Machine Co.,
and I. M. Singer & Co. (Oct. 24, 1856) (on file with author).
Antitrust Division, U.S. Department of Justice, “Antitrust Enforcement and Intellectual Property Rights:
Promoting Innovation and Competition” (April, 2007), available at www.justice.gov/atr/antitrust-enforcement-
and-intellectual-property-rights-promoting-innovation-and-competition.
Barnett, Jonathan M., 2009. “Property as Process: How Innovation Markets Select Innovation Regimes”, 119
Baron, Justus and Tim Pohlmanm, “The Effect of Patent Pools on Patenting and Innovation—Evidence from
Contemporary Technology Standards” (Northwestern Univ. Working Paper, 2015), available at www.law.
northwestern.edu/research faculty/searlecenter/innovationeconomics/documents/Baron_Pohlmann_effect_
of_patents.pdf.
Bohannan, Christina, 2011. “IP Misuse as Foreclosure”, 96 Iowa Law Review 475–528.
Bowman, Ward S., Jr., 1973. Patent and Antitrust Law: A Legal and Economic Appraisal (University of Chicago
Press).
Carlson, Steven C., 1999. “Patent Pools and the Antitrust Dilemma”, 16 Yale Journal on Regulation 359–99.
Choi, Jay Pil, 2003. “Patent Pools and Cross-Licensing in the Shadow of Patent Litigation”, Cesifo Working
Paper No. 1070, Category 9: Industrial Organization, November, available at www.cesifo.de.
Comment, 1931. “Patent Pools and the Sherman Act”, 40 Yale Law Journal 1297–303.
Cotter, Thomas F., 2009. “Patent Holdup, Patent Remedies, and Antitrust Responses”, 34 Journal of Corporate
Law 1151–207.
Cournot, Augustin, 1838. Researches into the Mathematical Principles of the Theory of Wealth (Nathaniel T.
Bacon trans., Augustus M. Kelley 2d ed. 1971).
Flamm, Kenneth, 2012. “A Tale of Two Standards: Patent Pools and Innovation in the Optical Disk Drive
Industry”, Working Paper, available at http://conference.nber.org/confer/2012/IPKE/flamm.pdf.
Gilbert, Richard J., 2004. “Antitrust for Patent Pools: A Century of Policy Evolution”, 2004 Stanford Technology
Law Review 3–51.
Goter, Phillip W., 2011. “Princo, Patent Pools, and the Risk of Foreclosure: A Framework for Assessing
Misuse”, 96 Iowa Law Review 699–735.
Heller, Michael A., 1998. “The Tragedy of the Anticommons: Property in the Transition from Marx to
Markets”, 111 Harvard Law Review 621–88.
Heller, Michael A. and Rebecca S. Eisenberg, 1998. “Can Patents Deter Innovation? The Anticommons in
Biomedical Research”, 280 Science 698–701.
Heller, Michael A. and Rick Hills, 2008. “Land Assembly Districts”, 121 Harvard Law Review 1465–527.
Hess, Charlotte and Elinor Ostrom, 2003. “Ideas, Artifacts, and Facilities: Information as a Common-Pool
Resource”, 66 Law and Contemporary Problems 111–45.
Hovenkamp, Erik and Herbert Hovenkamp, 2017. “Patent Pools and Related Technology Sharing”, in Rodger
D. Blair and D. Daniel Sokol, eds., The Cambridge Handbook of Antitrust, Intellectual Property, and High
Tech (Cambridge University Press).
Iden, Daniel J., 2011. “Combating Joint Ventures in Suppression: Taking Inventory of the Legal Arsenal”, 96
Joshi, Amol M. and Atul Nerkar, 2011. “When do Strategic Alliances Inhibit Innovation by Firms? Evidence
from Patent Pools in the Global Optical Disc Industry”, 32 Strategic Management Journal 1139–60.
Klitzke, Ramon A., 1985. “Patents and Monopolization: The Role of Patents Under Section Two of the
Sherman Act”, 68 Marquette Law Review 557–96.
Lampe, Ryan and Petra Moser, 2010. “Do Patent Pools Encourage Innovation? Evidence from the Nineteenth-
Century Sewing Machine Industry”, 70 Journal of Economic History 898–920.
Layne-Farrar, Anne and Josh Lerner, 2011. “To Join or Not to Join: Examining Patent Pool Participation and
Rent Sharing Rules”, 29 International Journal of Industrial Organization 294–303.
Lemley, Mark A. and Carl Shapiro, 2013. “A Simple Approach to Setting Reasonable Royalties for Standard-
Essential Patents”, 28 Berkeley Technology Law Journal 1135–66.
Lerner, Josh et al., 2007. “The Design of Patent Pools: The Determinants of Licensing Rules”, 38 RAND Journal
Letter from Joel I. Klein, Acting Assistant Attorney Gen., DOJ, to Gerrard R. Beeney, Esquire, Sullivan &
Cromwell (Jun. 26, 1997), available at www.usdoj.gov/atr/public/busreview/215742.pdf.
Letter from Joel I. Klein, Assistant Attorney General, Antitrust Div. of the U.S. Dep’t of Justice to Gerrard
R. Beeney, Sullivan & Cromwell (Jun. 26, 1997), available at www.justice.go v/archive/atr/public/busre-
view/215742.pdf.
Letter from Joel I. Klein, Assistant Attorney General, Antitrust Div. of the U.S. Dep’t of Justice to

Gerrard R. Beeney, Sullivan & Cromwell (Dec. 16, 1998), available at www.justice.gov/sites/default/files/atr/
legacy/2006/04/27/2121.pdf.
Letter from Joel I. Klein, Assistant Attorney General, Antitrust Div. of the U.S. Dep’t of Justice to Carey R.
Ramos of Paul, Weiss, Rifkind, Wharton & Garrison (Jun. 10, 1999), available at www.justice.gov/sites/
default/files/atr/legacy/2012/08/01/2485.pdf.
Lerner, Josh and Jean Tirole, 2004. “Efficient Patent Pools”, 94 American Economic Review 691–711.
Madison, Michael J., Brett M. Frischmann, and Katherine J. Strandburg, 2010. “Constructing Commons in the
Cultural Environment”, 95 Cornell Law Review 657–710.
Mattioli, Michael, 2012. “Communities of Innovation”, 106 Northwestern University Law Review 103–156.
Mattioli, Michael, 2014. “Power and Governance in Patent Pools”, 27 Harvard Journal of Law and Technology
421–65.
Mattioli, Michael, 2018. “Patent Pool Outsiders”, 33:1 Berkeley Technology Law Journal (forthcoming).
Merges, Robert P., 1996. “Contracting into Liability Rules: Intellectual Property Rights and Collective Rights
Organizations”, 84 California Law Review 1293–393.
Merges, Robert P., 2000. “Institutions for Intellectual Property Transactions: The Case of Patent Pools”, in
Rochelle Dreyfuss, Diane L. Zimmerman, and Harry First, eds., Expanding the Boundaries of Intellectual
Property: Innovation Policy for the Knowledge Society, available at www.law.berkeley.edu/files/pools.pdf.
Merges, Robert P. and Michael Mattioli, 2017. “Measuring the Costs and Benefits of Patent Pools”, 78 Ohio
State Law Journal 281.
Merges, Robert P. and Richard R. Nelson, 1990. “On the Complex Economics of Patent Scope”, 90 Columbia
Misa, Thomas J., 1999. A Nation of Steel (Johns Hopkins University Press, 2d ed.).
Moser, Petra, 2013. “Patents and Innovation: Evidence from Economic History”, 27 Journal of Economic
Mossoff, Adam, 2011. “The Rise and Fall of the First American Patent Thicket: The Sewing Machine War of
the 1850s”, 53 Arizona Law Review 165–211.
Ostrom, Elinor, 1990. Governing the Commons: The Evolution of Institutions for Collective Action (Cambridge
University Press).
Rai, Arti, 1999. “Regulating Scientific Research: Intellectual Property Rights and the Norms of Science”, 94
Northwestern University Law Review 77–152.
Searle Center Database, Data on Technology Standards, Industry Consortia, and Innovation, www.law.
northwestern.edu/research-faculty/searlecenter/innovationeconomics/data/technologystandards.
Shapiro, Carl, 2000. “Navigating the Patent Thicket: Cross Licenses, Patent Pools, and Standard Setting”, 1
Innovation Policy and the Economy 119–50.
Simon, Brenda and Ted Sichelman, 2017. “Data-Generating Patents”, 111 Northwestern University Law
377–438.
Stedman, John C., 1964. “The U.S. Patent System and Its Current Problems”, 42 Texas Law Review 450–97.
Strandburg, Katherine J., 2011. “Patent Fair Use 2.0”, 1 U.C. Irvine Law Review 265–306.
Vaughan, Floyd L., 1925. The Economics of Our Patent System (The Macmillan Co.).
Vaughan, Floyd L., 1956. The United States Patent System (Univ. of Oklahoma Press).
Ziedonis, Rosemarie Ham, 2004. “Don’t Fence Me In: Fragmented Markets for Technology and the Patent
Acquisition Strategies of Firms”, 50 Management Science 804–20.
Case List
Carpet Seaming Tape Licensing Corp. v. Best Seam, Inc. (1982), 694 F.2d 570.
E. Bement & Sons v. Nat’l Harrow Company (1902), 186 U.S. 70.
Hartford-Empire Co. v. U.S. (1945), 323 U.S. 386.
Standard Oil Co. (Indiana) v. United States (1931), 51 S. Ct. 421.
Standard Sanitary Mfg. Co. v. United States (1912), 226 U.S. 20.
United States v. Hartford-Empire Co. (1942), 46 F. Supp. 541.
United States v. Line Materials Co. (1948), 333 U.S. 287.
A Study of the Antitrust Laws of the United States, and their Administration, Interpretation, and Effect
Pursuant to S. Res. 61: Hearings Before the Subcommittee on Antitrust and Monopoly of the Committee on
the Judiciary, United States Senate, 84th Cong. 1st Sess. (1955).
Clayton Act, 38 Stat. 730 (1914) (codified at 15 U.S.C. §§ 12, 13, 14–21, 22–27 (1988)).

Compulsory Licensing of Patents: Hearing on H.R. 9259, H.R. 9815, and H.R. 1666 Before the Subcomm. on
Compulsory Licensing of the H. Comm. on Patents, 75th Cong. 38 (1938).
Investigation of Concentration of Economic Power: Hearings Pursuant to Public Resolution No. 113 Before
the Temp. Nat’l Econ. Comm., 75th Cong. 302 (1938).
Pooling of Patents: Hearings on H.R. 4523 Before the H. Comm. on Patents, 74th Cong. (1935).
Sherman Act, 647, 26 Stat. 209 (1890) (codified at 15 U.S.C. §§1–7 (2014)).

11. Empirical analyses related to university patenting
Arvids A. Ziedonis* 7
Contents
I. Introduction
II. A Brief History of U.S. University Patent Policies, 1925–80
III. University Patenting, 1925–80
IV. The Bayh–Dole Act of 1980
V. University Software Patenting and “Hold-Up” Concerns
VI. “Anti-Commons” Concerns
VII. The Market for University Patents
VIII. Conclusion and Future Directions
References
I. INTRODUCTION
Patenting by universities has seen a marked increase in the past two decades. According to
the National Science Foundation (NSF), patents issued by the United States Patent Office
(USPTO) to U.S. academic institutions more than doubled from 2293 in 1996 to 5990 in
2014, the most recent year the NSF has compiled data (National Science Board, 2016).
As a share of all patents granted, academic institutions accounted for about 2 percent in
the same time period. Patenting by universities of faculty inventions has an even longer
history in the United States, however, stretching back to the early twentieth century. For
much of that time, especially until 1980, the year of passage of the seminal Bayh–Dole
Act, the appropriateness of this activity as one of the many missions of U.S. universities
was itself a subject of debate. This chapter begins with a brief outline of this debate and
summarizes university patenting through this period. It then discusses the Bayh–Dole
Act, which facilitated patenting and licensing of federally funded university inventions.
The chapter concludes by describing the empirical research on university patenting in the
last 20 years, highlights some of the unresolved issues within this literature, and suggests
new avenues for research.
* Professor of Strategy and Innovation, Faculty of Economics and Business, KU Leuven,

ziedonis@kuleuven.be.
256

Empirical analyses related to university patenting 257
II. A BRIEF HISTORY OF U.S. UNIVERSITY PATENT

POLICIES, 1925–801
In 1907, Frederick Cottrell, a chemistry professor at the University of California,

Berkeley, received a patent for his electrostatic precipitator, an air pollution mitigation
technology for industrial smokestacks. A common practice by academic patenters at
the time was to make inventions available for public use. Cottrell, however, believed that
clearly defined intellectual property rights were necessary in order to provide firms with
the incentives to undertake the costly research, development, and marketing activities
necessary to bring an embryonic invention such as the precipitator to the market (Cottrell,
1912). He subsequently founded the Research Corporation, an organization intended to
manage not only his own patentable discoveries, but also intellectual properties emanating
from other universities and research institutions.
Public universities, particularly land-grant institutions, were among the earliest univer-
sities to seek patents for faculty inventions in the expectation that state taxpayers and the
local economy would benefit from their research. The first university-affiliated research
foundation was the Wisconsin Alumni Research Foundation (WARF). In 1924, Dr. Harry
Steenbock of the University of Wisconsin developed a method for increasing the vitamin
D content of food and drugs via the process of irradiation. Steenbock decided to patent
his findings, based partly on his desire to protect the public from unscrupulous or incom-
petent firms and from monopolization of the industry by a private patentee (Apple, 1996).
Steenbock offered to assign his patent to the University of Wisconsin, but the University
felt that the creation of a specialized office to handle patents was not worth the necessary
investment (Apple, 1989). Instead, Steenbock convinced several alumni to create WARF,
an entity affiliated with but legally separate from the University of Wisconsin that would
accept assignment of patents from University faculty, license these patents, and return
part of the proceeds to the inventor and the University.
WARF’s success and substantial licensing income influenced the development of
patent policies at other U.S. universities during the 1930s, when many institutions revised
and formalized their patent policies in response to several forces. First, the Depression’s
effects on university finances forced many institutions to seek new sources of income,
and the widely acknowledged profitability of WARF offered an attractive model. Second,
the post-World War I growth in industrially funded research at universities led some
universities to clarify or codify policies affecting ownership of faculty inventions. Land-
grant universities were the most active in this area. Purdue University, the University of
Minnesota, and Cornell University established affiliated but legally separate research
foundations that were similar to WARF during the late 1920s and early 1930s.
In 1933, the American Association for the Advancement of Science (AAAS)
Committee of Patents, Copyrights, and Trademarks investigated whether patenting
was necessary for the transfer of academic research to industry. Echoing the concern
raised earlier by Professor Cottrell, the Committee concluded that “discoveries or
inventions which are merely published and thus thrown open equally to all, unless of
1
This historical account draws heavily on Ivory Tower and Industrial Innovation by David
Mowery, Richard Nelson, Bhaven Sampat, and Arvids Ziedonis (Mowery et al. 2004).

great importance to the industry, are seldom adopted” (AAAS, 1934; p. 9), and that
“ordinarily no manufacturer or capitalist would be willing today to risk his money,
and expend time and energy in developing on a commercial scale a new product or
process without being assured that his investment in developing the invention would be
protected in some measure” (p. 10).
The Committee also supported Dr. Steenbock’s earlier view, concluding that a scientist
could not expect that sole publication of the technical details of his or her invention would
yield social benefits, because of the presence of “patent pirates” that could “wrongfully
appropriate his work” and “deny the public what he thought he gave it,” either by charging
monopoly prices or by withholding the invention from use (p. 10).
The Committee report also addressed the tension between university patenting and
the “open science” norms and institutions of academe (Merton, 1973). Objections to
patenting based on this view included the assertion “that it is unethical for scientists
or professors to patent the results of their work” (p. 8). Although the Committee did
consider the risk that patenting by university scientists would hinder communalism and
bias academic inquiry away from basic research, a concern echoed in more recent debates
over university patenting (Dasgupta and David, 1994; Henderson et al., 1998), its report
dismissed these concerns.
The debates over university patent policies in the 1930s treated medical patents as a
special case. Opposition to medical patents was widespread, based on the argument that
patents restricted the use of new discoveries and therefore had no place in the medical
community (Weiner, 1986). Opponents of medical patents also expressed concern over
perceptions of university profiteering at public expense in the field of public health
(McKusick, 1948). The AAAS Committee acknowledged the “special” nature of patents
in the field of public health, but suggested that the benefits of patenting discussed above,
particularly those associated with the use of patents as a “quality control” mechanism,
were sufficient to warrant patenting of such discoveries.
The primary motivations for patenting, therefore, were the protection of the public
interest and the preservation of academic institutions’ reputations. The prospect of licens-
ing income influenced the entry by several universities into patenting and licensing during
the 1930s and subsequently; but through much of the 1925–80 period, many academic
scientists and administrators preferred to avoid direct involvement in the management
of these patents. These tensions were reflected in the patent policies adopted by leading
pre-1940 institutional patenters and licensors.
Concerns regarding ownership of patents resulting from publicly funded research also
arose in Congress, particularly during the 1940s debate over post-war U.S. science and
technology policy. One side of the debate was represented by Senator Harley Kilgore
(D-W. Va.), who argued that the federal government should retain title to patents
resulting from federally funded research and place them in the public domain (Kevles,
1973). According to Kilgore, allowing private contractors to retain patents represented
a “giveaway” of the fruits of taxpayer-funded research to large corporations, reinforcing
the concentration of technological and economic power. A major proponent of the
opposite viewpoint was Vannevar Bush, the Director of the wartime Office of Scientific
Research and Development, who argued that allowing contractors to retain patent rights
would preserve their incentives to participate in federal R&D projects and to develop
commercially useful products based on government-funded research.

The post-war debate highlighted the central questions in government patent policy
toward federally funded research for the next three decades. Supporters of the retention
of intellectual property rights by government agencies argued that allowing contractors
(rather than government agencies) to retain title to patents resulting from federally funded
research favored large firms at the expense of small business. Moreover, they asserted that
such a policy would raise prices for the fruits of taxpayer-funded research. Supporters
of allowing contractors to retain title to patents resulting from federally funded research
argued that failure to do so would make it difficult to attract qualified firms to perform
government research and that absence of title would reduce incentives to invest in com-
mercial development of these inventions.
World War II and the Cold War that followed transformed the structure of the U.S.
national innovation system, including research at U.S. universities. Formerly funded
largely by state governments, the federal Department of Agriculture, and industry,
academic research experienced a surge of federal funding. As the growth in university-
industry research links had done during the 1920s and 1930s, increased federal funding
of university research strengthened two motives for university involvement in patenting.
First, the expanded scale of the academic research enterprise increased the probability
that universities would produce patentable inventions. Second, many federal research
sponsors required the development of a formal patent policy.
By the late 1940s, virtually all major U.S. universities had developed patent policies
(McKusick, 1948). In a survey of U.S. universities conducted in the late 1950s, Palmer
(1962) found that 85 institutions had adopted or revised patent policies between 1940–55.
Typically, patents from government-funded research were governed by the policies of the
sponsoring agency; privately sponsored research was managed on a case-by-case basis,
although sponsors often received preferential treatment. Significant differences nonethe-
less remained among U.S. universities in their assertion of rights to faculty inventions and
their willingness to pursue patent licenses.
1960s-era U.S. university patent policies and procedures reflected a continuing ambiva-
lence toward patents. Many institutions maintained avoidance in direct involvement
in patenting. For example, Columbia University’s policy left patenting to the inventor
and patent administration to the Research Corporation, stating that “it is not deemed
within the sphere of the University’s scholarly objectives” to hold patents (Mowery et
al., 2004). Harvard University, the University of Chicago, Yale University, the University
of Pennsylvania, and Johns Hopkins University adopted similar positions. All of these
universities discouraged or prohibited medical patents. Other universities allowed patents
on biomedical inventions only if it was clear that patenting would be in the public inter-
est. This institutional ambivalence toward patenting began to change during the 1960s,
although the prohibitions on medical patenting at Columbia, Harvard, Johns Hopkins,
and Chicago were not dropped until the 1970s.
The 1970s were a period of significant change in U.S. universities’ patent policies and
practices. The Research Corporation, which by 1970 was administering inventions for
over 200 institutions, began to encourage and assist its client universities in developing
capabilities to manage the early stages of the technology transfer process, particularly
invention screening and evaluation. In addition, a consequence of the growth in federal
support for biomedical research was the emergence during the 1970s of molecular biology
as a field characterized by numerous advances in basic science that promised commercial

applications of considerable interest to industry. The growth of biomedical research

increased universities’ interest in capturing revenues from licensing of biomedical patents
in an era of slower growth in overall federal research funding and accelerating growth in
research costs. U.S. universities increasingly began to seek patents for faculty inventions
and to manage their patenting and licensing activities themselves (Weiner, 1986).
Federal patent policies also changed during this period. Before the 1960s, federal
agencies lacked any policy beyond case-by-case negotiations for dealing with requests
from universities for title to inventions resulting from federally funded research. In the
mid-1960s the Department of Defense (DOD) began to grant title to inventions resulting
from DOD-financed research to universities with “approved” patent policies (i.e., policies
that required faculty to report inventions resulting from sponsored research) and allowed
universities to obtain patents on these inventions. Two other federal sources of academic
research funding—the Department of Health, Education, and Welfare (HEW) and the
NSF—began to negotiate Institutional Patent Agreements (IPAs) with universities in
1968 and 1973, respectively. IPAs eliminated the need for case-by-case reviews of the
disposition of individual academic inventions resulting from federally funded research
and facilitated licensing of such inventions on an exclusive or non-exclusive basis.
III. UNIVERSITY PATENTING, 1925–80
The heterogeneity among U.S. universities’ patent policies is reflected in considerable

differences in the size and characteristics of the patent portfolios that each university
acquired during the 1925–80 period. Patents issued to the nation’s 16 leading pre-1940
research universities (Geiger, 1986) accounted for only half of the academic patents
(excluding Research Corporation patents) issued during 1925–45. Ivy League institu-
tions avoided direct institutional involvement in patenting. Individual faculty from these
universities, however, used the Research Corporation for patent administration beginning
in the late 1920s. Several other private schools active in both engineering and the applied
sciences before World War II, such as the California Institute of Technology and Stanford
University, accounted for a substantial number of patents during 1925–45. Public univer-
sities were more heavily represented in patenting than private universities during this time,
however, both within the top research universities and more generally.
The 1970s represented the most dramatic period of change in U.S. university patenting
during the 1925–80 period. Overall university patenting grew significantly and became less
concentrated; more and more universities chose to manage their patents themselves; and
biomedical inventions increased in importance within university patenting and licensing.
During 1925–80, U.S. universities’ share of all domestically assigned U.S. utility patents
grew from zero to approximately 1 percent. After growing during the 1940s and 1950s, the
number of university patents remained roughly constant until the 1970s, when university
patenting increased significantly. The total number of university patents issued during
the 1970s alone was one-and-a-half times the total number of university patents issued in
the previous two decades.
The pre-1940 period reflected the dominance of chemistry in patenting activity,
reflecting inter-war academic research in fields including agricultural chemistry, industrial
chemistry, and chemical engineering, as well as the strong university-industry linkages in

those fields of research (Rosenberg and Nelson, 1994). Nearly 60 percent of the chemical
patents issued to universities during this period were organic compounds. Some of these
compounds were vitamins and could also be classified as biomedical inventions. During
the interwar period, U.S. universities made significant contributions to vitamin research
and synthesis (Apple 1996), and the most lucrative academic patent licenses of this period
involved vitamin-related inventions. These included the Williams-Waterman method for
synthesizing vitamin B1 (developed at the University of California and administered
by the Research Corporation), Milas’ vitamin A and B formulations (developed at
the Massachusetts Institute of Technology (MIT) and administered by the Research
Corporation), and Steenbock’s method for producing vitamin D (administered by WARF).
Increased federal financial support for biomedical research during the postwar era is
reflected in the growing share of biomedical patents during and after the 1970s. Much
of the growth in university patenting during the 1970s occurred in the biomedical area.
Non-biomedical university patents increased by 90 percent from 1968–70 to 1978–80, but
biomedical university patents tripled in number during this period. This rapid growth
in biomedical patents also reflected the expansion of the IPA program associated with
the major biomedical funding agency (HEW) during the 1970s. The increased share of
the biomedical disciplines within overall federal academic R&D funding, the dramatic
advances in biomedical science that occurred during the 1960s and 1970s, and the strong
industrial interest in the results of this biomedical research all affected the growth of
university patenting during this period.
IV. THE BAYH–DOLE ACT OF 1980
The Bayh–Dole Act (Public Law 96-517) became effective on July 1, 1981, creating a
uniform federal patent policy for universities and small businesses that gave them the
rights to any patents resulting from grants or contracts funded by any federal agency.2
The federal government retained a non-exclusive royalty-free license to any such patents
and also retained “march-in” rights to license or practice the invention when contrac-
tors’ licensing policies failed to promote utilization or where doing so was necessary for
public health or safety. The Act and subsequent regulations also incorporated policies
governing the timetable for disclosures of invention to the funding agency and for filing
patent applications. These regulations required that universities share licensing royalties
with inventors and mandated a preference for small businesses in the award of licenses
by universities and other research performers. The logic underpinning the Bayh–Dole
legislation was that patents and exclusive licensing were necessary to create incentives
for private firms to undertake the necessary investments to develop and commercialize
“embryonic” or early-stage university inventions (Mowery et al., 2004; Rai et al., 2009).
Mowery et al. (2001, 2004) argue that the Bayh–Dole Act was one part of a broader
shift in U.S. policy toward stronger intellectual property rights during the 1980s. Many
2
The Act was extended in 1984 by Public Law 98-620, which allowed universities to assign
their property rights to third parties and expanded the types of inventions that universities could
own (Henderson et al., 1998).

U.S. universities were active patenters and licensors long before the passage of the Bayh–
Dole Act, and expanded these activities during the 1970s and early 1980s in response
to significant advances in biomedical research and changes in the legal treatment of
patents on life forms. Moreover, the Court of Appeals for the Federal Circuit (CAFC),
established in 1982 to serve as the court of final appeal for patent cases throughout the
federal judiciary, soon emerged as a strong champion of patent holder rights. In addition,
a U.S. Supreme Court decision in 1980, Diamond v. Chakrabarty, upheld the validity of
a broad patent in the new industry of biotechnology, opening the door to patenting the
organisms, molecules, and research techniques emerging from biotechnology. Mowery et
al. conclude that the origins and effects of Bayh–Dole must be viewed in the context of
this larger shift in U.S. policy toward intellectual property rights.
One of the first studies of the effect of the Bayh–Dole Act on university patenting
was Henderson et al. (1998). Like Mowery et al. (2001, 2004), these scholars observed
that patenting by universities was rising prior to the passage of Bayh–Dole, in the order
of 15 times between 1965 and mid-1992. They also found that university patenting was
increasingly concentrated within drug and medical fields compared to a random sample,
from about 15 percent in 1965 to about 35 percent by 1992.3 In an analysis of university
patent characteristics, Henderson et al. compared 12 804 patents assigned to universities
during that period to a 1 percent random sample of all U.S. patents issued during the
same time. Using “forward citations” (future citations that reference the focal patent as
prior art), they constructed measures of patent “importance” (the number of times a
patent is cited) and “generality” (the dispersion of patent classes from which citing patents
reside) for university and control sample patents between 1965–88 (between 1975–88 for
generality). Overall, they found that university patents were both more important (more
highly cited) and more general (cited by patents from a more diverse set of technical fields)
than were the patents in the random sample. Differences in both measures increased
throughout the 1970s until the early 1980s. After that time, differences in importance and
generality between university random sample patents rapidly declined and were no longer
statistically significant. Henderson et al. attributed the relative decline of the “quality”
of university patents in the early 1980s to two factors. First, the number of patents
originating from smaller institutions was increasing. According to the authors, smaller
institutions’ patents are historically less cited than those from larger institutions. More
significantly, they found that an increasing number of patents held by all universities
received few or no citations at all. They attributed the overall decline in university patent
quality to the higher prevalence of these “low quality” patents. Based on these findings,
Henderson et al. (1998) argued that while one effect of the Bayh–Dole Act may have been
to increase overall university patenting and associated technology transfer to the private
sector, the relative decline in quality suggested the generation of commercially valuable
3
Azoulay et al. (2007b) also found increasing patenting by full-time medical school faculty
after Bayh-Dole—122 patents were granted to medical school faculty in 1976, rising to 2,664 in
1999. (Patents granted to medical school faculty as a share of all academic patents also increased
from 29 percent in 1976 to 53 percent in 2003.) Although a greater number of patents were held by
faculty in clinical departments than in basic science departments, this was due to the larger number
of clinical faculty. As a share, only 3.5 percent of clinical faculty members held patents, compared
to 12.1 percent of faculty from basic science departments.

inventions had not increased as a result of the Act. Thus, they concluded that the passage
of the Bayh–Dole Act was not associated with a successful shift by universities of their
research to a more commercial arena.
Based on an analysis of patenting by the Stanford University and the University of
California, two active patenters prior to the Bayh–Dole Act, Mowery et al. (2001) also
found little evidence that the Act shifted universities toward more applied research.
Mowery et al. (2001) observed a movement in patent licensing income from primarily
agricultural inventions in the mid-1970s to biomedical inventions by the mid-1990s. These
scholars interpreted much of this change due to the increasing commercial importance of
biomedical research, along with the general strengthening of intellectual property rights
beyond the Bayh–Dole Act.4
Thursby and Thursby (2002) also examined the factors underpinning the growth in
university patenting and licensing subsequent to the Bayh–Dole Act. They considered
whether increased patenting and licensing were due to a shift in research agendas by
university scientists toward more commercially relevant outcomes or whether this
increase could be explained by an increased propensity to patent and license previous
inventions within their existing research portfolios. Based on Association of University
of Technology Managers (AUTM) data on invention disclosures, patents, and licenses
from 64 North American universities between 1994–98, the authors conducted a total
factor productivity analysis to estimate the relative contribution of these three inputs
into the growth of academic commercial activity. While they observed a modest growth
in invention disclosures by faculty, most of the increase in commercial activity was due to
an increase in patent applications. Interestingly, they discovered a negative relative growth
rate in licensing compared to disclosure and patenting activity. The authors interpreted
these patterns as consistent with an increasing propensity by university administrators
to commercialize academic inventions, coupled with a declining marginal commercial
appeal of patent and licenses. To examine whether a shift in faculty research agendas was
behind this increase, the authors supplemented this analysis with a survey of 112 firms
licensing university inventions, inquiring about their motivations for licensing university
technology. Based on this survey, they found that these firms attributed an increase in
licensing to a growing interest by university administrators in commercializing inventions
as well as a receptivity to bringing in outside research, rather than a shift in faculty
research toward the needs of business. Similar to Henderson et al. (1998) and Mowery
et al. (2001), the Thursby and Thursby (2002) findings are consistent with the intended
effects of the Bayh–Dole Act to encourage commercial application of university research,
yet show little evidence that this increase was associated with a shift of faculty attention
toward more applied research.
Mowery and Ziedonis (2002) also examined the effect of the Bayh–Dole Act on uni-
versity patenting behavior. Like Henderson et al. (1998), Mowery and Ziedonis measured
the importance and generality of university patents before and after the Bayh–Dole Act.
Also consistent with Henderson et al., these scholars observed a general decline in the
4
Recent work by Love (2014) emphasizes a subsequent shift in university technology transfer
from the licensing of biomedical patents to licensing of computing and telecommunications
patents.

quality of university patents after the Act. However, Mowery and Ziedonis found that
the quality of patents issued to Stanford University and the University of California
(identified in the previous study as two prominent “pre-Bayh–Dole” patenters) did not
decline. They further analyzed the importance and generality of all patents issued to
U.S. universities, dividing these institutions into “incumbents” (those with ten or more
patents prior to Bayh–Dole) and “entrants” (those with no or few patents before 1980).
Compared to a control group of non-academic patents matched by date and technology
class, Mowery and Ziedonis found that the overall quality of patents among incumbent
universities did not decline after Bayh–Dole. The overall decline in patent quality
therefore was driven by the entry of new universities with limited patenting experience.
Mowery and Ziedonis interpreted these findings as suggesting that while Bayh–Dole
may have encouraged entry into patenting and licensing activity by universities that were
previously not active patenters, it had little effect on the quality of the inventions patented
by long-time patenters.
The Mowery and Ziedonis (2002) findings that a decline in the quality of university
patents after the Bayh–Dole Act was primarily due to the entry of inexperienced academic
patenters, rather than a shift by incumbents toward more applied research as a result of
the Act that could lead toward less significant patents, raises the question of whether
the quality of patents issued to entrant institutions could improve as they gain patenting
experience. Mowery et al. (2002) examined citations to patents issued to universities
during the 1980s–90s to investigate whether there was such convergence. Like Mowery
and Ziedonis (2002), they found little evidence that the quality of patents issued to
universities with significant patenting experience prior to Bayh–Dole declined after the
Act. Moreover, they found that quality of entrant university patents, while initially less
cited than established university patents, converged toward the latter part of the 1980s
and early 1990s to exhibit similar quality on average.5 These authors investigated possible
sources of the improvement in entrant patent quality, considering three possibilities: (1)
whether the entrant had a prior relationship with the Research Corporation, (2) whether
it had extensive early experience in patenting after Bayh–Dole, and (3) whether the entrant
university established a formal technology transfer office for patent management and
licensing immediately after the Act. They found none of these sources to be correlated
with an improvement in patent quality for entrant universities, leading them to conclude
that the convergence was driven by more informal mechanisms such as learning from the
experience of more established universities (through the movement of personnel or the
dissemination of “best practices” by AUTM, for example).
Research by Pierre Azoulay, Waverly Ding, and Toby Stuart examined patenting by
university faculty in the life sciences at the level of the individual scientist (Azoulay et
al., 2007a; 2009). Based on a panel dataset of 3862 scientists with Ph.Ds in scientific
disciplines related to commercial biotechnology, they found that on average, academic
scientists who co-authored journal articles with industrial scientists were 25 percent more
likely to patent. They also found that early career patenting by scientists had increased
5
Owen-Smith and Powell (2003) also found that the impact of university patent portfolios in
the life sciences increased from the 1980s through the 1990s, attributing this increase to a greater
“connectedness” between university researchers and the life sciences industry.

between 1968–99, suggesting that patenting “has become to be recognized as a legitimate

form of scientific output in the academic life sciences” (Azoulay et al., 2007a, p. 614).
Lastly, they found that patenting often appeared to be preceded by “a flurry of scientific
activity” (p. 615). These scholars interpreted this result as suggesting that publications
and patents are coincident outputs of the same scientific activity, rather than the view
that patents reflect an effort by scientists to monetize pre-existing accomplishments and
professional reputations.
In follow-on work, Azoulay et al. examined more fully the effect of faculty patenting on
the intensity and characteristics of publication in scientific journals. Using the same data-
set as in their previous study, these scholars found that patenting was positively related
to the rate of journal publications. They also found that patenting was positively related
to the quality of publications based on journal impact factors, although this finding is
weakly statistically significant (Azoulay et al., 2009). Lastly, unlike the institutional-level
analyses of Henderson et al. (1998), Mowery et al. (2001), and Thursby and Thursby
(2002), Azoulay et al. (2009) did find evidence that following patenting, scientists shifted
their research priorities toward topics with greater commercial application, although the
magnitude of this effect was modest.
Despite increased patenting by universities after the Bayh–Dole Act, university research
outcomes are disseminated to industry through a multitude of other channels. As Mowery
et al. (2004) and Kenney et al. (2014) discuss, those channels include publications, hiring
of graduates, access to sponsored research results, the formation of start-up firms, and
consulting.6
In an effort to place faculty patenting as a dissemination channel of university research
in perspective, Agrawal and Henderson (2002) surveyed and interviewed professors in
the mechanical engineering and electrical engineering and computer science departments
at MIT regarding their patenting activity. Despite having the highest propensity to
patent among all the departments at the university (with the exception of the biology
department), survey respondents reported that many faculty did not patent at all, and
that MIT professors published more scholarly papers than received patents. Agrawal and
Henderson found, however, that patents received by professors in these departments were
correlated with higher-impact research as measured by scholarly citations.
Zucker et al. (1998) reported a similar result within biotechnology in a study of 55
“star-scientists,” finding that published research output by faculty with patents was
significantly more highly cited than was research by faculty that did not patent. Together
with the Agrawal and Henderson (2002) findings in electrical engineering, these results
suggest that patenting may complement basic academic research, rather than substitute
for it, thus further alleviating concerns that the Bayh–Dole Act serves as an encourage-
ment to faculty to shift their research agendas.
6
Research by Thursby et al. (2009) indicates that consulting with industry underpins a signifi-
cant fraction of faculty patenting. They found that 26 percent of a sample of 5811 patents listing
university researchers at Carnegie Foundation classified Research I universities in 1993 as inventors
were assigned not to their university employers, but to private firms. Moreover, the patents assigned
to firms were less “basic” and more “applied” than those assigned to universities. Thursby et al.
interpreted these results as suggesting that many faculty actively engage in consulting, from which
the client firm would own the resulting property rights.

Other scholars have examined the conditions under which patenting is neither desirable
nor necessary for transfer of academic research results to industry. Jeannette Colyvas and
her colleagues (Colyvas et al. 2002) conducted case studies of the commercialization of 11
invention disclosures arising from research conducted by faculty at Columbia University
and Stanford University. Through interviews and examination of archival records of the
commercialization process, these scholars sought to determine the extent to which patent-
ing and exclusive licensing were necessary for commercialization. Of these 11 inventions,
eight were in the life sciences (biotechnological, pharmaceutical, or medical device),
and the remaining three were electronic or software inventions. Six of the life science
disclosures were licensed exclusively and two were licensed non-exclusively. All of these
inventions were either patented or had a patent application under review. The electronic
invention was also patented and one of the two software inventions was patented (the
other was protected by copyright). Based on their review, Colyvas et al. (2002) ascertained
that patents and exclusive licenses were necessary to induce firms to make the necessary
investments to develop the “embryonic” inventions in their study. For inventions that were
“ready to use,” however, these scholars concluded that patents and exclusive licenses, while
they generated revenue for the university, were not critical in bringing these inventions to
market.
More recently, Brian J. Love (2014) surveyed 2387 faculty in the top 20 (as ranked by
the 2013 U.S. News and World Report) electrical engineering and computing science
departments (including the MIT department surveyed by Agrawal and Henderson)
regarding their patenting activity, receiving 269 responses, or an 11.3 percent response
rate. Love’s evidence is consistent with the view that patenting assists the commer-
cialization of university research outcomes in these disciplines. Forty-two percent
of respondents stated that patents facilitated their ability to commercialize research,
although 12 percent reported that patents were detrimental in that process. Among the
subset that had previously started firms, a greater percentage felt that patenting by the
university hurt their efforts to commercialize research, however (24 percent versus 12
percent for the overall sample). Faculty were also asked whether patenting affected their
ability to work as consultants to industry, another form of commercialization activity.
Twenty-two percent stated that the ability to patent their research helped them find
consulting opportunities, while 18 percent stated that patenting hindered those activities.
Start-up founders, however, had the opposite view, on average, with 15 percent stating
that university patenting helped their consulting activities, but almost twice as many
reporting that it harmed those efforts.
Respondents to Love’s survey were less positive that patenting supported their ability
to conduct underlying research. Over 50 percent of respondents “strongly disagreed” and
almost 15 percent “disagreed” with the statement, “The ability to patent my university
research encourages me to do more research than I would otherwise” (p. 315). In contrast,
only 10 percent agreed or strongly agreed with that statement. Moreover, almost 60 per-
cent strongly disagreed with the statement, “The ability to patent my university research
encourages me to do higher quality research than I would otherwise” (with 15 percent
agreeing). Not surprisingly, 76 percent of respondents ranked “enjoyment/curiosity/
desire to advance knowledge” as the prime motivator in conducting research. Half of
surveyed faculty members reported that university patenting had no effect on their abil-
ity to attract research funding, while the remaining faculty were evenly split on whether

such patenting negatively or positively affected such ability (approximately 20 percent of

respondents on either side). A greater proportion of respondents (25–30 percent) felt that
university patenting negatively affected their abilities to collaborate across institutions,
while fewer than 5 percent felt that it facilitated such collaboration.
Kenney et al. (2014) also question whether patenting by the university is essential to
transfer research results to industry within the engineering and computer science disci-
plines. They report that in the electrical engineering and computer science department
at the University of California, Berkeley, patenting by faculty was an infrequent activity.
Quoting David Patterson, a central figure in the development of RISC and RAID tech-
nologies, he “probably never filed a single patent at Berkeley” (p. 117). Moreover, Kenney
et al. report that the Dean of the College of Engineering believed that patenting by faculty
stood in the way of faculty relationships with industry, arguing that “if they’re thinking
you’re going to go off and patent it, they’re not going to talk to you, but if you tell them
what you’re doing and they’re confident that you’re not going to steal their stuff and write
a patent, they’ll tell you more about their stuff ” (p. 118). The Kenney et al. (2014) and
Love (2014) findings reveal the importance of “two-way” interaction between academia
and industry. Universities may be a source of new discoveries for industry, but they
rely on ideas, advances, and insights from industry in turn. The insertion of intellectual
property rights into that relationship may have detrimental effects on the effectiveness of
that interaction.
V. UNIVERSITY SOFTWARE PATENTING AND “HOLD-UP”

CONCERNS
During the 1980s and 1990s, a series of court decisions increasingly expanded patent
protection for software. In 1981 the Diamond v. Diehr (450 U.S. 175 1981) decision by the
Supreme Court gave patent protection for software that controlled the physical process
for curing rubber. And in a 2002 decision in State Street v. Signature Financial Group
(149 F.3d 1368 Fed. Cir. 1998), the CAFC deemed all software that provided a “useful”
result to be patentable. As a result, software patenting as a fraction of all U.S. patent-
ing increased from essentially zero in the mid-1970s to 3 percent to 25 percent in 2006,
depending on the definition of a software patent (Hall and MacGarvie, 2010).
Patenting of software by universities also increased during the 1980s and 1990s, accord-
ing to a study by Arti Rai, Bhaven Sampat, and John Allison (Rai et al., 2009). These
scholars found that software patenting by universities rose by more than a factor of ten in
the wake of the Bayh–Dole Act, increasing from 37 software patents assigned to universi-
ties in 1982 to 396 in 2002. Moreover, the percentage of software patents relative to all
patents received by universities rose from 9 percent to 13 percent in the same time period.
Computer science and electrical engineering departments generated 52 percent of the
patented software inventions at the top 15 institutions that were issued software patents in
2002. Other disciplines also generated software patents, however, including neuroscience,
robotics, radiology, chemistry, and mechanical engineering, among others. Overall, more
than one-third of software patents in their sample had biomedical applications.
To investigate potential explanations for these increases, Rai et al. estimated a “patent
production function” for software patents based on the amount of software R&D, overall

R&D, and non-software patent propensity for each university in the sample.7 They con-
ducted cross-sectional and panel regression analyses of these factors using cross-sections
at five-year intervals between 1982 and 2002. The most consistent overall finding from
these regressions is that software patenting has a strong relationship with the level of
non-software patenting at the university.
Rai et al. (2009) surmise that the positive relationship between the level of overall
patenting and software patenting at the university level could reflect “economies of scale.”
That is, the marginal cost of obtaining an additional patent (including a software patent)
is lower for universities that have larger technology transfer operations (p. 1542). If the
logic of the Bayh–Dole Act is that such patents are necessary to induce firms to undertake
investment in commercializing early-stage or embryonic inventions and such incentives are
not necessary to develop software inventions, the authors contend that granting exclusive
licenses for patented university software inventions could be undesirable from a social
welfare perspective. These scholars also argue that the justification for patents as a mecha-
nism for promoting the transfer of “tacit” knowledge in technology transfer between the
inventor and licensee may be less potent for the commercialization of software inventions,
as they tend to be codified.
The most significant critique of university software patenting raised by Rai, Allison,
and Sampat is that software patents may enable the university or its exclusive licensee to
extract “rents” from other firms without facilitating or promoting technology transfer.
This problem can be exacerbated by “hold-up” (Lemley (2007a) and Ryan and Frye (2017)
raise similar concerns).8 They cite Eolas Technologies, Inc. v. Microsoft Corp. (399 F.3d
1325 Fed Cir. 2005), where Eolas, as a licensor of a University of California software
patent, threatened injunctive relief to extract rents larger than the value of the contribu-
tion made by the University or Eolas. They also describe other anecdotal evidence that
some universities have engaged in hold-up or other actions that hindered rather than
facilitated commercialization efforts, but fail to find a significant difference in the rates of
litigation between software and non-software patents in their sample. Although this result
prevents them from concluding that the frequency of hold-up is greater for university
software patents than non-software patents, Rai et al. note that hold-up may still be
7
Rai et al.’s sample is based on “doctoral” and “research” universities as designated by the
Carnegie Commission of Higher Education.
8
Lemley (2007b) provides a detailed description of the hold-up problem. Briefly, in sectors
such as information technology (which includes computer software) where firms developing
innovations rely on many patented inventions to commercialize a marketable product (Lemley
cites the example of an Intel microprocessor requiring clearance of 500 different patent
rights), a single patent holder can extract rents in excess of the contribution of its patent by
threatening “injunctive relief.” Under this doctrine, if a patent is found to be infringed and the
infringed patented technology cannot be easily removed from the process, the patentee can be
awarded an injunction that prevents the infringer from selling the product. Often the infringer
has made substantial ex ante investments in the technology and re-designing the product or
manufacturing process to exclude the infringed technology would be cost prohibitive. Such
“irreversible investments” gives the patent owner substantial bargaining power that often drives
“the licensing settlement value to a percentage that is much greater than it would be in a system
in which we calculated the value that the inventor had actually contributed to the product”
(p. 154).

more likely with patents and exclusive licenses for software to the extent that they are less
necessary for commercialization than for other technologies.
Despite the concerns raised and evidence provided in Rai et al. (2009), university
technology transfer offices differ in their approaches to software protection. Moreover,
increasing numbers of universities appear to be limiting the use of patents as protection
for software inventions. As mentioned above, the Dean of the College of Engineering at
Berkeley resisted patenting of engineering inventions (which include computer science
and software inventions). At the campus level, Rai et al. cite a change in policy at Berkeley
that endorses an “open source” approach to software commercialization. They also
report examples of software technologies broadly licensed by Stanford University and
the University of Washington without the use of patents.
Because the data assembled by Rai et al. end in the late 1990s, they are unable to address
more recent decisions on the incidence and use of software patents and licenses by univer-
sities, such as in the Bilski v. Kappos (130 S. Ct. 23218 2010), Mayo Collaborative Servs. v.
Prometheus Labs, Inc. (132 S. Ct. 1289 2012), and Alice Corp. Pty Ltd. v. CLS Bank Int’l
(134 S. Ct. 2347 2014) cases that roll back the patent eligibility of software. More research
needs to be done to understand current practices by universities in this realm.
VI. “ANTI-COMMONS” CONCERNS
A related class of concerns is that the expansion of intellectual property rights covering
“use-inspired” basic research may produce an “anti-commons” effect, particularly in
the life sciences (Eisenberg, 1996; Heller and Eisenberg, 1998).9 Similar unease over
the expansion of intellectual property rights over scientific research has been raised by
David (2004) and Nelson (2004). This body of work argues that intellectual property and
the exclusion of access to “upstream research” that was formerly in the public domain
threaten the ability of scientists to build on prior scientific discoveries. Moreover, the
granting of narrow property rights on pieces of knowledge that are distributed among
multiple parties can impose transaction costs when access to dispersed rights is needed to
further discovery and development (Eisenberg, 1996; Ziedonis, 2004).
The well-known 1980s discovery of the “Oncomouse” by Harvard University scientist
Philip Leder is often held up as an example of the dangers of granting intellectual
property rights to upstream research results (Murray, 2010; Murray and Stern, 2007).
Leder developed a mouse genetically engineered to be susceptible to cancer, which proved
to be valuable as a research tool to scientists studying the disease. The resulting research
was published in the journal Cell and broad patent rights were subsequently granted.
The Du Pont Corporation exclusively licensed the Oncomouse and aggressively enforced
its property rights, even demanding “reach-through” rights (royalties for any discoveries
or products developed using the Oncomouse), review of scientific research prior to
9
Donald Stokes (1997) coined the term “Pasteur’s quadrant” describing such research.
Research within Pasteur’s quadrant follows a quest for fundamental understanding that
simultaneously has practical application. Much biomedical research is thought to fall within this
category.

publication, and limits on breeding and sharing of the mice among scientific colleagues.
These limits were viewed by many in the scientific community as significant obstacles to
follow-on scientific research (Murray, 2010; Murray and Stern, 2007).
One of the earliest empirical tests of the anti-commons effect was conducted by
Fiona Murray and Scott Stern (2007). These scholars first constructed a sample of 340
peer-reviewed scholarly papers published in the scientific journal Nature Biotechnology
between 1997–99. For each of these articles, the authors determined whether a patent
covering the knowledge underpinning the article had been issued, thus constructing
“patent-paper” pairs. One hundred and sixty-nine of the articles were linked to a patent.
Using papers not linked to a patent to generate a control group, the authors conducted a
difference-in-difference analysis to compare the effect of the issuance of a patent on sub-
sequent journal citations to the paired paper. They found that a patent was associated with
a subsequent 10–19 percent decline in citations to papers in the subject group compared to
the unpatented control group. Moreover, this decline was larger for papers associated with
universities or other public research institutions. Although Murray and Stern were unable
to identify the precise institutional mechanisms underpinning their results, these findings
are consistent with an anti-commons effect in the patenting of academic research.10
Walsh et al. (2005) examined the effect of patenting on access to research results by
414 biomedical researchers in universities, governmental agencies, and research insti-
tutes. These scholars also found patenting to be a common occurrence among research
scientists. Twenty-two percent of respondents had filed for a patent within the two years
prior to the survey. Thirty-five percent of respondents had participated in a business
activity such as participating in a start-up, receiving licensing income, or negotiating
patent rights over their inventions. Despite their own patenting and commercial activity,
however, few of these respondents reported that patents held by others had restricted
their access to critical biotechnological knowledge required in their research. Only 5
percent of respondents reported regularly monitoring whether knowledge inputs were
patented, and 8 percent stated that they believed that they had conducted research using
patented knowledge.11 Of those reporting that they were aware of relevant intellectual
property, only five delayed the completion of an experiment by more than a month,
and only four modified their research design to avoid infringement. None of these
scientists reported that they had completely abandoned a line of research as a result of
patents held by other parties.12 These results suggest that an anti-commons effect from
patents may not be present in practice, contrasting with the findings of Murray and
Stern (2007).
10
Follow-on work by Fehder et al. (2014) found that the negative effect of patenting on the
cumulative citation of paired papers was concentrated in the initial years after a journal’s existence
(based on pairs involving papers published in Nature Biotechnology and Nature Materials).
11
Lei et al. (2009) also found that academic scientists largely ignore prior patents in a
survey of agricultural biologists at four research universities—77 of 85 respondents reported
that they had not checked whether a research tool that they would need in planned research was
patented.
12
Walsh et al. (2005) also surveyed 93 academic genomics researchers conducting signal
protein research, a patent-intensive area with strong commercial interest. Among these scientists,
only 3 percent reported abandoning a project due to a patent held by a third party.

Denial of access to research tools and materials appears to be a greater concern than
patent infringement for the scientists surveyed by Walsh et al. (2005). At least 75 percent
of respondents requested access to a material or tool in the two years prior to the survey,
averaging seven requests for materials to other academic scientists and two requests to
industry. Nineteen percent reported that their most recent request had been denied.
These denials caused significant delays in making progress on research, according to
respondents—one in six scientists reported a delay of at least one month, and one in
14 abandoned a research project completely as a result of difficulty in gaining access
to materials. A regression analysis revealed that the presence of a patent covering the
requested material had no effect on whether the request was fulfilled.
Historically, exchange of research materials typically involved a letter in which the pro-
vider requested acknowledgement in any resulting publication and that the material not
be passed onto a third party (McCain, 1991). In 1997–98, however, the National Institutes
of Health Working Group on Research Tools investigated factors that contributed to
increasing difficulties confronting researchers in obtaining access to research tools for use
in biomedical research (Eisenberg, 2001). The Working Group surveyed and interviewed
scientists in academia and industry, and university technology transfer professionals.
Based on its investigation, the Working Group determined that increasing formalism in
the transfer of research tools was an outgrowth of the increased patenting in biomedical
research (see also Streitz and Bennett (2003)). Such transfers are typically governed by
“material transfer agreements” (MTAs)—agreements that govern the transfer and use
of research materials, methods, and data for use in scientific research—and increasingly
include restrictions such as pre-publication approval, or the assertion of rights to future
inventions made using the provided tool through “reach-through” or “grant-back”
provisions (Eisenberg, 2001; Marshall, 1997). Also contributing to delays in access to
materials were diverse participants conducting basic research that included scientists in
academia as well as those employed by biotechnology and pharmaceutical firms, making
negotiations more difficult (Eisenberg, 2001). Academic agricultural biologists surveyed
by Lei et al. (2009) also highlighted delayed or blocked access to research tools, rather than
patents per se as impediments to research. Respondents in this survey attributed these
delays to the imposition of MTAs, mandated by their institutions to govern the transfer
of materials. Unlike the scientists in the Walsh et al. study, the scientists surveyed by Lei et
al. attributed the increasing use of MTAs to increased intellectual property protection—a
majority agreed with the statement, “Overall, the intellectual property protection of
research tools is having a negative impact on research in your area.” Lei et al. conclude
from this sentiment that “proliferation of patent protection is a key driver of the increase
in university enforcement of the use of MTAs for exchanges of research materials” (Lei
et al., 2009, p. 38).
Scott Stern presents a different viewpoint of the use of MTAs in the transfer of research
materials in scientific research. In a study of biological resource centers (BRCs)—non-
profit depositories for biological materials)—Stern (2004) finds that contamination of
widely used cell lines in biomedical research has ruined major research programs and
rendered results useless. Stern contends that the use of MTAs has facilitated the exchange
of materials by BRCs and helped ensure the quality of materials exchanged, concluding
that “putting MTAs in place at the time of patent approval lowers the cost of mutually
beneficial transactions between the developers of materials and follow-on research and

widens the availability of patented biomaterials” (Stern, 2004, pp. 96–97). Walsh et al.
(2003) take a similar position, arguing that the use of MTAs may simplify transactions
and increase researcher access.13
Despite the controversy over increasing formalism and use of MTAs in transferring
research materials among scientists, and the survey-based evidence discussed above, few
studies based on archival evidence on the effect of MTAs have been conducted. One
exception is Mowery and Ziedonis (2007). These scholars investigated the relationship
between patenting and the use of MTAs governing the outward transfer of materials at
the University of Michigan. They examined whether invention disclosures associated
with an MTA would be more or less likely to be patented or licensed than disclosures not
linked to an MTA. They also examined the effect of MTAs on the intensity and frequency
of patent citations. Their results find little evidence that MTAs restrict knowledge flow or
commercialization of university research. The proportion of MTA-linked disclosures that
are patented is significantly greater than the proportion of matching non-MTA-linked
inventions disclosed on similar dates. MTA-linked patents were more heavily and more
frequently cited than patents arising from control disclosures. Lastly, they found little
evidence that the presence of an MTA decreased demand for a license. As the authors
caution, the small sample size (83 disclosures linked to at least one MTA) and inability
to carefully control for possible selection bias between the treated and control samples of
the Mowery and Ziedonis study prevent any strong conclusions from being made from
this study.
Follow-on work by Thompson et al. (2018) utilizes a larger dataset of invention
disclosures between 1997–2007 at the University of California. These scholars examine
whether licenses affect the communication and dissemination of scientific knowledge by
examining the journal citations to published articles linked to patented invention disclo-
sures. Based on a sample of licensed patents versus a control sample of unlicensed patents,
Thompson et al. find that licenses do not appear to impede knowledge flow—the number
of journal citations made to articles linked to licensed patents is not significantly different
than the number of citations made to articles linked to unlicensed patents. For research
tools, however, Thompson et al. provide a less sanguine view, finding that licenses appear
to impede knowledge flow. They identify a research tool as an invention that is associated
with an MTA. Based on that definition, journal articles linked to licensed patents with
MTAs receive significantly fewer citations than unlicensed patents with MTAs.
VII. THE MARKET FOR UNIVERSITY PATENTS
The economic value of innovation has been shown in several studies to follow a skewed
distribution, with a minority of innovations accounting for a large share of value (e.g.,
the distribution of profits in pharmaceuticals (Grabowski and Vernon, 1990, 1994)). The
Mowery et al. (2001) study discussed earlier found that the commercial value of inventions
13
“...commercialization of research materials may actually increase access by creating market-
based institutions for distributing them rather than relying on gift exchanges among researchers”
(Walsh et al., 2003, p. 322).

discovered at Columbia University, Stanford University, and the University of California

was also highly concentrated. For example, these scholars found that between 1986–90,
the percentages of invention disclosures generating patents were 18.6 percent, 23.2
percent, and 20.4 percent, respectively. Similarly, the percentages of disclosures licensed
yielding income were 12.3 percent, 22.4 percent, and 7.4 percent. The shares of income
generated by the top five earners were even more concentrated, ranging from 92–94
percent at Columbia for the period between 1985–95 to 47–79 percent at the University
of California for the period between 1970–95.
Scherer and Harhoff (2000) found a similar pattern for income resulting from the
licensing of 118 Harvard University patents. Of that sample, the top 10 percent earning
patents accounted for 84 percent of the value. More generally, they found that the top 10
percent of patents generated between 91.5 percent and 93 percent of royalties between
1991–94 for a sample of patents from six research universities. Similarly, citing AUTM
data from 2012, Valdivia (2013) reports that in that year eight universities (the top 5
percent) earned 50 percent of the total university licensing income and the top 10 percent
(16 universities) captured 75 percent of total licensing income.
Other studies have attempted to directly measure the private value of patents. Hall et
al.’s (2005) analysis that links citations to patents to estimate the market value of publicly
traded manufacturing firms (as measured by “Tobin’s Q”), and the Hall and MacGarvie
(2010) study of software patenting and firm market value are two prominent examples.
Sampat and Ziedonis (2004) is one of the few studies attempting to estimate the economic
value of university patents. These scholars employed a sample of patents at Columbia
University and the University of California to estimate the degree to which citations could
predict the likelihood that a patent would be licensed by a private firm. Although they find
that the extent to which a patent is cited is correlated with the probability that a patent is
licensed, they fail to find a linkage between citations and the level of licensing revenues.
Unfortunately, the small size of their sample in the revenue analysis (56 patents) limits the
conclusions that can be drawn from this study.
Current research by Hsu et al. (2019) employs estimates of patent value using stock
market reactions to patent issuance by Kogan et al. (2017) to estimate the portion
of private patent value captured by universities. Based on estimates benchmarking
university and corporate patents, Hsu et al. find that universities capture 5–9 percent of
the economic value of their patent portfolios on average, although this percentage varies
widely across universities. Needless to say, more research is needed, but the difficulty of
obtaining sensitive licensing revenue data in order to conduct valuation studies has held
back progress in this area.
More headway has been made in understanding the factors that facilitate markets for
university patents and determining the types of firms that commercialize university tech-
nology. In a series of papers examining the licensing of 1397 patents covering inventions
at MIT between 1980–96, Scott Shane (Shane 2001a, b, 2002) examined the strength of
intellectual property rights on whether the invention (1) would be licensed, and (2) if so,
to a start-up or an established firm. Shane (2001a) argues that important inventions have
more economic value and thus are more liable to justify the foundation of a new firm.
In contrast, “incremental” inventions are less likely to spur an entrepreneurial enterprise
and thus more likely to be developed by an established firm. Lastly, because entrepreneurs
are often disadvantaged relative to existing firms due to a lack of complementary assets

such as manufacturing facilities or distribution systems (Teece, 1986), patents with narrow
scope are less likely to protect the entrepreneur from appropriation by others, and thus
are less likely to be licensed by an entrepreneur. Shane finds support for these hypotheses
based on the MIT data.
Shane (2001b) further examines the effectiveness of the patent “regime” to enable
appropriation of the value of the technology and its effect on whether the licensed patent
would be exploited through a start-up. Using a measure originally developed through a
survey of R&D managers by Levin et al. (1987) of the effectiveness of patents to enable
appropriation of their value within a technology field, Shane finds that greater patent
“strength” is associated with a greater likelihood of licensing to a start-up firm. He also
finds that start-ups are more likely to be the licensee in technological fields complemen-
tary assets such as marketing or distribution are less important for commercialization,
supporting the arguments of Teece (1986).
In the third paper of the series, Shane (2002) continues the examination of the
effectiveness of patents on university patent licensing. He argues that strong patent
rights facilitate the market for university technology by reducing transaction costs due
to adverse selection, moral hazard, and hold-up (Arora et al., 2001). Using the same
dataset of MIT inventions as in the previous studies, he finds that university inventions
are more likely to be licensed when patents are effective in facilitating the appropriation
of economic returns from the patented invention. Moreover, he finds that university
patents are more likely to be licensed to non-inventors when patent rights are strong, as
strong rights facilitate the transfer of related tacit knowledge (Arora, 1995). In contrast,
he finds that when patent rights are weak, university inventions are more likely to be
commercialized by inventor-founded start-up firms.14 Shane interprets this outcome as a
“second-best” solution—that is, failure in the market for knowledge due to weak patent
rights resulting in commercialization being undertaken by a start-up that often lacks the
requisite complementary assets necessary for a successful commercial outcome.
Lowe and Ziedonis (2006) revisited the Shane (2002) finding regarding the performance
of start-ups versus established firms in commercializing patented university technology.
Based on patenting and licensing data from the University of California, these scholars
found that economic returns from licenses to start-ups were greater than those generated
by more established companies commercializing similar technology, in stark contrast to
Shane (2002).15 Moreover, start-ups appeared to reach commercial sales from licensed
patents more quickly than established firms. Lowe and Ziedonis (2006) also found that
the majority of returns associated with the technologies licensed to start-ups were realized
subsequent to the acquisition by an established firm, leading them to conclude that many
academic start-ups serve as a “transitional” organizational form bridging technology
development of embryonic technologies in an academic setting toward ultimate commer-
cialization via the market for technology, thus overcoming Shane (2002)’s “second-best”
outcome.
14
In follow-on work based on similar data, Dechenaux et al. (2015) find that stronger patent
rights are also associated with a lower likelihood of contract termination.
15
Start-ups appeared to take longer than existing firms to terminate efforts toward technolo-
gies that ultimately failed, however.

VIII. CONCLUSION AND FUTURE DIRECTIONS
This chapter has surveyed the empirical literature on the patenting of university research.
Academic patenting in the United States has a long and controversial history, spanning
most of the twentieth century. Congressional action in 1980 simplified the process by which
universities could obtain patents covering federally funded research (through passage of
the Bayh–Dole Act), thus encouraging academic patenting and the commercialization of
university research. As pointed out by Mowery et al. (2004), however, the Bayh–Dole Act
was only one part of a broader shift toward stronger intellectual property rights in the
1980s. Coupled with the increasing commercial importance of research within Pasteur’s
quadrant, these expanded intellectual property rights resulted in the greater prevalence
and acceptance of university patenting that we see today.
This acceptance has not come without continuing controversy, however. Empirical
research has examined whether the Bayh–Dole Act was associated with a shift by uni-
versity researchers toward more applied research. Most of this work suggests that while
the Act was associated with increasing commercial activity by universities, there is little
evidence of movement away from basic scientific research.
More recent work has highlighted possible frictions and a potential chilling effect on
cumulative innovation induced by the insertion of intellectual property rights on upstream
scientific research (e.g., Eisenberg, 1996; Heller and Eisenberg, 1998). Other scholars have
questioned whether intellectual property rights are always necessary to transfer university
research findings to the private sector (e.g., Colyvas et al., 2002; Rai et al., 2009; Kenney
et al., 2014). These questions are still being debated within the literature.
This chapter has also cast a spotlight on the lack of concern of potential patent infringe-
ment by academic researchers highlighted by Walsh et al. (2005) and Lei et al. (2009). This
indifference by university scientists may reflect the de facto research exemption historically
granted by U.S. industry to university researchers in many areas of academic research.
But the increasing overlap of industrial and academic research, particularly in biomedical
fields, may increase the reluctance by industry to maintain such a policy (Eisenberg, 2001).
The 2002 Madey v. Duke (307 F.3d 1351 (Fed. Cir. 2002), cert. denied, 123 S. Ct.
2639 2003) decision of the CAFC effectively ended this informal research exemption
enjoyed by university researchers. In ruling on an infringement suit filed by an academic
researcher against Duke University, the CAFC argued that basic and applied research
was part of the “central business” of a university. Since universities benefited financially
and otherwise from their research activities, the CAFC reasoned, it was equitable under
the law for a patent holder to require that the university license patented material used in
research (Mowery et al., 2004). Dreyfuss (2004) suggests that a research exemption could
still be granted to universities if they agreed not to patent research results that rely on
patented material. Rai and Eisenberg (2003) propose that the Bayh–Dole Act be amended
to strengthen federal agencies’ discretion and ability to limit the patenting or restrictive
licensing of critically important scientific discoveries. On the other hand, Weschler (2004)
points out that despite the Madey v. Duke decision, given beneficial ties between academia
and industry, it is in industry’s interest to preserve university non-commercial research,
thus in practice to continue an informal research exemption.
An informal research exemption notwithstanding, interviews of technology transfer
office directors and other senior personnel by Jacob Rooksby (Rooksby, 2013) suggest

that universities often are uncomfortable with initiating infringement litigation because
it may incur costly legal fees, be perceived as being overly litigious, or conflict with
the university’s mission of public service. Despite these reservations, participation in
infringement lawsuits by universities is often unavoidable. For example, high-profile
conflicts between universities and industrial firms such as the University of Minnesota’s
dispute against Glaxo Wellcome and its patents covering its drug Ziagen, the University
of Rochester’s attempt to enforce its Cox-2 inhibitor patent against pharmaceutical firms,
and Genentech’s alleged theft of University of California IP suggest that, given overlap-
ping scientific interest between academia and industry, particularly over research within
Stokes’ (1997) “Pasteur’s quadrant,” frictions are likely to continue.
Such frictions could result in a chilling effect on patenting and licensing by universities.
Indeed, a study of licensing by 116 universities over the period from 1991–2000 by Shane
and Somaya (2007) found that each year of litigation spent by universities in patent
enforcement was correlated with an 18 percent decline in the number of licenses issued in
the subsequent year on average. Moreover, these scholars found this effect to be limited to
exclusive licensing. Interviews with university technology transfer professionals revealed
that due to the greater complexity of exclusive licenses and more frequent exclusive
licensing to small firms, litigation was more disruptive to exclusive than non-exclusive
licensing activity. Less is known about the costs imposed on the private sector as a result
of university litigation, calling for more research in this area.
It is also unclear whether more aggressive patenting by universities “crowds out”
inventive activity in the private sector. An extensive literature has examined the effect of
government funding of R&D on private R&D (see David et al. (2000) for a survey of
the evidence). Few studies have focused specifically on university research, however. One
exception is Adams (2000), who found that rather than substituting for private sector
R&D, university research serves as an inducement to industry, although this stimulus is
limited to learning about academic research. Likewise, Toole (2007) found that academic
research stimulated and thus complemented research by the pharmaceutical industry.
With respect to university patenting per se, the most notable study is Jaffe (1989). Jaffe
found that “geographically mediated” spillovers from university patenting induced
corporate patenting, particularly within biomedical technologies, electronics, optics, and
nuclear technologies. Given that universities have greatly expanded their patenting in
the 25 years since Jaffe’s early work, this area presents a promising opportunity for new
research. Clearly, on this and other important issues highlighted in this chapter, much
empirical work remains to be done.
REFERENCES
Adams, J. D. 2000. “Endogenous R&D Spillovers and Industrial Research Productivity.” NBER Working Paper
Series, 1–40.
Agrawal, A. and R. Henderson. 2002. “Putting Patents in Context: Exploring Knowledge Transfer from MIT”,
48:1 Management Science 44–60.
American Association for the Advancement of Science (AAAS). 1934. The Protection by Patents of Scientific
Discoveries: Report of the Committee on Patentsk Copyrights, and Trademarks. New York: Science Press.
Apple, R. D. 1989. “Patenting University Research: Harry Steenbock and the Wisconsin Alumni Research
Foundation”, 80 ISIS 375–394.
Apple, R. D. 1996. Vitaminia: Vitamins in American Culture. New Brunswick, NJ: Rutgers University Press.

Arora, A. 1995. “Licensing Tacit Knowledge: Intellectual Property Rights and the Market for Know-How”, 4
Economics of Innovation and New Technology 41–59.
Arora, A., A. Fosfuri, and A. Gambardella. 2001. “Markets for Technology and their Implications for
Corporate Strategy” in The Economics of Innovation and Corporate Strategy. Cambridge, MA: MIT Press.
Azoulay, P., W. Ding, and T. Stuart. 2007a. “The Determinants of Faculty Patenting Behavior: Demographics
or Opportunities?”, 63:4 Journal of Economic Behavior and Organization 599–623.
Azoulay, P., W. Ding, and T. Stuart. 2009. “The Impact of Academic Patenting on the Rate, Quality, and
Direction of (Public) Research Output”, 57:4 The Journal of Industrial Economics 637–676.
Azoulay, P., R. Michigan, and B. N. Sampat. 2007b. “The Anatomy of Medical School Patenting”, 357:20 New
England Journal of Medicine 2049–2056.
Colyvas, J., M. Crow, A. Gelijns, R. Mazzoleni, R. R. Nelson, N. Rosenberg, and B. N. Sampat. 2002. “How
Do University Inventions Get Into Practice?”, 48:1 Management Science 61–72.
Cottrell, F. G. 1912. “The Research Corporation, An Experiment in Public Administration of Patent Rights”,
4 (December) The Journal of Industrial and Engineering Chemistry 864–867.
Dasgupta, P. and P. A. David. 1994. “Toward a New Economics of Science”, 23 Research Policy 487–521.
David, P. A. 2004. “Can “Open Science” be Protected from the Evolving Regime of IPR Protections?”, 160
Journal of Institutional and Theoretical Economics 9–34.
David, P. A., B. H. Hall, and A. A. Toole. 2000. “Is Public R&D a Complement or Substitute for Private R&D?
A Review of the Econometric Evidence”, Research Policy 497–529.
Dechenaux, E., B. Goldfarb, S. Shane, and M. Thursby. 2015. “Appropriability and Commercialization:
Evidence from MIT Inventions”, 54:5 Management Science 893–906.
Dreyfuss, R. 2004. “Protecting the Public Domain of Science: Has the Time for an Experimental Use Defense
Arrived?”, 46:3 Arizona Law Review 457–472.
Eisenberg, R. S. 1996. “Public Research and Private Development: Patents and Technology Transfer in
Government-Sponsored Research”, 82 Virginia Law Review 1663–1727.
Eisenberg, R. S. 2001. “Bargaining Over the Transfer of Proprietary Research Tools” in Dreyfuss, R. C.,
Zimmerman, D. L., and First, H., eds., Expanding the Boundaries of Intellectual Property, 223–249. Oxford,
UK: Oxford University Press.
Fehder, D. C., F. Murray, and S. Stern. 2014. “Intellectual Property Rights and the Evolution of Scientific
Journals as Knowledge Platforms”, 36:C International Journal of Industrial Organization 383–394.
Geiger, R. L. 1986. To Advance Knowledge: The Growth of American Research Universities. New York, NY:
Grabowski, H. and J. Vernon. 1990. “A New Look at the Returns and Risks to Pharmaceutical R&D”, 36:7
Grabowski, H. G. and J. M. Vernon. 1994. “Returns to R&D on New Drug Introductions in the 1980s”, 13
Journal of Health Economics 383–406.
Hall, B. H., A. Jaffe, and M. Trajtenberg. 2005. “Market Value and Patent Citations”, 36:1 RAND Journal of
Economics 16–38.
Hall, B. H. and M. MacGarvie. 2010. “The Private Value of Software Patents”, 39:7 Research Policy 994–1009.
Heller, M. A. and R. S. Eisenberg. 1998. “Can Patents Deter Innovation? The Anticommons in Biomedical
Research”, 280:5364 Science 698–701.
Henderson, R. M., A. B. Jaffe, and M. Trajtenberg. 1998. “Universities as a Source of Commercial Technology:
A Detailed Analysis of University Patenting”, 80:1 The Review of Economics and Statistics 119–127.
Hsu, D. H., P-H. Hsu, T. Zhou, and A. A. Ziedonis. 2019. “Benchmarking U.S. University Technology
Commercialization Efforts: A New Approach.” Working Paper. Available at: “https://urldefense.proofpoint.
com/v2/url?u=https-3A__ssrn.com_abstract-3D3264090&d=DwMGaQ&c=yHlS04HhBraes5BQ9ueu5zKhE
7rtNXt_d012z2PA6ws&r=Lh_iVEf8nJlsegqOAObzUWj9qBpkVtZvRTFGcfuiJUBvpXSb2gNrr0jUY96kFb
ei&m=B33YL1-FTtfKD98FyjMQnq0xb0JQSSlk24RRZowmz_I&s=4yyEwWrr2IGp7p-4HiPldcoy4LAOt4
mmelvN7o36ScA&e=” https://ssrn.com/abstract=3264090 or “https://urldefense.proofpoint.com/v2/url?u=
http-3A__dx.doi.org_10.2139_ssrn.3264090&d=DwMGaQ&c=yHlS04HhBraes5BQ9ueu5zKhE7rtNXt_
d012z2PA6ws&r=Lh_iVEf8nJlsegqOAObzUWj9qBpkVtZvRTFGcfuiJUBvpXSb2gNrr0jUY96kFbei&m=B
33YL1-FTtfKD98FyjMQnq0xb0JQSSlk24RRZowmz_I&s=TaYZ8D1yojywiSBAzeWyLid41I7ZFu2RHjT4
CUT3mEA&e=” http://dx.doi.org/10.2139/ssrn.3264090.
Jaffe, A. B. 1989. “Real Effects of Academic Research”, 79:5 The American Economic Review 957–970.
Kenney, M., D. C. Mowery, and D. Patton. 2014. “Electrical Engineering and Computer Science at UC Berkeley
and in the Silicon Valley”, in Kenney, M. and Mowery, D. C., eds., Public Universities and Regional Growth,
97–126. Stanford, CA: Stanford Business Books.
Kevles, D. J. 1973. “The National Science Foundation and the Debate over Postwar Research Policy”, 68:241
ISIS 5–26.
Kogan, L., D. Papanikolaou, A. Seru, and N. Stoffman. 2017. “Technological Innovation, Resource Allocation,
and Growth”, 132:2 The Quarterly Journal of Economics 665–712.

Lei, Z., R. Juneja, and B. D. Wright. 2009. “Patents Versus Patenting: Implications of Intellectual Property
Protection for Biological Research”, 27:1 Nature Biotechnology 36–40.
Lemley, M. A. 2007a. “Are Universities Patent Trolls?”, 18:3 Fordham Intellectual Property Media and
Lemley, M. A. 2007b. “Ten Things To Do About Patent Holdup of Standards (And One Not To)”, 48 Boston
Levin, R., A. Klevorick, R. Nelson, and S. Winter. 1987. “Appropriating the Returns from Industrial Research
and Development”, 1987:3 Brookings Papers on Economic Activity 783–831.
Love, B. J. 2014. “Do University Patents Pay Off ? Evidence from a Survey of University Inventors in Computer
Science and Electrical Engineering”, 16 Yale Journal of Law and Technology 285–343.
Lowe, R. A. and A. A. Ziedonis. 2006. “Overoptimism and the Performance of Entrepreneurial Firms”, 52:2
Marshall, E. 1997. “Need a Reagent? Just Sign Here ...”, 278:5336 Science 212–213.
McCain, K. W. 1991. “Communication, Competition, and Secrecy: The Production and Dissemination of
Research-Related Information in Genetics”, 16:4 Science, Technology, and Human Values 491–516.
McKusick, V. L. 1948. “A Study of Patent Policies in Educational Institutions, Giving Specific Attention to the
Massachusetts Institute of Technology”, 245:3 Journal of the Franklin Institute 193–225.
Merton, R. K. 1973. The Sociology of Science: Theoretical and Empirical Investigations. Chicago, IL: The
Mowery, D. C. and A. A. Ziedonis. 2002. “Academic Patent Quality and Quantity before and after the Bayh–
Dole Act in the United States”, 31 Research Policy 399–418.
Mowery, D. C. and A. A. Ziedonis. 2007. “Academic Patents and Materials Transfer Agreements: Substitutes
or Complements?”, 32:3 The Journal of Technology Transfer 157–172.
Mowery, D. C., R. R. Nelson, B. N. Sampat, and A. A. Ziedonis. 2001. “The Growth of Patenting and
Licensing by U.S. Universities: An Assessment of the Effects of the Bayh–Dole Act of 1980”, 30 Research
Policy 99–119.
Mowery, D. C., R. R. Nelson., B.N. Sampat, and A. A. Ziedonis. 2004. Ivory Tower and Industrial Innovation:
University-Industry Technology Transfer Before and After the Bayh–Dole Act in the United States. Stanford,
Mowery, D. C., B. N. Sampat, and A. A. Ziedonis. 2002. “Learning to Patent: Institutional Experience,
Learning, and the Characteristics of U.S. University Patents After the Bayh–Dole Act, 1981–1992”, 48:1
Murray, F. 2010. “The Oncomouse that Roared: Hybrid Exchange Strategies as a Source of Distinction at the
Boundary of Overlapping Institutions”, 116:2 The American Journal of Sociology 341–388.
Murray, F. and S. Stern. 2007. “Do Formal Intellectual Property Rights Hinder the Free Flow of Scientific
Knowledge?”, 63:4 Journal of Economic Behavior and Organization 648–687.
National Science Board. 2016. “Science and Engineering Indicators 2016”, Technical Report NSB-2016-1,
Arlington, VA.
Nelson, R. R. 2004. “The Market Economy, and the Scientific Commons”, 33:3 Research Policy 455–471.
Owen-Smith, J. and W. W. Powell. 2003. “The Expanding Role of University Patenting in the Life Sciences:
Assessing the Importance of Experience and Connectivity”, 32:9 Research Policy 1695–1711.
Palmer, A. M. 1962. University Research and Patent Policies, Practices, and Procedures. Washington D.C.:
National Academy of Sciences–National Research Council.
Rai, A. K. and R. S. Eisenberg. 2003. “Bayh–Dole Reform and the Progress of Biomedicine”, 91:1 American
Scientist 52–59.
Rai, A. K., J. R. Allison, and B. N. Sampat. 2009. “University Software Ownership and Litigation: A First
Examination”, 87 North Carolina Law Review 1519–1570.
Rooksby, J. H. 2013. “When Tigers Bare Teeth: A Qualitative Study of University Patent Enforcement”, 46
Akron Law Review 169–205.
Rosenberg, N. and R. R. Nelson. 1994. “American Universities and Technical Advance in Industry”, 23:3
Ryan, C. J. and B. L. Frye. 2017. “An Empirical Study of University Patent Activity”, NYU Journal of
Intellectual Property Entertainment Law, Forthcoming: 1–28.
Sampat, B. N. and A. A. Ziedonis. 2004. “Patent Citations and the Economic Value of Patents”, in Moed, H.
F., Glänzel, W., and Schmoch, U., eds., Handbook of Quantitative Science and Technology Research, 277–298.
Secaucus, NJ: Springer-Verlag New York, Inc.
Scherer, F. M. and D. Harhoff. 2000. “Technology Policy for a World of Skew-Distributed Outcomes”, 29
Shane, S. 2001a. “Technological Opportunities and New Firm Creation”, 47:2 Management Science 205–220.
Shane, S. 2001b. “Technology Regimes and New Firm Formation”, 47:9 Management Science 1173–1190.
Shane, S. 2002. “Selling University Technology: Patterns from MIT”, 48:1 Management Science 122–137.

Shane, S. and D. Somaya. 2007. “The Effects of Patent Litigation on University Licensing Efforts”, 63:4 Journal
of Economic Behavior and Organization 739–755.
Stern, S. 2004. Biological Resource Centers: Knowledge Hubs for the Life Sciences. Washington, D.C.: Brookings
Institution Press.
Stokes, D. E. 1997. Pasteur’s Quadrant. Basic Science and Technological Innovation. Washington, D.C.:
Brookings Institution Press.
Streitz, W. D. and A. B. Bennett. 2003. “Material Transfer Agreements: A University Perspective”, 133 Plant
Physiology 10–13.
Teece, D. J. 1986. “Profiting from Technological Innovation: Implications for Integration, Collaboration,
Licensing and Public Policy”, 15:6 Research Policy 285–305.
Thompson, N. C., A. A. Ziedonis, and D. C. Mowery. 2018. “University Licensing and the Flow of Scientific
Knowledge”, 47:6 Research Policy 1060–1069.
Thursby, J. G. and M. C. Thursby. 2002. “Who Is Selling the Ivory Tower? Sources of Growth in University
Licensing”, 48:1 Management Science 90–104.
Thursby, J., A. W. Fuller, and M. Thursby. 2009. “US Faculty Patenting: Inside and Outside the University”,
38:1 Research Policy 14–25.
Toole, A. A. 2007. “Does Public Scientific Research Complement Private Investment in Research and
Development in the Pharmaceutical Industry?”, 50 (February) Journal of Law and Economics 81–104.
Valdivia, W. D. 2013. “University Start-Ups: Critical for Improving Technology Transfer”, Technical Report,
Brookings Institution.
Walsh, J. P., A. Arora, and W. M. Cohen. 2003. “Working through the Patent Problem”, 299:5609 Science 1021.
Walsh, J. P., C. Cho, and W. M. Cohen. 2005. “View from the Bench: Patents and Material Transfers”, 309
(September 23, 2005) Science 2002–2003.
Weiner, C. 1986. “Universities, Professors, and Patents: A Continuing Controversy”, 89:2 Technology Review
33–43.
Weschler, C. 2004. “The Informal Experimental Use Exception: University Research after Madey v. Duke
University”, 79:4 New York University Law Review 1536–1569.
Ziedonis, R. H. 2004. “Don’t Fence Me In: Fragmented Markets for Technology and the Patent Acquisition
Strategies of Firms”, 50:6 Management Science 804–820.
Zucker, L. G., M. B. Darby, and J. Armstrong. 1998. “Intellectual Capital and the Firm: The Technology of
Geographically Localized Knowledge Spillovers”, 36 (January) Economic Inquiry 65–86.

PART III
PATENT LAW DOCTRINES

12. Empirical studies in patentability
Ronald Mann* and Christopher Cotropia**
16
Contents
I. Introduction
II. Patentability Requirements
III. Quantifying the Validity of Issued Patents
A. The Economist Papers
1. Jean Lanjouw and Mark Schankerman, “Patent Quality and
Research Productivity: Measuring Innovation with Multiple
Indicators” (2004)
2. David Popp et al., “Technology Variation vs. R&D Uncertainty:
What Matters Most for Energy Patent Success?” (2013)
B. John Allison and His Co-Authors
1. John R. Allison et al., “Valuable Patents” (2004)
2. John R. Allison and Ronald J. Mann, “The Disputed Quality of
Software Patents” (2007)
3. John R. Allison et al., “Extreme Value or Trolls on Top? The
Characteristics of the Most-Litigated Patents” (2009)
4. John R. Allison et al., “Patent Quality and Settlement Among
Repeat Patent Litigants” (2011)
5. Ronald J. Mann and Marian Underweiser, “A New Look at Patent
Quality: Relating Patent Prosecution to Validity” (2012)
6. Michael Risch, “A Generation of Patent Litigation” (2015) and
Michael Risch, “The Layered Patent System” (2015)
IV. Judicial Outcomes and Implementation
A. Validity in General and Novelty (35 U.S.C. § 102)
1. General studies on validity
B. Novelty-Specific Studies
1. Dennis D. Crouch, “Is Novelty Obsolete? Chronicling the
Irrelevance of the Invention Date in U.S. Patent Law” (2009)
2. David S. Abrams and R. Polk Wagner, “Poisoning the Next
Apple? The America Invents Act and Individual Inventors” (2013)
C. Non-obviousness (35 U.S.C. § 103)
1. Results in non-obviousness decisions prior to KSR v. Teleflex
2. Results in non-obviousness decisions after KSR v. Teleflex
3. Non-obviousness standard as applied by the courts
* Albert E. Cinelli Enterprise Professor of Law, Columbia Law School.

** Professor of Law, University of Richmond School of Law.
281

4. Hindsight bias in non-obviousness decision-making

D. Disclosure Requirements (35 U.S.C. § 112)
1. John R. Allison and Lisa Larrimore Ouellette, “How Courts
Adjudicate Patent Definiteness and Disclosure” (2015)
V. Conclusion
References
I. INTRODUCTION
The current empirical literature studying the patentability requirements focuses on two
broad research questions. The first investigates whether a patent’s validity is empirically
quantifiable. This literature attempts to discern indicators of patent validity and patent
quality. The second area of focus in the literature is the courts, and how they evaluate a
patent’s (or patent application’s) validity. This literature examines judicial and administra-
tive outcomes and the content of these decisions reaching these outcomes in an attempt
to quantify a decision-maker’s attitude toward patentability and the specific standards
and analysis used in deciding patentability. Given that the literature is generally divided
between these two avenues of research, this chapter mimics this organization.
II. PATENTABILITY REQUIREMENTS
The basic patentability requirements fall into three general categories. There are initial
tests to see whether the patent’s claims, which define the invention, encompass eligible
patentable subject matter and meet the requisite utility (35 U.S.C. § 101). Patent law then
tests the claimed invention against the state of technology at the time of patenting via
the novelty and non-obviousness requirements. Novelty (35 U.S.C. § 102), and the related
statutory bar standards, ensure that the invention has not been previously made and that
the inventor has not improperly delayed filing the patent application claiming the inven-
tion. Non-obviousness (35 U.S.C. § 103), in contrast, ensures that patentable inventions
evidence a degree of technological advancement over the “prior art”—what has been
done before—and patent law was needed to prompt the invention’s creation. Another set
of patentability requirements (35 U.S.C. § 112) compares the claimed invention to the rest
of the patent disclosure, ensuring that the patent’s specification enables the invention,
adequately describes the invention, and discloses its best mode. There is also a require-
ment that the patent’s claims be definite.
All of these patent requirements are initially evaluated by the United States Patent
and Trademark Office (USPTO) at the application phase. Once the USPTO believes
one or more of the patent claims meet the standards, and thus are valid, the application
issues. The patentability requirements can be tested again in a United States district court
via challenges by alleged infringers. The USPTO may even re-evaluate its own work in
post-issuance reviews such as inter partes review (IPR) decided by the Board of Patent
Appeals and Interferences (BPAI). Appeals from these challenges, and initial denials of
patentability, are heard by the United States Court of Appeals for the Federal Circuit, a
national appellate court tasked in 1982 with hearing all patent appeals. It is against this

Empirical studies in patentability 283
basic backdrop on the patentability requirements and the institutions that apply and
review them that the empirical literature on patentability sits.
III. QUANTIFYING THE VALIDITY OF ISSUED PATENTS
Quantifying the quality of a body of patents is a challenging task, because the quality
of any particular patent depends on a comparison of that patent to pre-existing patents
and prior art. In theory, at least, an investigator could take up a group of patents and
examine them carefully, one by one, to form an educated opinion about the quality of the
patents; not surprisingly, none of the academic literature has taken that tack. A second
approach would be to take the body of patents, identify objective attributes of those
patents that might correlate with some metric of quality, and examine the extent to which
those attributes indirectly shed any light on the inventive step of the subject patents. To
the extent the literature sheds any quantitative light on this question, it is the second
path that the literature has taken. That path, of course, does not directly examine the
validity of the patents at all, much less the likelihood that a particular patent is “novel” or
“enabled” or covers patentable subject matter or the like. Rather, as the discussion below
makes clear, that literature analyzes the correlation between certain patent attributes
and the patent’s value, frequency of litigation, or the like. Still, as the discussion below
suggests, the literature by its end will shed some light on empirical assessment of patent
validity.
With that caveat, the following discussion is segmented, for expository purposes, into
two distinct lines of papers. The first line includes two unrelated papers by economists.
The second discusses an extended series of five papers co-authored by John Allison and
his various co-authors and a recent paper that continues this line of analysis.
A. The Economist Papers
1. Jean Lanjouw and Mark Schankerman, “Patent Quality and Research Productivity:
Measuring Innovation with Multiple Indicators” (2004)
It might seem odd, after suggesting that the two economist papers are “miscellaneous,” to
start with them. But we do so because the first of them is undoubtedly the seminal paper
in this area: Lanjouw and Schankerman’s “Patent Quality and Research Productivity”,
which applies a four-factor index of quality to a body of 100 000 patents applied for
from 1975–93, spread across seven technology areas. As we will see in the discussion
that follows, much of the effort of constructing such a paper is in the selection of the
patent attributes that the paper will analyze. In this paper, the first in the area, Lanjouw
and Schankerman analyze four patent attributes: number of claims, forward citations,
citations in the application, and family size. In light of developments in later literature,
it bears emphasis that Lanjouw and Schankerman do not offer any empirical justifica-
tion for the factors that they select; they offer only general reasons to believe that those
indicators should correlate with the innovative quality of the patent. Analytically, they
model the four attributes as evidence of a single latent factor, which they characterize as
quality. Lanjouw and Schankerman rest their analysis on an index constructed from a
linear combination of the values of the four attributes for each patent.

Table 12.1 Changes in patent quality index
Percentile Drugs Chemicals Electronics Mechanical

Top 5% 24.5 27.0 19.7 16.0
10% 21.3 24.3 20.1 16.7
50% 23.2 21.6 21.9 19.7
90% 27.1 24.1 25.3 23.0
95% 27.9 24.9 25.9 23.8
Notes: Each cell gives the mean percentage increase in the average quality for a randomly drawn (without
replacement) sample of 787 patents per year in each technology field, 1975–93. 787 is the minimum number
of patents in any year in any technology area. Three technologies are excluded due to small numbers in the
early years. Similar patterns hold for sub-periods 1975–84 and 1984–93, although changes are smaller for the
second period.
Source: Lanjouw and Schankerman (2004), “Patent Quality and Research Productivity”, Table 5.
Lanjouw and Schankerman draw two major conclusions. The first is that the index
increased substantially during the time period of their analysis, for each of the industries
that they examined. Table 12.1 is illustrative.
Lanjouw and Schankerman use that increase in quality to suggest that the decline
during their study period of the rate of patenting to R&D investment is offset substan-
tially by the increase in the average quality of each patent that was obtained. Analyzing
the question separately for each of the four industries that they study, they conclude that
the increased quality offsets about one-third of the observed drop in the ratio of patents
to R&D investment.
Interestingly, Lanjouw and Schankerman observe no significant differences at the
firm level, suggesting that firms do not in an observable way differentiate their strategies
between rarer higher-quality patenting and more frequent lower-quality patenting. They
do observe, however, significant relationships between the quality index of a firm’s patents
and its market capitalization (taking account of the firm’s level of patenting and R&D
expenditure), which suggests (intriguingly to this observer) that the market might well
have some ability to account for the innovative value of a firm’s patents, even for large
firms likely to have scores of patents.
2. David Popp et al., “Technology Variation vs. R&D Uncertainty: What Matters Most
for Energy Patent Success?” (2013)
The second of the two economist papers, Popp et al.’s 2013 paper on “Energy Patent
Success” examines patents in the vein of knowledge transmission—from the perspective
that patents that receive more forward citations are more successful from a social per-
spective because their advances have fostered more substantial related work. Within the
domain of energy patents, the authors can examine a population of all of the patents with
priority dates from 1971–2008 (about 800 000 patents). Using IBM’s Delphion database,
they collect the objective data points evident from the face of the patents. They then use
that data to identify predictors of success. Importantly, because they are focused on the
knowledge-transmission function of patents, they use so-called “forward” citations—
citations in later patents—as their sole indicator of success. Responding to the difficulty

that most patents have no forward citations at all, the authors use count data regression
techniques and quantile regressions.
The principal findings of the paper relate to three attributes. First, they link increased
forward citations (“better” patents for their purposes) with issuance to a domestic cor-
poration (as opposed to a government or foreign patentee) and to the number of claims.
Conversely, they link the number of references in the patent to lowered forward citations.
B. John Allison and His Co-Authors
The principal work from law professors using empirical analysis to assess the quality of
issued patents is a lengthy line of papers all of which are authored by John Allison or one
of his several co-authors. His centrality to the ongoing project rested on his willingness to
undertake the incredibly laborious task of reviewing patents one by one to collect detailed
information about the nature of the claimed invention that could not be captured with the
more “automated” data collection that uses standardized fields of information apparent
from the face of the patents.
1. John R. Allison et al., “Valuable Patents” (2004)

The first of these papers (“Valuable Patents”) was published in 2004, almost simultane-
ously with Lanjouw and Schankerman. For this paper, Allison was joined by Mark
Lemley, Kimberly Moore (now Judge Moore), and R. Derek Trunkey. Like his other
papers, the paper does not examine the “quality” or “correctly issued” questions directly,
but rather focuses on the value of patents. Working from the premise that patents
involved in litigation are more likely, as a class, to be more valuable than patents that are
not involved in litigation, “Valuable Patents” compares litigated patents to non-litigated
patents, relying on the inference that the characteristics of litigated patents are the
characteristics that make them valuable.
The focus of the study is a dataset of about 7000 patents involved in litigation that ter-
minated in 1999–2000. The authors compare those patents both to a population of about
3 million patents issued from 1963–99 and also to a smaller dataset of 1000 patents hand-
collected by the authors. The larger group of patents comes from the National Bureau of
Economic Research (NBER) Patent Citations Data File, which allows comparison of the
patents on the objective characteristics that can be extracted by automation from the face
of the issued patents. With regard to the 1000 patent samples, the authors also collected
data (not available in the NBER file) on foreign citations, non-patent prior art (NPPA),
nationality of the owners and whether the owners were a small entity (“SBE status”), and
also applied a more refined specification of the area of technology (the results of Allison’s
patent-by-patent analysis).
Relying on the results of a straightforward logit regression, the authors emphasize
several characteristics associated with patents involved in litigation. The first part of the
study examines a number of patent attributes commonly discussed in the prior literature
and, for the most part, confirms that they relate to litigation. Here, the authors conclude
that the number of claims, total citations, citations to U.S. patents, citations to NPPA,
self-citations, and forward citations all correlate directly with litigation.
The authors also observe from their regressions several attributes that relate to litiga-
tion that were new to the literature. In this part of the paper, the authors discuss positive

Table 12.2 Patent attributes and litigation (Logit)
Patent attribute Estimate

Claims 0.015 (0.0006)***
Prior art references 0.087 (0.004)***
Forward citations 0.049 (0.0016)***
Inventors −.064 (.01)***
Years in prosecution .038 (.013)***
Foreign assignee −.67 (.068)***
Source: John R. Allison et al. (2004), “Valuable Patents”, Table 4. Table reports estimates (standard errors).
N=3,500,000.
relations between litigation and the number of continuations applications, the number of
patents in the “family,” the time spent in prosecution, and the SBE status of the inventor
(i.e., small-inventor patents are more often litigated) Conversely, they find that litigation
is inversely related to the age of the patent, the number of inventors on the patent, and
to having a foreign issuer.
Table 12.2 summarizes the central findings.
2. John R. Allison and Ronald J. Mann, “The Disputed Quality of Software Patents”
(2007)
Second in the line of papers co-authored by John Allison is his 2007 paper with Ronald
Mann, which responds to the controversy over the propriety of patenting software
by looking specifically at software patents. This paper follows directly from “Valuable
Patents”, but marks a substantial analytical step forward in the literature. Specifically,
it is the first paper to use a set of objectively determined patent attributes to assess the
“quality” of a group of issued patents. As we shall see, there are several problems with
the methodology—later papers, for example, will undermine the relevance of the patent
attributes that the paper analyzes. But that does not undermine the methodological
contribution.
The specific topic is the quality of software patents—a group of patents frequently
vilified for their lack of quality. To assess the situation, the authors collect all of the
34 000 patents held by any firm listed in the “Software 500” at any point in the five years
from 1998–2002. Allison then examines the patents with a view to determining whether
each should or should not be regarded as a patent on a “software” invention. Applying
a definition that limits “software” patents to those in which at least one claim element
covers data processing, Allison separates the “software” from the “nonsoftware” inven-
tions, concluding that about 62 percent of the patents held by firms in the Software 500
are true software patents.
The paper emphasizes the distinction between “value” (the economic returns on a
patent), which obviously correlates highly with litigation, and “quality,” in the sense that
the patent is valid. The paper then argues that five patent characteristics can be regarded
as correlating with the “quality” (which for these authors means validity) of the patent:
prior art references, claims, forward citations, maintenance fees, and patent families.
Because their dataset is almost entirely domestic patents, patent families are not a useful

Table 12.3 Mean attributes of software and non-software patents
Attribute Software patents Non-software patents

Claims 23 (15) 18 (12)
Prior art references 19 (31) 14 (17)
Forward citations 1.1 (1.5) 0.9 (1.2)
Source: John R. Allison and Ronald J. Mann (2007), “The Disputed Quality of Software Patents” Table 2.
N=20,000.
indicator for their project. Similarly, in a dataset dominated by recently issued patents,
maintenance fees are not useful.
Accordingly, the body of the paper compares the claims, prior art references, and for-
ward citations of the software patents to those of the non-software patents. Although the
paper presents detailed findings about the different types of references and the different
types of NPPA, the key finding of the paper is that the software patents have significantly
more claims, prior art references, and forward citations than the non-software patents.
For each of those key findings, Table 12.3 presents the mean value, with the standard
deviation in parentheses.
The paper tests the robustness of those findings in two ways. First, it reports the results
of a variety of models (OLS, Poisson, firm-level fixed-effects), controlling for industry
and year of issuance, among other things. The regressions suggest that the findings related
to claims and prior art references are robust, but that the findings related to forward cita-
tions are less substantial. Second, a parallel set of models compares the software patents
of the pure software firms to those of the mixed software firms (those with less than 80
percent of their revenue from software sales). The findings are quite similar: the software
patents of the software firms have markedly more claims, prior art references, and forward
citations than the software patents of the mixed firms.
3. John R. Allison et al., “Extreme Value or Trolls on Top? The Characteristics of the
Most-Litigated Patents” (2009)
Allison is joined by Mark Lemley and Joshua Walker for the 2009 paper (“Extreme
Value”). That paper essentially replicates the methodology of “Valuable Patents”, with
two crucial changes: an updated dataset and a separate examination of the most heavily
litigated patents. The authors used the Stanford IP Litigation Clearinghouse to identify
all patents litigated eight times between 2000–07, a group of 106 patents. The authors
then collected a random control set of 106 patents litigated a single time during the same
period.
The authors used logistic regression to identify significant differences between the two
groups of patents. As summarized in Table 12.4, the findings are even starker than those
summarized in “Valuable Patents”. The mean number of claims in the most-litigated
patents is about 50 percent higher than in the once-litigated patents, and the mean num-
bers of applications, forward citations, domestic references, and foreign references for the
most-litigated patents are more than twice those of once-litigated patents.
As in “Valuable Patents”, the authors draw few inferences about the implications of their

Table 12.4 Most-litigated and once-litigated patents
Patent attribute Most litigated Once litigated

Applications 4.3 (4.9) 2.0 (1.4)
Forward citations 32 (42) 14 (23)
Claims 39 (45) 24 (19)
US references 61 (109) 23 (31)
Foreign references 9 (19) 4 (8)
NPPA 53 (111) 6 (16)
Source: John R. Allison et al., “Extreme Value or Trolls on Top? The Characteristics of the Most-Litigated
Patents”, Tables 2 and 3. N=212. Table reports the mean (standard deviation). All attributes have a p-value of
0.0001 or less.
findings for the “quality” of the patents; they make it quite plain that they aim to examine
the private value of the patents, not their social value or validity. They do, however, examine
the technologies from which the most-litigated patents come (disproportionately IT) and
the types of firms that own them (predominantly small enterprises). On the last point, 84
percent of once-litigated patents—but only 46 percent of the most-litigated patents—are
held by firms that participate in the market in which they are enforcing the patent.
4. John R. Allison et al., “Patent Quality and Settlement Among Repeat Patent
Litigants” (2011)
The next paper, “Patent Quality and Settlement Among Repeat Patent Litigants” (“Repeat
Patent Litigants”) is essentially a match to 2009’s “Extreme Value”, with the same three
authors (Allison, Lemley, and Walker) and an analytical purpose that builds directly on
“Extreme Value”. Thus, where “Extreme Value” studied the patents that were litigated
most frequently, this paper studies the patent owners that litigate most frequently. The
findings here are fascinating, and arguably more directly relevant to the validity question
than anything else in the Allison canon.
The most provocative finding is that the most-litigated patents are much more likely
than the once-litigated patents to be held invalid if the case proceeds to trial: specifically,
the once-litigated patents (for purposes of “Repeat Patent Litigants”, a dataset of 343
patents) withstand trial-stage validity challenges 47 percent of the time (26/55), while the
most-litigated patents survive only 11 percent of the time (8/75). What makes that finding
even more provocative is its overlap with three closely related findings. Specifically, SBEs
are much less likely to prevail in litigation forced to a judgment (23 percent) than larger
entities (56 percent); non-practicing entities (NPEs) are much less likely to prevail (9
percent) than entities producing products that use the patent (50 percent), and software
patents are much less likely to prevail (13 percent) than non-software patents (51 percent).
The number of observations is sufficiently large that those results are not only substan-
tively, but also statistically significant. The multivariate analysis testing those findings sug-
gests that the relationships are robust. The biggest problem is that there is so much overlap
among the unsuccessful categories (SBEs, NPEs, and software patents) that the multivari-
ate logistic regression is unable to unpack the relative weight of the three indicators. The
authors conclude that patentees with those three characteristics—a substantial group

among the most litigious patentees—are those consuming the most social resources, as
they burden the federal judicial system with so much apparently meritless litigation.
The authors emphasize the oddity that the most litigious litigants seem to be the least
savvy at anticipating an unsuccessful verdict, noting that they are so much more likely to
proceed to judgment unsuccessfully than litigants with less repetitive experience. At first
glance, the finding seems to be in tension with the common theme of game theory that
“repeat players” gain valuable experience and connections that give them an advantage
over single-time players. For purposes of this summary, we underscore one possibility that
Allison, Lemley, and Walker examine: the possibility that the patentees are “crazy like a
fox,” litigating to judgment in circumstances where that is appropriate.
Although Allison, Lemley, and Walker do not find the hypothesis compelling—
emphasizing the lack of supportive data—it does appeal to me. The key fact here (as
Allison, Lemley, and Walker recognize) is the possibility that the most-litigated patents
are those for which a favorable verdict promises the greatest potential bump in patent
value. If a patent has an immense potential payoff ($500 million), but only a slight chance
of prevailing (10 percent), it is easy to see why the plaintiff might press to judgment more
often than plaintiffs negotiating with defendants over more balanced disputes: a defend-
ant advised that it has a 90 percent chance of prevailing seems most unlikely to agree to
the large settlement that a purely rational calculation of present values might suggest. If
the plaintiffs want to obtain that recovery, they will have to roll the dice to obtain it. If
that seems like a strategy with a highly variable outcome, it should surprise none of us
that the litigants pursuing this strategy are “pooling” the risks: assembling portfolios that
contain large numbers of these “shoot-the-moon” patents of dubious validity.
Stepping back from the analysis that the authors present, the paper provides a big
step forward in the literature of the relation between “value” and “quality.” As the
authors note, the patents are at the far right tail in their possession of the traditional
economic attributes associated with value, which resonates with their place among the
most frequently litigated patents. What is more important for our purposes, though, is
that patents of remarkably low quality (probable validity) dominate this group of notably
valuable patents: patents that succeed only 10 percent of the time when pressed to judg-
ment. By underscoring a sector in which validity and value are plainly not linked, the
paper illustrates the importance of reading “Valuable Patents” as written—a discussion
of features associated with high value, not a discussion of features associated with a high
likelihood of validity.
5. Ronald J. Mann and Marian Underweiser, “A New Look at Patent Quality: Relating
Patent Prosecution to Validity” (2012)
The last of the papers in this line (“Patent Quality and Validity”) comes from Ronald
Mann, co-author with John Allison on one of the earlier papers, and Marian Underweiser
from IBM. Their paper departs from the earlier line of Allison papers in two ways directly
relevant to the topic at hand. First, as compared to the earlier Allison papers, which
analyze characteristics associated with value (principally correlated with litigation) as
opposed to analyzing patentability directly, Mann and Underweiser undertake to identify
features of patents that are observable at the time of issuance that relate to validity.
Second, they do not limit their data collection to patent attributes observable on the face
of the patents; rather, they examine the entire prosecution histories.

Table 12.5 Patent attributes and validity
Patent attribute Coefficient (S.E.)

Claims .35 [.06]**
Applicant IDS filings .22 [0.03]**
Examiner added references (.10) [0.01]**
Rejections (unlisted references) .37 [0.10]*
Foreign references .17 [0.03]*
Application date (years) (.15) [0.01]*
Source: Mann and Underweiser (2012), “A New Look at Patent Quality: Relating Patent Prosecution to
Validity”, Table 3. Table reports the logistic coefficient [bootstrapped standard error]. **p<0.01; *p<0.05
For that purpose, they collect a dataset of all patent validity decisions in the Federal
Circuit from 2003–09, a total of 266 cases adjudicating the validity of 366 patents. For
each of those patents, they supplemented a dozen datapoints from the face of the patent,
with 30 additional datapoints collected from the prosecution history, including such
things as the details of the original application, changes between the original application
and the issued patent, information disclosure statements filed by the inventor, office
actions and their basis, and continuation applications. Mann and Underweiser then use a
logistic regression to assess the relation between the patent attributes in their dataset and
the validity of the patent (as determined by the Federal Circuit). Table 12.5 summarizes
the most substantial findings.
Mann and Underweiser interpret the findings against a model that treats the quality of
the patent (its likely validity) as a function of the invention, the effort of the applicant,
and the effort of the examiner. In that framework, the number of claims and the number
of foreign references relate both to the breadth of the invention and to the effort of the
applicant in defining the invention; Mann and Underweiser’s findings suggest a positive
relation with validity. The number of information disclosure statement filings is evidence
of applicant effort, which also seems to relate directly to validity. Conversely, the numbers
of examiner-added references and rejections based on unlisted references are evidence of
work that the examiner is doing that the applicant should have done in the first instance.
As such, it is not clear in the abstract whether they would contribute to validity (because
the work is being done) or undermine validity (because the applicant should have done the
work before filing the application). In this dataset, though, the examiner-added references
appear to relate inversely to validity (suggesting the applicant’s failure to understand the
relevant prior art), while rejections based on new references relate directly to validity
(suggesting that the examiner’s affirmative step in response to the unlisted art usefully
limited the scope of the patent).
6. Michael Risch, “A Generation of Patent Litigation” (2015) and Michael Risch, “The
Layered Patent System” (2015)
Michael Risch recently published two empirical studies using similar methodologies as
those described above. Risch randomly selects 1311 patent cases filed in the 1990s and
2000s, which involve 791 patents. The dataset contains cases filed by NPEs and non-

NPEs. In addition to observing aspects of the patent litigation, Risch investigates patent
characteristics associated with findings of patent validity and invalidity. Performing a
step-wise regression, Risch finds a negative, statistically significant relationship between
backward citations and a holding of invalidity. Risch postulates that this potentially
counterintuitive finding may be due to the increase in prior art citation during re-
examinations of these patents. And Risch observes a statistically significant relationship
between a patent’s exposure to re-examinations and a co-pending litigation invalidating
the patent.
IV. JUDICIAL OUTCOMES AND IMPLEMENTATION

Another avenue of empirical study of patent validity focuses on institutions and decision-
making, instead of patent quality. These studies examine decisions and results at the
Federal Circuit, district courts, and the USPTO. The studies vary between observing
the ultimate outcome decisions and focusing on the specific methods and approaches the
deciding body took in making the decision. These studies are summarized below, grouped
by the specific aspect of validity they investigate.
A. Validity in General and Novelty (35 U.S.C. § 102)
A handful of studies looked at all categories of invalidity. These broad studies are
supplemented by focused examination of decisions under first-to-invent compared to
first-to-file regimes.
1. General studies on validity
1.1. John R. Allison and Mark A. Lemley, “Empirical Evidence on the Validity of
Litigated Patents” (1998) Allison and Lemley (1998) study 299 patents litigated in 239
cases over an almost eight-year period from early 1989–96. The study includes cases from
both district courts and the Federal Circuit over this period. The study examines both the
outcome on validity of the litigated patents and the success rate of the validity challenges.
The study looked at all bases for invalidity challenges, including obviousness and novelty.
Allison and Lemley (1998) find that 162 (54 percent) of the patents were found valid,
while 138 (46 percent) were found invalid. Of these decisions, the largest number of
invalidity determinations (58) were made on the basis of obviousness—42 percent of all
cases finding invalidity, and the third largest number (37) were made on the basis of
novelty—26.8 percent. Together, these grounds for invalidity accounted for 138 out of 191
total determinations of invalidity. These were also the most popular grounds for asserted
invalidity over the period studied—160 assertions of obviousness and invalidity and 91
under novelty.
Allison and Lemley (1998) calculate the likelihood of success on these invalidity issues.
Regarding invalidity based on lack of novelty, 37 of the 91 decisions on novelty found
the patent invalid (40.7 percent). On non-obviousness, 58 of the 160 decisions found the
patent obvious (36.3 percent).
Allison and Lemley (1998) identify the following two findings in their data as

“ interesting”: first, that patents are only slightly more likely to be found valid than
invalid; and second, that challenges to validity based on novelty are more successful then
obviousness challenges.
1.2. John Allison, Mark Lemley, and David Schwartz, “Understanding the Realties of
Modern Patent Litigation” (2014) Allison et. al (2014) expand on the earlier Allison
and Lemley (1998) study. Allison et. al. (2014) code every substantive decision in all
district court patent cases filed in 2008 and 2009. Regarding novelty and obviousness,
Allison et. al. (2014) find 154 motions for summary judgment of invalidity based on lack
of novelty, with 31 (20 percent) being successful. There were 149 motions for summary
judgment of invalidity of obviousness, with 31 finding invalidity (20 percent). The overall
invalidity win rates were slightly higher, with 31.1 percent of the novelty challenges being
successful and 27.8 percent of the obviousness challenges being successful.
Allison et. al. note that challenges to validity on the basis of obviousness and lack
of novelty have decreased over the years. And most motions for invalidity, including
those based on obviousness and lack of novelty, usually fail. Finally, the validity rate is
essentially the same as observed by Allison and Lemley (1998)—43 percent as compared
to the earlier observed 46 percent.
1.3. Brian J. Love and Shawn Ambwani, “Inter Partes Review: An Early Look at the
Number” (2014) Love and Ambwani (2014) examine the IPR mechanism instiuted
under the America Invents Act (AIA). They collect data on IPRs filed over an almost
two-year period, starting from the beginning of the AIA, which includes 979 petitions.
They also track co-pending litigations involving the patents under IPR review. Love and
Ambwani (2014) find that 84 percent of requests for IPR are granted and if the IPR
reaches a final decision, 77 percent of those decisions invalidate or disclaim all claims
under review. Love and Ambwani (2014) note that “it is still too early to draw a firm
conclusion about IPR’s impact.” Further study—including examining data available from
the USPTO on IPRs, available at www.uspto.gov—can be done.
B. Novelty-Specific Studies
1. Dennis D. Crouch, “Is Novelty Obsolete? Chronicling the Irrelevance of the Invention
Date in U.S. Patent Law” (2009)
Crouch (2009) investigates the usage of the novelty requirement, specifically pre-AIA
§ 102(a), and whether patent applicants attempted to “swear behind” or allege an earlier
date of invention prior to filing. Crouch (2009) uses three sources in his empirical exami-
nation of § 102(a):
the prosecution history files of 21,000+ patent applications filed in the past decade; a survey of
1,000+ patent practitioners regarding their use of the novelty provisions of the Patent Act; and
a collection of 11,000,000+ prior art references cited in recently-issued patents.
Crouch (2009) finds that during prosecution, most patent applicants contend with
§ 102(a) prior art that could be overcome by claiming an invention date prior to filing, yet
most patentees do not make such a claim (fewer than 0.1 percent of the cases studied).
The result is most likely driven by the fact that it is administratively difficult to prove an

earlier date of invention, given that 77 percent of attempts to antedate references were
rejected by the USPTO. All of this suggests that, even prior to the AIA, the USPTO was
operating under a de facto first-to-file system.
2. David S. Abrams and R. Polk Wagner, “Poisoning the Next Apple? The America
Invents Act and Individual Inventors” (2013)
Abrams and Wagner (2013), in an attempt to project the impact of the adoption of the
AIA and a first-to-file system, examine the impact of an earlier Canadian law change
to first-to-file and compared it to pre-AIA activity in the United States. Abrams and
Wagner use a difference-in-difference analysis over a ten-year period (1984–93), focusing
particularly on smaller inventors. They find a significant drop in the share of patent grants
to individual inventors in Canada coinciding with the implementation of a first-to-file
system. These results occurred while there was no measurable change in patent quality in
Canada or alternative explanations. Figure 12.1 reports their results.
Abrams and Wagner (2013) conclude that these results suggest that the move to a
first-to-file system under the AIA may reduce individual inventor patenting in the United
States in a similar way as occurred in Canada.
C. Non-obviousness (35 U.S.C. § 103)
These studies are divided between those focused on non-obviousness decisions prior
to the Supreme Court’s non-obviousness decision in KSR v. Teleflex and those looking
at cases after. This division occurred in the literature because researchers attempted to
0.2 Canadian Grants U.S. Grants

Individual Inventor Share
0.15
0.1
0.05
1/1/1984 1/1/1986 1/1/1988 1/1/1990 1/1/1992 1/1/1994
Date
Source: Abrams and Wagner (2013), “Poisoning the Next Apple? The America Invents Act and Individual
Inventors”, Table 4.
Figure 12.1 Individual inventor share in U.S. and Canadian patent grants

quantify the impact the KSR decision had on both the approach to, and the results on,
questions of non-obviousness.
1. Results in non-obviousness decisions prior to KSR v. Teleflex
1.1. Glynn S. Lunney, Jr., “E-Obviousness” (2001) Lunney (2001) empirically studies
Federal Circuit decisions on invalidity, particularly focusing on non-obviousness, in eight
time periods spread over a 50-year window. Six of these periods were prior to the Federal
Circuit’s creation and two were after. For these periods, Lunney (2001) uses the third
empirical approach detailed above—determining the percentage of cases where invalidity
of a patent was found, when invalidity was at issue. Of those patents that were found
invalid, the percentage of such patents where invalidity was based on obviousness is then
determined.
Lunney (2001) finds that the invalidity rate (i.e., percentage of patents held invalid
where validity was at issue and decided) fell from 50 percent in 1981–82, just prior to the
Federal Circuit’s creation, to 39.34 percent in 1984–85, shortly after the Federal Circuit’s
creation, and then to 25 percent in 1994–95. And of these invalidity findings, where
invalidity was at issue and decided, a great majority (66.67 to 79.49 percent) were based on
a finding of obviousness prior to the Federal Circuit’s creation, while after only 20 percent
were due to obviousness. The full results of these later findings are show in Figure 12.2.
Lunney (2001) concludes that the data “demonstrates that, among the available bases
for challenging a patent’s validity, obviousness has become particularly disfavored.” While
noting that “we must be careful in drawing inferences from an examination of appellate
case results along, these results strongly suggest that obviousness is much less central to
appellate determinations of patent validity under the Federal Circuit.”
100
80
60
40
20
0
6
5
94
95
96
96
97
98
98
99
–1
–1
–1
–1
–1
–1
–1
–1
44
55
64
66
75
81
84
94
19
19
19
19
19
19
19
19
Source: Lunney, Jr (2001), “E-Obviousness”, Figure 2.
Figure 12.2 Percentage of invalid patents found invalid for obviousness

0.8 75.9%
0.6
34.3%
0.4
0.2 Pre-Federal Circuit Post-Federal Circuit

0
6
1
94
95
96
96
97
98
98
98
98
99
99
99
99
99
00
–1
–1
–1
–1
–1
–1
–1
–1
–1
–1
–1
–1
–1
–1
–2
44
55
64
66
75
81
84
86
88
90
92
94
96
98
00
19
19
19
19
19
19
19
19
19
19
19
19
19
19
20
Source: Lunney (2004), “Patent Law, the Federal Circuit and the Supreme Court: A Quiet Revolution”,
Figure 2.
Figure 12.3 Percentage of failure results due to ruling that patent was invalid or
otherwise unenforceable
1.2. Glynn S. Lunney, “Patent Law, the Federal Circuit and the Supreme Court: A Quiet
Revolution” (2004) Lunney (2004) further explores the Federal Circuit’s decisions on
invalidity in comparison to the appellate decisions prior to the court’s creation. Lunney
(2004) expands the periods studied and observes that “patent invalidity is significantly less
likely to be the reason why a claim of patent infringement fails under the Federal Circuit.”
Figure 12.3 shows that after the Federal Circuit’s creation, invalidity accounted for just
one-third of the failure rate for patentees at the Federal Circuit, in contrast to it being
about 75 percent of the reason for failure prior.
Lunney (2004) states that the reason for this drop is the “sharply reduced role for the
nonobviousness requirement” in the post-Federal Circuit era. Lunney 2004 observes that
prior to the Federal Circuit’s creation, a patent’s obviousness accounted for 64.8 percent
of the invalidity decisions, while after, obviousness accounted for only 14.6 percent of
such determinations.
1.3. Sean M. McEldowney, “New Insights on the ‘Death’ of Obviousness: An Empirical

Study of District Court Obviousness Opinions” (2006) McEldowney (2006) studies
non-obviousness decisions by district courts before KSR. Specifically, the study examines
321 published district court opinions that reached the question of obviousness on 407
different utility patents. McEldowney (2006) includes these non-obviousness opinions for
two different time periods: the first (1970–75) before the Federal Circuit’s creation and the
second (1995–2000) after the Federal Circuit’s creation.
McEldowney (2006) finds a decrease in the 1990s of both district courts reaching the
issue of non-obviousness and finding patents obvious once the issue is reached. Table 12.6
provides detailed information on this finding.
McEldowney (2006) observes that in light of this study, and the conclusions of other
such as Lunney (2001), “the overwhelming conclusion is that, under the Federal Circuit’s
reign, courts are less likely to invalidate patents as obvious when faced with the issue.”
McEldowney (2006) indicates that these findings may “prove the existence of a pro-patent
bias in the Federal Circuit” that has seeped down into the district courts.

Table 12.6 Number of patent suits filed and number of suits reaching obviousness,
1970–75 and 1995–2000
Year Total Patent Patents % of Patents Suits Reaching % of Suits

Suits Filed Reaching Reaching Obviousness Reaching
Obviousness Obviousness Obviousness
1970 1020 64 6.3% 50 4.9%
1971 920 27 2.9% 25 2.7%
1972 900 32 3.6% 24 2.7%
1973 800 37 4.6% 28 3.5%
1974 930 44 4.7% 34 3.7%
1975 900 30 3.3% 27 3.0%
1970s Totals 5470 234 4.3% 188 3.4%
1995 1600 35 2.2% 28 1.8%
1996 1840 25 1.4% 20 1.1%
1997 2112 33 1.6% 22 1.0%
1998 2218 33 1.5% 25 1.1%
1999 2318 27 1.2% 21 1.9%
2000 2484 20 0.8% 17 0.7%
1990s Totals 12 572 173 1.4% 133 1.1%
Source: McEldowney (2006), “New Insights on the ‘Death’ of Obviousness: An Empirical Study of District
Court Obviousness Opinions”, Table 3.
McEldowney (2006) also examines the effect of technological complexity on the rates of
obviousness. The study divides the non-obviousness decisions into levels of complexity
and calculates the rate of obviousness (the number of non-obviousness determinations
that result in a finding of obvious) for two time periods: the 1970s and 1990s. In the
1970s, McEldowney (2006) finds little variation in rate of obviousness amongst the four
complexity categories, from category 1 (very simple) with a 58 percent rate of obvious-
ness, to category 4 (very complex) with a 50 percent rate. In the 1990s, the rate does vary
depending on technological complexity, with category 1 with a 39 percent rate, category 2
(simple) with a 51 percent rate, category 3 (complex) with a 27 percent rate, and category
4 with a 16 percent rate. The rate of obviousness drops between the 1970s and 1990s for
all categories but category 2, simple technologies.
1.4. Christopher A. Cotropia, “Nonobviousness and the Federal Circuit: An Empirical

Analysis of Recent Case Law” (2007) Cotropia (2007) also examines Federal Circuit
decisions on invalidity, specifically decisions regarding non-obviousness, and comes to a
different conclusion than Lunney (2001, 2004). Cotropia (2007) studies all Federal Circuit
decisions on non-obviousness over a four-year period—from 2002–05—that included 102
determinations on non-obviousness. Cotropia (2007) took the first and second approaches
to studying these decisions, looking at both the court’s ultimate finding on non-obviousness
and the court’s disposition on the lower tribunal’s decision (affirmance, reversal, or vacation).
Cotropia (2007) finds that of the 67 final determinations on non-obviousness, 37
(56.06 percent) found the patent non-obvious, while 29 (43.93 percent) found the patent

obvious. When focusing on the court’s disposition in relation to the lower tribunal, 62.5
percent (30) of the non-obvious determinations were affirmed, while 6.25 percent (three)
were reversed and 31.25 percent (15) were vacated. In comparison, Cotropia (2007)
finds that 48.15 percent (26) of the obvious determinations were affirmed, while 12.96
percent (seven) were reversed and 38.89 percent (21) were vacated. Including all of the
non-obviousness decisions over the period, Cotropia (2007) finds that “[t]he percentage
of patents found nonobvious is only slightly higher than those found obvious—36.27%
compared to 28.43%.” Cotropia (2007) could not find a significant difference between
the Federal Circuit’s handling of a lower court’s finding of non-obvious as compared to
a finding of obvious.
When investigating appeals from the USPTO, Cotropia (2007) observes 86.79 percent
(46) obvious determinations were affirmed, while 5.66 percent (three) were reversed
and 7.55 percent (four) were vacated. Cotropia (2007) notes that the obviousness rate
and affirmance rate on appeals from the USPTO are both statistically significant and
extremely high.
Cotropia (2007) concludes that:
These results suggest that the court has not lowered the nonobviousness requirement and may
be maintaining a higher nonobviousness requirement in the case of the USPTO. But this conclu-
sion is greatly tempered by the fact that the cases studied may all include litigants and patent
applicants that have already taken into account a lowered nonobviousness standard. Thus, any
inference into the strength of the nonobviousness requirement from this data is weak at best.
Cotropia (2007) recognizes that the results are impacted by litigants’ selection of which
cases to litigate and appeal to the Federal Circuit, particularly on non-obviousness.
Cotropia (2007) concludes that “this line of reasoning does not fully discount the
significance of the macro-level studies results,” for four reasons. “First, for this line of
reasoning to hold and a low nonobviousness standard to exist, all of the patents in the
study must hover around this lower nonobviousness standard and are, as a result, invalid
under the ‘correct’ standard,” and this simply cannot be the case. “Second, this critique
rests on the assumption that litigants have a good understanding of exactly where the
nonobviousness bar sits,” which is unlikely at best. Third, “if the Federal Circuit had
lowered the nonobviousness bar, and these results simply record litigants’ understanding
of this lowering, the Federal Circuit would have to communicate the lowering in some
way,” and finding as many patents obvious as non-obvious does not send that message.
And finally, Cotropia (2007) concludes that the results of the study, at the very least,
discount any empirical proof to support critics who argue that the Federal Circuit has
lowered the non-obviousness requirement.
1.5. Lee Petherbridge and R. Polk Wagner, “The Federal Circuit and Patentability: An
Empirical Assessment of the Law of Obviousness” (2007) Petherbridge and Wagner
(2007) perform content analysis on Federal Circuit written opinions concerning non-
obviousness issued over a 15-year period from 1990–2005. Petherbridge and Wagner
(2007) include 480 Federal Circuit analysis of non-obviousness. In addition to focusing
on the specific standard used to determine non-obviousness, Petherbridge and Wagner
(2007) take the first and second empirical approach noted above and report data regarding
the Federal Circuit’s disposition of non-obviousness decisions.

Petherbridge and Wagner (2007) find that 42.2 percent (178) of decisions result in
a non-obvious finding, compared to 57.8 percent (244) of decisions of obviousness.
Petherbridge and Wagner (2007) also report a high affirmance rate of decisions on non-
obviousness, with 312 (65.0 percent) decisions affirmed compared to 110 (22.9 percent)
reversed and 57 (11.9 percent) vacated. These results are not broken down by the lower
court’s decision—obvious or non-obvious—that was being appealed.
Petherbridge and Wagner (2007) note that a “clear majority of the analyses” find a
patent obvious. Based on this data, the study observes that “the Federal Circuit seems
to have little problem finding claims obvious.” Furthermore, the rate of reversal is much
lower than the reversal rate across all patent cases (47.3 percent), suggesting that doctrinal
stability in this area “is no worse (and may well be better) than many other issues in
patent law.” The study concludes that the non-obviousness doctrine is “relatively clear
and stable.”
Petherbridge and Wagner (2007) also find little variation in non-obviousness determi-
nations divided by technologies. The study:
discovered a nearly uniform distribution of obvious and nonobvious outcomes between the
mechanical (57.1% obvious/42.9% nonobvious), chemical (58.7% obvious/41.3% nonobvious),
and biotechnological arts (58.3% obvious/41.7% nonobvious). The electronic arts differed
somewhat, showing 46.9% obvious outcomes and 53.1% nonobvious outcomes.
The results are “consistent with the overall frequency of an obvious disposition” in the
Petherbridge and Wagner (2007) Federal Circuit decision dataset, “which is 57.8%.”
2. Results in non-obviousness decisions after KSR v. Teleflex
2.1. Jennifer Nock and Sreekar Gadde, “Raising the Bar for Nonobviousness: An
Empirical Study of Federal Circuit Case Law Following KSR” (2010) Nock and
Gadde (2010) study all Federal Circuit decisions issued in the immediate two-and-a-half
years following the Supreme Court’s KSR decision. Nock and Gadde (2010) include
65 Federal Circuit decisions reaching the question of non-obviousness after KSR. The
study takes the first and second approach to evaluating the Federal Circuit’s handling
of non-obviousness—looking at both the ultimate outcome of the court’s decision
(non-obvious or obvious) and the disposition relative to the lower tribunal (affirmed,
reversed, or vacated).
Nock and Gadde (2010) observe 21 decisions finding the patent(s) non-obvious (32
percent), while 41 find the patent(s) obvious (63.1 percent). The Federal Circuit did not
reverse any lower-court findings of obviousness and affirmed 81 percent (30 of 37) of the
lower-court decisions of obviousness. In contrast, 53 percent (21 of 40) of lower-court
decisions of non-obviousness were affirmed. Figure 12.4 illustrates these results.
Nock and Gadde (2010) also study the Federal Circuit’s handling of appeals from the
BPAI after KSR. Notably, the study finds that the court affirms 83 percent (15) of appeals
from BPAI that found the patent obvious, reverses none, and vacates 17 percent (3).
Nock and Gadde (2010) conclude that:
These findings demonstrate that findings of obvious are unlikely to be overturned on appeal, and
similarly, inventions are more likely to be held obvious than nonobvious by the Federal Circuit.

81.1%
52.5% Affirm
Reverse
Vacate
27.5%
20.0% 18.9%
0.0%
Non-Obvious Obvious
Source: Nock and Gadde (2010), “Raising the Bar for Nonobviousness: An Empirical Study of Federal
Circuit Case Law Following KSR”, Figure 1.
Figure 12.4 Federal Circuit disposition of cases post-KSR, by outcome below
Those who were concerned that patentability standards had grown too lax under the TSM test
should be satisfied that the Federal Circuit is now less likely to uphold patents on innovations
that tread close to the prior art.
2.2. Glynn S. Lunney, Jr. and Christian Johnson, “Not So Obvious After All: Patent Law’s
Nonobviousness Requirement, KSR, and the Fear of Hindsight Bias” (2012) Lunney and
Johnson (2012) also examine non-obviousness decisions at the Federal Circuit after KSR.
Lunney and Johnson (2012) study Federal Circuit decisions on invalidity immediately
prior to KSR (2004–06) and immediate thereafter (2008–10). Lunney and Johnson (2012)
take the same approach as Lunney (2001) and Lunney (2004): determining the percentage
of invalidations by the Federal Circuit based on a finding of obviousness. Figure 12.5
reports these results.
Lunney and Johnson (2012) find:
some recovery in the importance of the nonobviousness doctrine following KSR, as one would
expect. From 2008 through 2010, obviousness accounted for just less than 20% of the losses
patentees experienced. Compared to the immediately preceding period of 2004 through 2006,
this represents a healthy rebound. Yet, despite the bump, obviousness today accounts for only a
small fraction of the losses for which it accounted in the pre-Federal Circuit era and is no longer
the dominant doctrine it once was.
Lunney and Johnson (2012) expressly criticize the approaches taken in Cotropia (2007)
and Petherbridge and Wagner (2007) for not taking into account selection bias. As
Lunney and Johnson (2012) note, “[w]hen we use appellate resolutions as our data set,
we are observing only those patent disputes that parties have chosen to litigate rather
than settle. . . . As a practical matter, these selection effects tend to lead to an equilibrium
success rate in decided cases.”
Thus, any inference from the ratio of obvious to non-obvious outcomes or the affirm-
ance and reversal of either is irrelevant because, as Lunney and Johnson (2012) explain,
“the average success rate merely reflects the underlying considerations that lead parties to

80%
70%
60%
50%
Invalidity
40%
Obviousness
30%
20%
10%
0%
FC
0
00
00
01
e-
–2
–2
–2
Pr
84
04
08
19
20
20
Source: Lunney and Johnson (2012), “Not So Obvious After All: Patent Law’s Nonobviousness
Requirement, KSR, and the Fear of Hindsight Bias”, Figure 1.
Figure 12.5 Reasons for losing: percentage of losses due to invalidity or unenforceability

of patent and due to obviousness
select a case for resolution through litigation rather than settlement.” This is why looking
at the percentage of obviousness challenges amongst invalidity challenges, and the suc-
cess of such challenges, provides more insight into the strength of the nonobviousness
doctrine. “If a court weakens the nonobviousness requirement, making obviousness much
harder for a defendant to prove, then defendants will emphasize other defenses, such as
noninfringement, and we should see obviousness account for fewer patent-holder losses
accordingly.”
2.3. Jason Rantanen, “The Federal Circuit’s New Obviousness Jurisprudence: An

Empirical Study” (2013) Rantanen (2013) also studies Federal Circuit non-obviousness
decisions after KSR. Rantanen (2013) examines all Federal Circuit decisions five years
before and five years after KSR—from April 30, 2007 to April 30, 2012. During this
period, Rantanen (2013) found 583 distinct obviousness analyses by the Federal Circuit.
Rantanen (2013) took both the first and second approach to analyzing the Federal
Circuit’s non-obviousness determinations.
Table 12.7 reports the outcome of the Federal Circuit analysis in the study and shows
a statistically significant increase in obvious findings after KSR.
Rantanen (2013) also examines the Federal Circuit’s affirmance, reversal, and vacation
rates for the different dispositions of the lower tribunals on non-obviousness. Rantanen
(2013) finds that the higher rate of obviousness findings by the court is due to the greater
deference granted to district court findings of obviousness. Table 12.8 reports these
findings.
Rantanen (2013) concludes that:

Table 12.7 Federal Circuit determinations of obviousness before and after KSR
Nonobvious Obvious
Appeals arising from District Courts and ITC
Pre-Certiorari 92 (54%) 77 (46%)
Post-Opinion 57 (43%) 75 (57%)
Expected if no change 72 60
Appeals arising from PTO
Pre-Certiorari 20 (17%) 98 (83%)
Post-Opinion 4 (4%) 93 (96%)
Expected if no change 16 81
Notes:
Chi-square for actual post-opinion versus expected p<0.01 for appeals arising from District Courts and ITC.
Chi-square for actual post-opinion versus expected p<0.01 for appeals arising from the PTO.
Source: Rantanen (2013), “The Federal Circuit’s New Obviousness Jurisprudence: An Empirical Study”, Table 1.
Table 12.8 Federal Court dispositions of district court and ITC findings of non-obvious
and obvious
Pre-certiorari post-opinion
Lower tribunal finding of nonobvious
Reversed 15 (15%) 18 (23%)
Affirmed 71 (72%) 54 (68%)
Vacated 13 (13%) 7 (9%)
Lower tribunal finding of obvious
Reversed 24 (23%) 6 (9%)
Affirmed 64 (60%) 57 (81%)
Vacated 18 (17%) 7 (10%)
Source: Rantanen (2013), “The Federal Circuit’s New Obviousness Jurisprudence: An Empirical Study”,
Table 4.
the pattern of the Federal Circuit’s affirmances since KSR does not evince a substantively changed
obviousness jurisprudence; instead, it suggests a jurisprudence in which the Federal Circuit is
giving greater deference when lower tribunals reach conclusions that patents are obvious.
Rantanen (2013) also examines non-obviousness determinations at the Federal Circuit by

technology type both before and after KSR. The study finds that the rate “at which the
Federal Circuit found inventions to be obvious went up in all areas of technology, with the
largest for electronic inventions.” Table 12.9 shows that in the ten years prior to KSR, “the
Federal Circuit found 21 out of 45 electronic inventions to be obvious. In half that much time
since KSR, the Federal Circuit has found 37 out of 58 electronic inventions to be obvious.”
Based on all of these studies, it appears that non-obviousness determinations are not
technology dependent, with the determinations rates staying relatively constant across
technologies.

Table 12.9 Technologies analyzed for obviousness by the Federal Circuit pre- and post-
KSR result in appeals from district courts and ITC
Nonobvious Obvious
Pre-Certiorari
Biological 3 (75%) 1 (25%)
Chemical 14 (61%) 9 (39%)
Electronic 24 (53%) 21 (47%)
Mechanical 51 (53%) 46 (47%)
Post-Opinion
Biological 4 (67%) 2 (33%)
Chemical 14 (52%) 13 (48%)
Electronic 21 (36%) 37 (64%)
Mechanical 18 (44%) 23 (56%)
Table 6.
3. Non-obviousness standard as applied by the courts

Another area of empirical research on non-obviousness focuses on the content of the
jurisprudence itself. That is, the studies look to determine the actual legal standards
used by courts, in most instances the Federal Circuit, to reach their validity determina-
tions and whether such standards negate any “hindsight bias” by the decision-maker.
Some studies go further to investigate whether there is any correlation between the
legal standards used and the outcome reached. Many of these studies focus on the
non-obviousness doctrine, specifically the Federal Circuit application of the teaching,
suggestion, motivation (TSM) test prior to KSR and that court’s application of KSR
after the Supreme Court’s decision.
A majority of the studies performing content analysis on non-obviousness determina-
tions did so to determine whether the Federal Circuit’s TSM test and the Supreme Court’s
reason to combine standards in KSR impact the outcome in a given case. A group of
studies—Cotropia (2006), Cotropia (2007), and Petherbridge and Wagner (2007)—focus
on Federal Circuit case law just prior to KSR; while other studies—Rantanen (2013) and
Cotropia (2014)—examine the Federal Circuit’s implementation of KSR.
3.1. TSM and non-obviousness analysis prior to KSR Cotropia (2006) studies Federal
Circuit non-obviousness case law to determine the degree of flexibility in the TSM test.
Cotropia (2006) examines written Federal Circuit non-obviousness opinions over a three-
year period—August 2002 to September 2005—prior to the Supreme Court’s decision in
KSR. Of the 69 non-obviousness opinions during this period, Cotropia (2006) observes a
full discussion of TSM in only 12 cases. Twenty of the cases do not mention TSM at all,
while 37 are presented with only prior art-based TSM evidence.
Of the cases in which the Federal Circuit could choose amongst the sources of TSM,
Cotropia (2006) observes that the court focuses only on the prior art—what Cotropia
(2006) calls the “narrow suggestion test”—when either the undocumented evidence of
TSM is not detailed enough or the complexity of the technology involved is high. If the

technology is less complex and/or the undocumented evidence of TSM is very detailed, a
“broad”, or flexible, TSM test is used.
Cotropia (2006) concludes that the empirical findings detail a flexible TSM test that
adapts to the detail of the TSM evidence and the complexity of the technology at issue.
Cotropia (2007) also examines the usage of TSM by the Federal Circuit prior to KSR.
Cotropia (2007) finds that TSM is relevant to the non-obviousness determination in only
25 patents in the population studied, while 54 patents’ non-obviousness was determined
regardless of the presence of TSM. Specifically, failure to find a TSM results in a finding
of non-obviousness, or a vacation of a finding of obviousness, in 32.91 percent of patents
appealed from patent infringement cases and only 11.1 percent of patents appealed from
the USPTO.
Cotropia (2007) concludes that TSM “plays a fairly small role in the [Federal Circuit’s]
nonobviousness jurisprudence.”
Petherbridge and Wagner (2007), in the content analysis they perform over a larger
population of Federal Circuit cases, observe similar results. “TSM analysis is used less
than half of the time (45%) in obviousness analyses by the Federal Circuit,” and this
analysis has no “observable effect on whether the reviewed analysis is affirmed and only
a modest impact (about 5%) on whether the patent is declared obvious.”
And while the rate of TSM application increases over time in Petherbridge and
Wagner’s (2007) dataset, the rate of finding patents non-obvious drops. The study also
finds that the number of sources for TSM—from prior art to nature of the problem being
solved—increases over time as well. These results, shown partially in Figure 12.6, “suggest
that TSM is not a great impediment to establishing obviousness. To the contrary, the more
Obviousness Analyses of the Federal Circuit, 1990–2005

[18-analysis legged average, n = 187]
100%
90%
Percentage of nonobvious results
80%
70%
60%
50%
40%
30%
20%
10%
0%
10 20 30 40 50 60 70 80 90 100 110 120 130 140 150 160 170 180
analysis number (n = 187)
Source: Petherbridge and Wagner (2007), “The Federal Circuit and Patentability: An Empirical Assessment
of the Law of Obviousness”, Figure 7.
Figure 12.6 Rate of non-obviousness when TSM applies

TSM appeared in Federal Circuit analyses, the less likely it was that the outcome of the
analysis was nonobvious.”
3.2. TSM and non-obviousness analysis after KSR Rantanen (2013) studies the usage
and impact of TSM and other “reason to combine” articulations by the Federal Circuit
both before and after KSR. Rantanen 2013 finds that the usage of TSM in Federal Circuit
opinions decreases after KSR and “lives on in diminished form” with a required “reason to
combine” or “reason to modify” the prior art that can be found in such things as common
sense or predictable uses of the prior art. Prior to KSR, the Federal Circuit used a form of
TSM in 58 percent of its opinions. After KSR, that usage fell to 29 percent, and the court’s
formal usage of TSM has essentially ceased. This decline is shown in detail in Figure 12.7,
with KSR being issued prior to Federal Circuit non-obviousness Opinion 258.
Rantanen (2013) next examines the Federal Circuit’s usage of a “reason to combine”
test for non-obviousness after KSR. The study finds a much more flexible non-obvious-
ness analysis, with multiple types of “reasons to combine” articulated in Federal Circuit
decisions. The distribution of these tests for non-obviousness is reported in Table 12.10,
broken down by type and both prior to certiorari being granted in KSR and after the
KSR decision.
In “Predictability and Nonobviousness in Patent Law After KSR” (2014), Cotropia
(2014) studies USPTO BPAI decisions, as well as district court and Federal Circuit
90.0%
80.0%
70.0%
60.0%
Percentage of Opinions
50.0%
40.0%
30.0%
20.0%
10.0%
0%
10
41
52
63
74
85
96
107
118
129
140
151
162
173
184
195
206
217
228
239
250
261
272
283
294
305
316
327
338
349
360
371
382
393
404
TSM Use (Lagged)
Figure 2.
Figure 12.7 Use of TSM framework in Federal Circuit obviousness opinions, 1996–2012

Table 12.10 Type of “reason to combine” articulation
Total Opinions Prior art only Prior art, PHOSITA, Particular sources
or problem not specified
Pre-Certiorari 241 14 (6%) 50 (21%) 79 (33%)
Post-Opinion 147 0 (0%) 4 (3%) 71 (48%)
Table 7.
ecisions, to ascertain their usage of KSR and how KSR has changed the non-obvious-
d
ness. Specifically, Cotropia (2014) studied PTAB decisions both before KSR and after
to determine the PTAB’s usage of predictability in non-obviousness decisions. Cotropia
(2014) finds that:
Prior to KSR, 110 of the 8183 nonobviousness decisions used the phrase “predictable” or some
variation thereof. Fifty-two of these used the phrase “predictable results” or some derivation.
And, during the pre-KSR period, there were no precedential decisions that focused on using
predictability in determining nonobviousness. In comparison, after KSR, the PTAB’s use of the
predictability terminology is quite common. As of August 27, 2012, 6455 of 31,414 nonobvious-
ness decisions invoked the term. More than three fourths, 4954 cases, used the phrase “predict-
able results” or some derivation. This is a change from 1.3% to 20.5% of the nonobviousness
determinations that invoked the term “predictability” and a change from 0.6% to 15.8% in use
of the phrase “predictable results.”
4. Hindsight bias in non-obviousness decision-making
4.1. Gregory N. Mandel, “Patently Non-Obvious: Empirical Demonstration that the

Hindsight Bias Renders Patent Decisions Irrational” (2006) and Gregory Mandel,
“Patently Non-Obvious II: Experimental Study on the Hindsight Issue Before the Supreme
Court in KSR v. Teleflex” (2007) In two papers, Mandel (2006, 2007) investigates the
extent and impact of hindsight bias on non-obviousness determinations. Both studies
performed original experiments to test the extent of hindsight bias and the effect to which
different legal standards impact such bias. Both studies gave participant mock jurors
hypothetical fact scenarios concerning an invention. The scenarios were formulated to
measure the impact of hindsight and determine whether the legal standard for determin-
ing non-obviousness mitigates that impact.
Mandel (2006) presented 247 first-year law students with two scenarios (one involving
a baseball bat and the other a fishing lure), including background regarding the field of
art of the invention, a variety of prior art references, and a description of the problem the
inventor was trying to solve. Each participant was asked three questions: (1) whether, in
light of the prior art and information provided in the scenario, a solution to the problem
was obvious to a person with ordinary skill in the relevant field; (2) the confidence the
respondent had in his or her answer to the obvious query (answered on a scale from 0
percent to 100 percent, with answers indicated in 10 percent increments); and (3) the
likelihood that the inventor in the scenario would achieve the invention (answered on a
scale from l—“not at all likely”—to 7—“extremely likely”).

Table 12.11 Participant responses
n Obvious Confidence Likelihood

(frequency) (mean) (mean)
Baseball foresight 42 10 (24%) 65.48 4.40
hindsight 42a 32 (76%) 75.71 5.41
debiasing 41 27 (66%) 70.98 5.00
Fishing Lure foresight 40 9 (23%) 72.00 4.05
hindsight 41 24 (59%) 66.34 4.66
debiasing 41 20 (49%) 67.32 4.49
Notes: a One baseball hindsight participaint did not answer the likelihood question.
Source: Mandel (2006), “Patently Non-Obvious: Empirical Demonstration that the Hindsight Bias Renders
Patent Decisions Irrational”, Table 1.
The experiment used a between-subjects experimental design, and each participant received
only one of three different conditions—(1) a foresight condition including all lead-up
conditions with the scenario character trying to solve the problem; (2) a hindsight condition
identical to the foresight condition with one additional sentence saying the scenario charac-
ter invented the solution and what the solution was; and (3) a debiasing condition identical
to the hindsight scenario including a statement from the Model Patent Jury Instructions
that warned the participant about using hindsight in answering the above questions.
The results from the experiment are set out in Table 12.11.
Mandel (2006) finds the results demonstrate that the hindsight bias significantly
influences non-obvious judgments. Participants who were not informed of the invention
were substantially more likely to judge a solution non-obvious than participants who
were informed what the invention was. Mandel (2006) concludes that the magnitude of
the hindsight bias in these patent scenarios is striking and is greater than that reported
for other legal judgments. Ex post knowledge of invention deeply affected participants’
conclusions regarding whether an invention was non-obvious ex ante. Equally significant,
Mandel (2006) finds, is that debiasing instructions based on actual model jury instructions
did not ameliorate the hindsight bias.
Mandel (2007) performed a similar experiment to that described in Mandel 2006,
this time using 384 general members of the public, who were eligible for jury service, as
participants. The experiment used facts from an actual, decided non-obviousness case,
and the participants were provided with similar scenario facts as those presented in
Mandel (2006). A between-subjects experiment was again used. There were two versions
of the scenario: one where the prior art included a TSM and one where it did not. Mandel
(2007) also presented participants with one of four conditions—(1) a foresight condition,
(2) a hindsight condition, (3) a hindsight condition and an instruction informing jurors
to follow the TSM test, and (4) a hindsight condition and an instruction on the Graham
factors. The same three questions asked in Mandel’s first study (2006) were asked to
participants in Mandel’s second study (2007).
The results are set out in Table 12.12.
Mandel’s second study (2007) confirms the findings of his first (2006)—that a hindsight

Table 12.12 Participant responses
Suggestion in Instruction n Obvious Confidence Likelihood

scenario (freq.) (mean) (mean)
Foresight no none 55 23 (42%) 68.73 4.78
yes none 55a 27 (49%) 71.00 4.94
Hindsight no none 55 39 (71%) 74.55 5.38
yes none 53 45 (85%) 74.91 5.77
no suggestion 55b 45 (83%) 71.45 5.56
yes suggestion 56a 42 (75%) 71.79 5.42
no Graham 55 43 (78%) 74.91 5.65
Notes:
a One participant in each of these conditions did not answer the likelihood question.
b One participant in this condition did not answer the obvious question.
Source: Mandel (2007), “Patently Non-Obvious II: Experimental Study on the Hindsight Issue Before the
Supreme Court in KSR v. Teleflex”, Table 1.
condition influenced participants’ decisions. In 2007, Mandel also finds that instructions
on the TSM test or the Graham factors did not influence participants’ ultimate conclu-
sions on non-obviousness.
D. Disclosure Requirements (35 U.S.C. § 112)
1. John R. Allison and Lisa Larrimore Ouellette, “How Courts Adjudicate Patent
Definiteness and Disclosure” (2015)
Allison and Ouellette (2015) empirically examine determinations made under three
§ 112 patentability requirements—enablement, written description, and indefiniteness.
They code 1144 decisions, at both the Federal Circuit and district court level, on these
validity challenges. In performing the coding, they use “a novel five-level scale so as to
capture significant subtlety in the strength of each decision,” and also code the technol-
ogy and industry of the patent subject to the § 112 challenge and a variety of litigation
characteristics.
Allison and Ouellette (2015) find that technology and industry do impact § 112 deci-
sions, “although fewer than the conventional wisdom suggests, and not always in the
direction that many have believed.” In their ordered logistic regression models (taking
other litigation characteristics into account), they find that “only electronics patents
performing better than those in other technology fields on written description and
enablement, and mechanical patents scoring better on enablement. Technology spectrum,
software patents other than business methods fared poorly on enablement.” With regard
to industry, their regression models show non-Abbreviated New Drug Application
pharmaceutical patents performed poorly under the written description and definiteness
requirements, while patents in the semiconductor industry were significantly more likely
to survive on written description grounds.
Regarding non-technology and industry factors, Allison and Ouellette (2015) observe
a claim was more likely to fall on indefiniteness grounds when the challenged claim

element was drafted in means-plus-function format. Similar results were found regarding
indefiniteness challenges in pre-Markman decisions. They also observe that district courts
as a group were significantly more likely than the Federal Circuit to uphold the validity of
patents on enablement and definiteness grounds in both technology and industry models.
V. CONCLUSION
The volume and breadth of empirical studies on the patentability requirements are not
surprising, given the importance of such standards to the patent system. There is, however,
much ground left to be covered by future researchers. With regard to empirically quantify-
ing patent quality, sound and efficient methodologies to take the noted first approach—
examining patents themselves and formulating an educated opinion on their quality—have
yet to be articulated. In addition, use of the ever-increasingly available USPTO data can be
used to shed further light on patent quality. And as to studying validity decisions, the AIA’s
new post-grant challenge mechanisms provide ample ore to be mined in future projects.
Furthermore, the impact of the recent rash of Supreme Court decisions on the patentable
subject-matter requirement and the AIA’s new first-to-file rules has yet to be fully studied.
And all of this does not account for further examination and usage of the studies described
above. There is much to keep investigators in this area busy.
REFERENCES
Abrams, David S. and R. Polk Wagner. 2013. “Poisoning the Next Apple? The America Invents Act and
Individual Inventors”, 65 Stanford Law Review 516–564.
Allison, John R. and Lisa Larrimore Ouellette. 2016. “How Courts Adjudicate Patent Definiteness and
Disclosure”, 65 Duke Law Journal 609–695.
Allison, John R. and Ronald J. Mann. 2007. “The Disputed Quality of Software Patents”, 85 Washington
Allison, John R., Mark A. Lemley, and David Schwartz. 2014. “Understanding the Realties of Modern Patent
Litigation”, 92 Texas Law Review 1769–1801.
Allison, John R., Mark A. Lemley, and Joshua Walker. 2009. “Extreme Value or Trolls on Top? The
Characteristics of the Most-Litigated Patents”, 158 University of Pennsylvania Law Review 1, 3 & n.3.
Allison, John R., Mark A. Lemley, and Joshua Walker. 2011. “Patent Quality and Settlement Among Repeat
Patent Litigants”, 99 Georgetown Law Journal 677–712.
Allison, John R., Mark A. Lemley, Kimberly A. Moore, and R. Derek Trunkey. 2004. “Valuable Patents”, 92
Cotropia, Christopher A. 2006. “Patent Law Viewed Through an Evidentiary Lens: The ‘Suggestion Test’ as a
Rule of Evidence”, 2006 BYU Law Review 1517–1588.
Cotropia, Christopher A. 2007. “Nonobviousness and the Federal Circuit: An Empirical Analysis of Recent
Case Law”, 82 Notre Dame Law Review 911–953.
Cotropia, Christopher A. 2014. “Predictability and Nonobviousness in Patent Law After KSR”, 20 Mich.
Telecommunications & Technology Law Review 391–437.
Crouch, Dennis D. 2009. “Is Novelty Obsolete? Chronicling the Irrelevance of the Invention Date in U.S. Patent
Law”, 16 Michigan Telecommunications & Technology Law Review 53–107.
Lanjouw, Jean and Mark Schankerman. 2004. “Patent Quality and Research Productivity: Measuring
Innovation with Multiple Indicators”, 114:495 Economics Journal 441–465.
Love, Brian J. and Shawn Ambwani. 2014. “Inter Partes Review: An Early Look at the Number”, 81 University
of Chicago Law Review Dialogue 93–107.
Lunney, Jr., Glynn S. 2001. “E-Obviousness”, 7 Michigan Telecommunications & Technology Law Review 363–422.

Lunney, Jr., Glynn S. 2004. “Patent Law, the Federal Circuit and the Supreme Court: A Quiet Revolution”, 11
Lunney, Jr., Glynn S. and Christian Johnson. 2012. “Not So Obvious After All: Patent Law’s Nonobviousness
Requirement, KSR, and the Fear of Hindsight Bias”, 47 Georgia Law Review 41–111.
Mandel, Gregory N. 2006. “Patently Non-Obvious: Empirical Demonstration that the Hindsight Bias Renders
Patent Decisions Irrational”, 67 Ohio State Law Journal 1391–1463.
Mandel, Gregory N. 2007. “Patently Non-Obvious II: Experimental Study on the Hindsight Issue Before the
Supreme Court in KSR v. Teleflex”, 9 Yale Journal of Law & Technology 1–43.
Mann, Ronald J. and Marian Underweiser. 2012. “A New Look at Patent Quality: Relating Patent Prosecution
to Validity”, 9 Journal of Empirical Legal Studies 1–32.
McEldowney, Sean M. 2006. “New Insights on the ‘Death’ of Obviousness: An Empirical Study of District
Court Obviousness Opinions”, 2006 Stanford Technology Law Review 4–7.
Nock, Jennifer and Sreekar Gadde. 2010. “Raising the Bar for Nonobviousness: An Empirical Study of Federal
Circuit Case Law Following KSR”, 20 Federal Circuit Bar Journal 369–408.
Petherbridge, Lee and R. Polk Wagner. 2007. “The Federal Circuit and Patentability: An Empirical Assessment
of the Law of Obviousness”, 85 Texas Law Review 2051–2110.
Popp, David, Nidhi Santen, Karen Fisher-Vanden, and Mort Webster. 2013. “Technology Variation vs. R&D
Uncertainty: What Matters Most for Energy Patent Success?”, 35:4 Resource and Energy Economics 505–533.
Rantanen, Jason. 2013. “The Federal Circuit’s New Obviousness Jurisprudence: An Empirical Study”, 16
Risch, Michael. 2015. “A Generation of Patent Litigation”, 52 San Diego Law Review 67–132.
Risch, Michael. 2016. “The Layered Patent System”, 101 Iowa Law Review 1535–1579.

13. Patent duration
Brian J. Love*
Contents
I. Introduction
II. The Law of Patent Duration
III. Theoretical and Empirical Study of the Patent Term
IV. Theoretical and Empirical Study of Patent Renewal
V. Conclusion
References
I. INTRODUCTION
The duration of a given patent’s life is a factor of two variables: first, the maximum term
of protection afforded by law; and second, its owner’s willingness and ability to comply
with periodic renewal obligations. This chapter reviews the theoretical and empirical
literature related to both.
II. THE LAW OF PATENT DURATION
In most of the industrialized world, including the United States, newly issued patents
receive a nominal term of protection that extends 20 years from the filing date of the
earliest application to which the patent claims priority. In the U.S., the patent term is
calculated in this manner for all patents issued from applications filed on or after June
8, 1995 as set forth in the Uruguay Round Agreements Act (URAA) (1994), legislation
passed in 1994 following ratification of the Agreement on Trade-Related Aspects of
Intellectual Property Rights (“TRIPS Agreement”) (1994). Prior to passage of the
URAA, patents were awarded a term of 17 years from the date on which the U.S. Patent
and Trademark Office (USPTO) issued the patent, rather than from the date on which the
patent’s original application was filed. Patents that either were in force on June 8, 1995 or
were subsequently issued from applications pending on that day are protected for 17 years
from issue or 20 years from filing, whichever period is longer—a choice that extended
these patents’ terms by about eight months on average (Lemley, 1994).1
* Associate Professor and Co-Director of the High Tech Law Institute, Santa Clara University
School of Law.
1
In the U.S., the 17-years-from-issue calculation dates back to the Patent Act of 1861. Patent
Act of 1861, 36th Cong., 2d session, 12 Stat. 246, § 16 (1861). The original term of protection set
forth in the Patent Act of 1790 was 14 years. Patent Act of 1790, Ch. 7, 1 Stat. 109, § 1 (1790).
310

Patent duration 311
Overall, the pre- and post-URAA term calculations produce similar results. In the late
1990s, the average patent issued about 2.8 years after filing (Allison and Lemley, 2002).
Today, though the number of filings has grown considerably, average pendency is only
about six months longer (Crouch, 2013). As a result, typical patents receive roughly 17
years of post-issuance protection using either formulation of the patent term.
The shift from an issue date-based calculation to a filing date-based calculation
principally impacts outliers issuing from applications that pended at the USPTO for
abnormally long periods of time, often due to delays intentionally caused by their owners.
Under the pre-URAA regime, it was possible for patents to issue many years—in some
instances many decades—after their original filing date and thereafter still receive 17 years
of protection.2 Indeed, some applications filed prior to June 8, 1995 reportedly remain
in active prosecution at the USPTO today (Crouch, 2010; Crouch, 2014a). Using this
strategy, a patentee could keep a patent tied up in prosecution until a company worth
suing independently invented similar technology and incurred substantial fixed costs.
This practice is often referred to as “submarine” patenting. Numerous scholars, includ-
ing Blount and Zarfas (1999) and Lemley and Moore (2002), have criticized submarine
patenting as a form of bad-faith gamesmanship. Under current rules, a patentee has less
incentive to stall because its own prosecution delays will shorten, and eventually consume,
the life of its patent rights.
For the same reason, though, delays caused by the government also tend to limit
patents’ effective lives. To mitigate this, U.S. patent law provides two mechanisms for
extending patent rights to account for unusually long periods of agency action or inac-
tion. First, the USPTO can—and routinely does—extend the lives of post-URAA patents
to account for the agency’s own delays during prosecution.3 As shown in Figures 13.1 and
13.2, these so-called term “adjustments” have been awarded to about 70 percent of all
patents issued in the last few years and during this same period have averaged about one
year in duration (Crouch, 2014b).
Second, patents covering pharmaceutical products can receive a term extension—called
a term “restoration” in this context—to recoup a portion of the time required to obtain
regulatory approval by the U.S. Food and Drug Administration (FDA).4 Since this
provision was established by the “Hatch-Waxman” Drug Price Competition and Patent
During the period 1836–61, this 14-year term was renewable for an additional seven-year period,
bringing the total possible term of protection to 21. Patent Act of 1836, Ch. 357, 5 Stat. 117, § 18
(1836). Various unsuccessful legislative attempts to modify the patent term have been proposed as
well. A few examples are discussed in Office of Technology Assessment (1981).
2
For example, the patents at issue in Symbol Technologies, Inc. v. Lemelson Medical,
Education and Research Foundation, L.P., 277 F.3d 1361 (Fed. Cir. 2002), were in prosecution for
roughly 40 years before they issued in the 1990s.
3
These delays are measured relative to a series of prosecution deadlines set forth in the
American Inventors Protection Act, Pub. L. No. 106-113, 113 Stat. 1501, 1501A-557-560 (1999)
(codified 35 U.S.C. § 154(b)). Under the Act, extensions may also be given for delays caused by
interferences, secrecy orders, and successful appeals.
4
A Hatch-Waxman term extension equals the entire regulatory review period (unless the
patentee failed to act with due diligence), up to five years or a maximum post-approval effective
term of 14 years. 35 U.S.C. § 156. An additional six-month extension is also given to patentees that
conduct pediatric clinical trials. 21 U.S.C. § 355a.

Percentage of Patents Awarded Patent Term Adjustment

90%
80%
70%
Percentage of Patents Awarded PTA (%)
60%
50%
40%
30%
20%
10%
0%
Aug-04 Aug-05 Aug-06 Aug-07 Aug-08 Aug-09 Aug-10 Aug-11 Aug-12 Aug-13 Aug-14
Figure 13.1 Percentage of U.S. patents issued each week from Jan. 2005 to Aug. 2014
that were awarded a term adjustment (Crouch, 2014b)
Term Restoration Act (1984), the USPTO has extended the lives of over 600 drug patents
by an average of about 3.5 years (Clift, 2008).5 In rare instances—seven times since
1980—Congress has also simply bypassed both mechanisms entirely by passing private
laws that extend the life of particular patents (Schacht and Thomas, 2002).6
Independently of the patent system, pharmaceutical companies also benefit from laws
that delay competitors’ ability to seek “abbreviated” approval of generic versions of a new
drug for a period of time following that drug’s approval by the FDA. Under the Hatch-
5
For the full text of this section of the Drug Price Competition and Patent Term Restoration
Act, see Pub. L. No. 98-417, 98 Stat. 1585 (1984). A list of patents that have been granted an exten-
sion is available via the USPTO website at www.uspto.gov/patents/resources/terms/156.jsp.
6
U.S. patent law also allows a patent’s term to be shortened. Under 35 U.S.C. § 253, a patent
applicant may voluntarily waive a portion of the patent term to which it might otherwise be
entitled. These so-called “terminal disclaimers” are frequently used by patentees to overcome an
“obviousness-type double patenting” rejection, which arises when a patentee attempts to patent
obvious variations of its own prior patented inventions.

Patent duration 313
Average Patent Term Adjustment (PTA)

20
18
16
14
Average PTA (Months)
12
10
0
Aug-04 Aug-05 Aug-06 Aug-07 Aug-08 Aug-09 Aug-10 Aug-11 Aug-12 Aug-13 Aug-14
Figure 13.2 Average duration of term adjustments awarded to U.S. patents issued each
week from Jan. 2005 to Aug. 2014 (Crouch, 2014b)
Waxman Act, this period of “data exclusivity”—so called because abbreviated applica-
tions rely on clinical trial data submitted with the original new drug application—is set at
five years for new chemically manufactured “small molecule” drugs.7 The Biologics Price
Competition and Innovation Act, passed in 2010, extended the period of data exclusivity
to 12 years for new “large molecule” drugs, or “biologics,” derived from living cells.8 These
two statutes effectively create a sui generis form of IP protection that, like a patent, can
provide a period of market exclusivity for new pharmaceuticals. Because the duration of
7
This provision is codified at 21 U.S.C. § 355(c)(3)(E)(ii), (j)(5)(F)(ii). New small molecules
that additionally qualify as “orphan drugs” for the treatment of rare diseases receive a seven-year
period of data exclusivity. 21 U.S.C. §§ 360aa-360ee. An additional six months of data exclusivity is
also given to new small molecules that were tested in pediatric populations. 21 U.S.C. § 355a. New
variations of previously approved small molecules receive only three years of data exclusivity. 21
U.S.C. § 355(c)(3)(E)(iii)-(iv), (j)(5)(F)(iii)-(iv)).
8
This provision is codified at 42 U.S.C. § 262.

data exclusivity is short relative to the patent term, the two terms run concurrently for
most pharmaceuticals. However, for at least some new drugs, FDA approval occurs so
long after patent filing that data exclusivity extends beyond the expiration of patent rights
to the drug. Eisenberg and Thomas (2014) provide a useful summary of data exclusivity
rules and their interplay with pharmaceutical patents.
However a patent’s term is calculated, in order to enjoy the full term of protection
afforded by statute, a patentee must also periodically renew its patent rights. In the U.S., a
patent will expire prematurely unless its owner makes three “maintenance fee” payments,
which are due 3.5, 7.5, and 11.5 years after the date of issue.9 Current maintenance fee
rates are set forth in the USPTO’s Fee Schedule,10 but only “large entities” pay the full
fee. “Small entities,” or patentees that have fewer than 500 employees, receive a 50 percent
discount,11 and “micro entities,” even smaller patentees that are new to the patent system,
receive a 75 percent discount.12
Failure to make a maintenance fee payment results in the patent’s premature
expiration. Over the last 20 years, more than half of all issued U.S. patents expired
early for this reason. As shown in Figure 13.3, between 80 and 90 percent of patents
are renewed once and about two-thirds are renewed twice. Only 40 to 50 percent of
patentees elect to take advantage of the full patent term by making all three payments
(Crouch, 2015).13
Thanks to widespread adoption of the TRIPS Agreement, basic rules for calculating
the patent term are virtually uniform worldwide. However, many nations recognize a
distinct form of quasi-patent right, often referred to as a “utility model” or “petty” patent.
These “patents” may be obtained with less scrutiny and receive a shorter term of protec-
tion, typically between 6–15 years.14 Summaries of petty patent regimes are available via
9
Patent owners are given a six-month grace period during which they can pay overdue
maintenance fees without consequence. 35 U.S.C. § 41(b). In addition, a payment made within
24 months after the grace period’s expiration may be accepted if the delay was “unintentional” or
“unavoidable.” Id. § 41(c). Maintenance fees were first added to the U.S. patent system in 1980,
Pub. L. No. 96-517, sec. 2, 94 Stat. 3015, 3017-18 (1980) (establishing maintenance fees for patents
issued from applications filed on or after December 12, 1980). This is relatively recent by interna-
tional standards. Renewal fees have a much longer history in Europe and elsewhere in the world.
Federico (1954) provides a useful summary of international renewal fee regimes in place during the
mid-twentieth century. Chien (2012) discusses an unsuccessful push by the railroad industry to add
renewal fees to the U.S. patent system in the late nineteenth century.
10
The USPTO Fee Schedule is available online at www.uspto.gov/web/offices/ac/qs/ope/
fee010114.htm. De Rassenfosse and van Pottelsberghe de la Potterie (2013) discuss the historical
fluctuation of patent fees, including those charged in the U.S.
11
For more detail on “small entity” status, see 37 C.F.R. 1.27.
12
Universities also receive the “micro entity” fee discount. Changes to Implement Micro
Entity Status for Paying Patent Fees, 77 Fed. Reg. 75019 (2012).
13
This rate of payment in the U.S. is much higher than that observed in many other nations,
including Germany (Lanjouw, 1998a) and France (Schankerman, 1998).
14
Though the U.S. does not have a petty patent system, it does have separate “patent” regimes
for the protection of designs and asexually reproduced plants. U.S. design patents receive a term
of protection that extends 14 years from the date of issue. 35 U.S.C. § 173. U.S. plants patents are,
like ordinary “utility” patents, protected for a term of 20 years from filing. 35 U.S.C. § 161. The
U.S. also provides sui generis protection (albeit not labelled a form of “patent” protection) for
sexually reproduced plants, which receive a term of 20 or 25 years from issue, 7 U.S.C. § 2483(b);

Patent duration 315
1000%
90%
80%
Percentage of Patents with Fees Paid
70%
60%
50%
40%
30%
20%
Paid First Fee
10% Paid Second Fee
Paid Third Fee
0%
01/1995 01/1999 01/2003 01/2007 01/2011 01/2015
Maintenance Fee Payment Deadline
Figure 13.3 Percentage of first, second, and third maintenance fee payments made
1995–2014 (Crouch, 2015)
the World Intellectual Property Organization (2015) and Richards (2010). Janis (1999)
also provides a critique of these regimes.
Because patent renewal requirements are not the subject of a treaty, the structure and
amount of these fees are not nearly as uniform. In fact, the U.S. is largely an outlier among
nations that are popular with patent applicants. For one, most other nations require
patent owners to renew their patents annually.15 In addition, U.S. maintenance fees are
relatively low by international standards, particularly viewed relative to population and
GDP. De Rassenfosse and van Pottelsberghe de la Potterie (2013) and Park (2010) both
compare patent fees across nations in absolute and relative terms.
semiconductor mask works, which are protected for ten years from registration, 17 U.S.C. § 904;
and boat hull designs, which also receive a term of ten years, 17 U.S.C. § 1305.
15
In Germany, for example, patent renewal fees are due “the third and each subsequent year
following the international filing date” (German Patent and Trademark Office, 2012). Similarly,
in the U.K. patents must be renewed “on the 4th anniversary of the filing date and every year
after that . . . up to 20 years” (United Kingdom Intellectual Property Office, 2015). Japan requires
renewal fees every year after filing (Japan Patent Office, 2012).

III.
THEORETICAL AND EMPIRICAL STUDY OF THE PATENT
TERM
The modern choice of a roughly 17-year patent term is more an artifact of history than a
welfare-maximizing calculation by policy-makers (White, 1956).16 However, scholars have
long recognized the patent term’s potential as a tool for calibrating the costs and benefits
of patent protection. Patent protection works to incentivize invention by excluding imita-
tors, thereby permitting the inventor to charge supracompetitive prices. In doing so, the
patent system makes society better off by allowing inventors to recoup the sunk cost of
R&D, which has the effect of boosting incentives to innovate to more socially optimal
levels. At the same time, however, the patent system also acts in various ways to reduce
social welfare and, thus, counteract its positive contribution to society.
One way to limit these welfare-reducing effects is to limit the duration of the patent
right itself. Perhaps the most obvious source of lost welfare is the deadweight loss that
results from supracompetitive prices. Nordhaus (1969, 1972) and Scherer (1972) show
that in a patent system where invention is not cumulative, the optimal patent term is one
that weighs incentives to innovate against the deadweight loss created by reduced competi-
tion. As Nordhaus (1969) explains:
As [the patent term] is increased, two opposite forces affect the level of economic welfare.
First, a longer life increases invention and thus gives on balance a larger amount of output for
a given level of inputs. This is a positive effect. Second, a longer life means that the monopoly
on information lasts longer and thus there are more losses from inefficiencies associated with
monopoly. The optimal life of a patent is that point at which the two forces balance at the margin.
Nordhaus’ model and ideas were subsequently extended by, among others, Kamien and
Schwartz (1974), who additionally consider the effects of rivalry among firms to invent
first and secure patent rights; and Klemperer (1990) and Gilbert and Shapiro (1990),
both of whom consider the optimal balance of patent term and claim scope.17 Green and
Scotchmer (1995) and O’Donoghue et al. (1998), among others, subsequently tackled
patent term’s intersection with another cost of the patent system—its potential to inhibit
future innovation by locking current technology away from those who might improve it
or otherwise utilize it in new research—by updating models of optimal patent protection
to account for the cumulative nature of innovation.
Building on this foundation, contemporary study of the patent term largely focuses on
the term’s potential to mitigate inefficiencies created by the patent system’s near-uniform
application across what, in reality, is a broad spectrum of industries and technologies.
Firms in different industries tend to rely on patent protection to different extents. As a
16
As White (1956) explains, the U.S. patent term is derived from the English Statute of
Monopolies of 1623, 21 Jac. 1, c. 3, which set the duration of protection at 14 years to match the
time required under English law to train two apprentice craftsmen.
17
More recent additions to the Nordhaus model include Partnoy (2001), who considers the
impact of financial variables, especially interest rates; and Duffy (2003), who considers the possibil-
ity that in some circumstances a longer term of protection can speed the pace of innovation and
thereby draw inventions into the public domain more quickly than would happen in a regime with
a shorter term.

Patent duration 317
result, many scholars have called for the patent system to operate on an industry-specific
basis more often than it currently does (Levin, et al., 1987; Carroll, 2006; Bessen and
Meurer, 2008; Burk and Lemley, 2009).18 Among other policy levers that might be used
to achieve this goal, a number of scholars have called for the adoption of industry- or
technology-specific terms of patent protection (Tabarrock, 2002; Johnson, 2006; Khoury,
2010; Roin, 2014)—a recommendation supported by empirical evidence such as that
gathered by Bilir (2014), who, in a study of multinational corporations’ manufacturing
locations, found a positive correlation between an industry’s product lifecycles and its
sensitivity to patent duration; Sukhatme and Cramer (2014), who found variation among
industries in a study of patent prosecution speed following passage of the URAA; and
Giummo (2014), who also found inter-industry differences in the temporal distribution of
returns on patented inventions as reflected in German employee compensation records.
As with the broader scholarly debate on industry exceptionalism, scholars studying
industry-specific patent terms have largely focused on what could be called the “end
points” of the industrial spectrum: software and computing on one side and pharma-
ceuticals and biotechnology on the other. Scholars who identify software and computing
as candidates for special treatment generally point to the disproportionate speed with
which software and other computer-related inventions and products iterate. Moore’s 1975
prediction that computing power would double every two years (Moore, 1975)—known
today as “Moore’s Law”—has so far proven to be accurate. As a result of this exponential
increase, products in the industry become obsolete quickly. According to one recent
report, half of all revenue in the semiconductor industry is derived from products that
have been on the market less than six months (Aboagye, et al., 2012).
Given the gap between product lifecycles in this industry and patent duration, several
scholars have argued that the term of protection should be reduced for software patents,
generally to somewhere between three and ten years (Menell, 1987; Phillips, 1992; Stern,
1993; Samuelson et al., 1994; Paley, 1996). Outside the academy, notable figures such as
Federal Circuit Judge Pauline Newman (Valek, 2008), entrepreneur and investor Mark
Cuban (Quinn, 2015), and Amazon.com founder Jeff Bezos (Richtel, 2000) have made
similar suggestions.
At the other end of the spectrum is the pharmaceutical industry.19 In contrast to the
rapid pace of product development in the computer industry, the development of new
drugs and medical treatments often requires many years of R&D plus several additional
years of testing to achieve regulatory approval to bring a product to market. The length
and complexity of this process have two effects. First, the unpredictable nature of R&D
and regulatory approval drive up the cost of drug development. Estimates of the average
cost of developing a new drug vary greatly, but generally exceed $100 million (Morgan
et al., 2011), with recent industry-sponsored estimates of the total cost of developing a
18
For a counterpoint to these arguments, see Wagner (2003) and National Research Council
(2004), both arguing in favor of patent uniformity.
19
Other industries that arguably fall nearby on the continuum include oil and gas exploration
and clean-energy technology (Roin, 2014). Abramowicz (2011) also describes the characteristics of
inventions that may be inadequately incentivized by the current patent system, including examples
such as electric car battery charging stations, carbon dioxide removal technologies, and supersonic
jet travel.

wholly new drug, or “new molecular entity,” exceeding $1 billion (DiMasi and Grabowski,
2007; DiMasi et al., 2014). Second, due to the duration of the regulatory approval process,
drug patents are often many years old before the drug they protect is actually approved for
sale. Even with the benefit of term extensions, pharmaceuticals receive only about 12 years
of market exclusivity on average once they are approved for sale (Grabowski and Vernon,
2000a; Clift, 2008; Hemphill and Sampat, 2012) and thereafter quickly lose market share
to competition from generics (Grabowski and Kyle, 1997; Aitken, et al., 2008).
Firms are able to mitigate both effects to some extent by serially patenting and mar-
keting variants of popular drugs and treatments (Hutchins, 2003; Eisenberg, 2007)—a
strategy commonly referred to as “evergreening.” But studies suggest that evergreening is
only moderately successful. While it provides an additional 7.5 years of nominal patent
protection on average for new drugs (Kapczynski et al., 2012), in practice follow-on, or
“secondary,” pharmaceutical patents are often successfully challenged by generic drug
companies and thus rarely lead to a significant extension of market exclusivity (Hemphill
and Sampat, 2012).
Moreover, while the lives of most computer-related products last but a few months, new
drugs often remain popular for decades after their introduction (Grabowski and Vernon,
1990, 2000b; DiMasi et al., 2004). As a result, compared to that of other industries, patent
protection for pharmaceuticals is relatively short both with respect to the absolute number
of years of market exclusivity that is possible and compared to the duration of product
lifecycles.
Pointing to the disproportionately large investment required to bring a product to
market in the pharmaceutical industry, combined with the relatively short period of
market exclusivity afforded by current patent law, scholars have suggested that the term
of protection ought to be increased for patents covering at least some new drugs. Roin
(2010), for example, points to treatments for early-stage cancer as particularly deserving
of additional protection because their approval requires especially long-term clinical
trials.20 A good deal of empirical evidence also supports the notion that the pharmaceuti-
cal industry is especially sensitive to the term of patent protection. Abrams (2009) found
a significant positive relationship between the URAA’s impact on the effective patent
term and the number of patents sought in various technology classes, especially the
biological sciences. Likewise, Budish et al. (2015) present evidence that pharmaceutical
R&D favors drugs and treatments that are likely to be approved relatively quickly over
those that generally require longer-term clinical trials. Relying on prior studies of the
drug industry’s exclusivity-focused funding priorities, Goldman et al. (2011) estimate that
giving pharmaceutical companies an additional four to seven years of market exclusivity
(via data exclusivity, rather than additional patent protection) would induce an additional
228 new drug approvals during the period 2020–60.
Despite the theoretical appeal of industry-specific patent terms, scholars have also
raised two important practical challenges to implementing such a system. First, the
TRIPS Agreement (1994) mandates that patent rights must “be available and . . .
20
Oddi (1989) has more broadly proposed the creation of a sui generis “Revolutionary Patent”
that would provide incontestable protection to “extraordinary” inventions for 34 years from the
patent grant date.

Patent duration 319
e njoyable without discrimination as to the field of technology”; and, moreover, requires

that signatories provide at least 20 years of protection. A member nation with a term of
protection that varies significantly across industries would seemingly violate at least one
of these requirements and thus could face trade sanctions (Love, 2013; Roin, 2014). That
said, as Chien (2013) explains, various aspects of existing U.S. and E.U. patent law already
violate TRIPS’s technology-neutrality requirement with no real consequence.
Second, any regime put in place would be vulnerable to manipulation by rent-seeking
patentees hoping to extend the duration of protection applicable to their respective
inventions, either by lobbying Congress or by convincing the USPTO that their inventions
belong in other categories with lengthier terms. As Roin (2014) explains, the boundaries
between industries and technologies are blurry:
Consider, for example, the classification of a brain-computer interface technology, which may be
akin to software, computer hardware, other electronics, medical devices, diagnostics, video game
technology, or any other field that is likely to use these inventions. Any system of technology-
specific patent laws will inevitably encounter difficult line-drawing questions of this nature.
As a result, some scholars have questioned whether a manageable system of variable

terms could actually be put into operation (Jaffe and Lerner, 2004; Long, 2008; Burk and
Lemley, 2009; Roin, 2014).
IV. THEORETICAL AND EMPIRICAL STUDY OF PATENT

RENEWAL
Compared to patent term, which has been the subject of scholarly commentary for
close to 50 years, patent renewal fees are new to the academic spotlight. Historically,
patent fees have generally been set at low, patentee-friendly levels that approximate the
cost of performing the ministerial acts to which they are tied, with little consideration
of policy implications. As a result, scholars and policy-makers long ignored them as all
but irrelevant to the proper functioning of the patent system (De Rassenfosse and van
Pottelsberghe de la Potterie, 2013).
Fees for renewing patent rights—something that itself requires no action on the part of
the USPTO—have by extension historically been explained (to the extent they have been
at all) as a mechanism for delaying a portion of the costs incurred during prosecution so as
to lower upfront application fees and thereby encourage indigent or risk-averse inventors
to seek protection (Federico, 1954; USPTO, 2000). As the USPTO (2000) explains:
the maintenance fee concept was originally adopted to provide patent holders flexibility in the
face of uncertainty before the fact as to whether or not the patent would be commercially viable.
Instead of requiring the entire investment up front, owners were given the option to pay out
gradually. . . .
To the extent that the USPTO views the patent renewal process in this way today, the
agency uses renewal fees to defer collection of about one-third of the upfront cost of
prosecution. According to the USPTO’s 2013 budget report, maintenance fee payments
made up 34 percent of the agency’s total revenue (USPTO, 2013). While a structure that

defers fees until after issuance may well induce marginal filings,21 it may also skew the
agency’s incentives to thoroughly examine applications. Frakes and Wasserman (2013)
find evidence that the USPTO’s heavy budgetary reliance on fees collected post-issuance
biases the agency toward granting patents.
A more modern take on renewal fees is to view them as a mechanism for optimizing
the patent system by fine tuning the duration of patent rights. As Scotchmer (1999) and
Cornelli and Schankerman (1999) first observed, a patent system with renewal require-
ments is, in essence, a patent system with a variable patent term selected for each patent by
its owner. Viewed in this way, renewal fees help the patent system more accurately match
the patent incentive with the value of the patented invention and, moreover, do so using
the patentee’s own private information about the invention’s value. Models developed by
these scholars have been advanced by, among others, Baudry and Dumont (2009) and
Danguy and van Pottelsberghe de la Potterie (2011), both of whom simulated optimal
fee amounts in Europe; Gans et al. (2004), who studied optimal fee structure for a “self-
funded” patent office that must cover its own costs; and Sukhatme (2014), who estimates
optimal U.S. renewal fees for various categories of technology.
In a recent extension of these ideas, a number of scholars have studied renewal fees’
ability to induce the expiration of especially weak or unimportant patents. Moore
(2005), Baudry and Dumont (2006), and Giummo (2014) each suggest that additional
maintenance fees may spur innovation by reducing both the “thicket” of issued patents
that entrepreneurs must (at least in theory) parse before introducing new products, and
the large backlog of pending applications that currently sits unexamined at the USPTO.
In a similar vein, several scholars have recently explored renewal fees’ potential to reduce
so-called “patent trolling” by entities that specialize in filing patent suits. Magliocca
(2007), for example, suggests that an increase in renewal fees could serve as a “dormancy
tax” that deters patentees from opportunistically sitting on their rights. Similarly, Love
(2013) presents empirical evidence that non-practicing patentees assert patents that, on
average, are significantly older than patents enforced by product-producing technology
companies, and thus recommends an increase in late-term patent renewal fees to induce
the expiration of patents that are most likely to be used in abusive lawsuits. Bessen and
Love (2013) argue that renewal fees could also be set at “Pigovian” levels that force paten-
tees to internalize the cost of suits filed by non-practicing patentees and, using litigation
cost data, estimate renewal fees that would accomplish this goal. Most recently, Olson
(2017) suggests that the USPTO charge patentee-specific maintenance fees that increase
with the number of non-practiced patents held in each patentee’s portfolio.22
21
De Rassenfosse and van Pottelsberghe de la Potterie (2013) provide a summary of the
literature related to the relationship between fee levels and patenting activity.
22
Another distinct line of scholarship studies patent renewal rates and fees, not for purposes
of formulating an optimal fee structure, but rather as a means to estimate and model the value of
patents and R&D across industries and time. This type of analysis dates back to early studies by
Pakes and Schankerman (1984) and Schankerman and Pakes (1986), both of which study patent
renewal rates primarily in France, Germany, and the UK. These early studies have been periodically
expanded and updated by, among others, Lanjouw et al. (1998b), Schankerman (1998), and Deng
(2011).

Patent duration 321
V. CONCLUSION
To date, scholars have largely studied renewal fees and the patent term in separate litera-
ture. However, this practice seems destined to fade as scholars increasingly view term-
focused and fee-focused scholarship as two sides of the same coin. Many recent empirical
studies have drawn from multiple lines of duration-related scholarship. Sukhatme (2014),
for example, uses data on both term sensitivity and renewal rates.
Among other trends in empirical research, the ever-increasing availability of detailed
data on patent litigation and licensing may be the most promising for future study of
patent duration. This data opens yet another avenue for complementary research into
firms’ use of, and returns from, patents over time. Love (2013) and Bessen and Love
(2013), for example, use litigation data to link patent enforcement and patent duration,
and Giummo (2014) uses royalty data reflected in employee compensation to examine
what licensing can tell us about optimizing the rules for patent duration.
On the theoretical front, the cutting edge of modeling optimal patent duration appears
to be a further refinement of the dynamic effects of altering the length of patent rights,
as reflected in recent scholarship by Duffy (2003) and Deng (2011), for example. Also,
scholars such as Partnoy (2001) and Baudry and Dumont (2009) continue to introduce
principles from financial economics into the literature.
Trends such as these aside, a good deal of core research in this area remains ripe for the
picking. Because the patent term is, and has been, controlled by an evolving statutory regime,
opportunities for natural experiments and event studies abound, and, to date, remain largely
untapped. Similarly, the patchwork of renewal fee regimes in place around the globe provides
ample opportunity for the comparison of firm behavior across borders, only some of which
has been examined thus far. Among existing studies as well, many of the most significant
were carried out in the 1990s, and thus seem overdue for replication with updated data.
REFERENCES
Aboagye, Aaron, et al. 2012. “Finding the Next $100 Billion in Semiconductor Revenues”, McKinsey on
Semiconductors Autumn.
Abramowicz, Michael. 2011. “Orphan Business Models: Toward a New Form of Intellectual Property”, 124
Abrams, David S. 2009. “Did TRIPS Spur Innovation? An Analysis of Patent Duration and Incentives to
Innovate”, 157 University of Pennsylvania Law Review 1613–47.
Aitken, Murry, et al. 2008. “Prescription Drug Spending Trends in the United States: Looking Beyond the
Turning Point”, 28 Health Affairs w151–w160.
Allison, John R. and Mark A. Lemley. 2002. “The Growing Complexity of the United States Patent System”,
Baudry, Marc and Beatrice Dumont. 2006. “Patent Renewals as Options: Improving the Mechanism for
Weeding Out Lousy Patents”, 28 Review of Industrial Organization 41–62.
Baudry, Marc and Beatrice Dumont. 2009. “A Bayesian Real Option Approach to Patents and Optimal Renewal
Fees”, Working Paper, https://halshs.archives-ouvertes.fr/hal-00419330/document.
Bessen, James and Brian J. Love. 2013. “Make the Patent ‘Polluters’ Pay: Using Pigovian Fees to Curb Patent
Abuse”, 4 California Law Review Circuit 84–91.
Bessen, James and Michael J. Meurer. 2008. Patent Failure: How Judges Bureaucrats, and Lawyers Put Innovation
at Risk. Princeton, NJ: Princeton University Press.
Bilir, L. Kamran. 2014. “Patent Laws, Product Lifecycle Lengths, and Multinational Activity”, 104 The
Blount, Steve and Louis S. Zarfas. 1999. “The Use of Delaying Tactics to Obtain Submarine Patents and Amend

Around a Patent That a Competitor Has Designed Around”, 81 Journal of the Patent and Trademark Office
Society 11–32.
Budish, Eric, et al. 2015. “Do Firms Underinvest in Long-Term Research? Evidence from Cancer Clinical
Trials”, 105 The American Economic Review 2044–85.
Burk, Dan and Mark A. Lemley. 2009. The Patent Crisis and How the Courts Can Solve It. Chicago, IL: The
Carroll, Michael W. 2006. “One for All: The Problem of Uniformity Cost in Intellectual Property Law”, 55
American University Law Review 845–900.
Chien, Colleen V. 2012. “Reforming Software Patents”, 50 Houston Law Review 325–88.
Chien, Colleen V. 2013. “Tailoring the Patent System to Work for Software and Technology Patents”, Working
Paper, http://papers.ssrn.com/sol3/papers.cfm?abstract_id=2176520.
Clift, Charles. 2008. “The Value of Patent Term Extensions to the Pharmaceutical Industry in the USA”, 5
Journal of Generic Medicines 201–08.
Cornelli, Francesca and Mark Schankerman. 1999. “Patent Renewals and R&D Incentives”, 30 RAND Journal
Crouch, Dennis. 2010. “Old School Submarine Patents”, Patently-O, http://patentlyo.com/patent/2010/12/old-
school-submarine-patents.html.
Crouch, Dennis. 2013. “Average Pendency of U.S. Patent Applications”, Patently-O, http://patentlyo.com/pat
ent/2013/03/average-pendency-of-us-patent-applications.html.
Crouch, Dennis. 2014a. “Old Applications; New Patents”, Patently-O, http://patentlyo.com/patent/2014/01/
old-patents.html.
Crouch, Dennis. 2014b. “USPTO Continues to Reduce Patent Term Adjustments”, Patently-O, http://patentlyo.
com/patent/2014/08/continues-patent-adjustments.html.
Crouch, Dennis. 2015. “Maintenance Fees 2015”, Patently-O, http://patentlyo.com/patent/2015/07/maintenance-
fees-2015.html.
Danguy, Jerome and Bruno van Pottelsberghe de la Potterie. 2011. “Cost-Benefit Analysis of the Community
Patent”, 2 Journal of Benefit-Cost Analysis art. 3.
de Rassenfosse, Gaétan and Bruno van Pottelsberghe de la Potterie. 2013. “The Role of Fees in Patent Systems:
Theory and Evidence”, 27 Journal of Economic Surveys 696–716.
Deng, Yi. 2011. “A Dynamic Stochastic Analysis of International Patent Application and Renewal Processes”,
29 International Journal of Industrial Organization 766–77.
DiMasi, Joseph A. and Henry G. Grabowski. 2007. “The Cost of Biopharmaceutical R&D: Is Biotech
Different?”, 28 Managerial and Decision Economics 469–79.
DiMasi, Joseph A., et al. 2004. “R&D Costs and Returns by Therapeutic Category”, 38 Drug Information
Journal 211–23.
DiMasi, Joseph A., et al. 2014. “Innovation in the Pharmaceutical Industry: New Estimates of R&D Costs”, Tufts
Center for the Study of Drug Development, http://csdd.tufts.edu/files/uploads/Tufts_CSDD_briefing_on_RD_
cost_study_-_Nov_18,_2014..pdf.
Duffy, John F. 2003. “A Minimum Optimal Patent Term”, Working Paper, http://ssrn.com/abstract=354282.
Eisenberg, Rebecca S. 2007. “The Role of the FDA in Innovation Policy”, 13 Michigan Telecommunications and
Eisenberg, Rebecca S. and Jay Thomas. 2014. “Patents and Regulatory Exclusivity in the USA”, in Anthony J.
Culyer, ed., Encyclopedia of Health Economics. Philadelphia, PA: Elsevier 443–52.
Federico, P. J. 1954. “Renewal Fees and Other Patent Fees in Foreign Countries”, 36 Journal of the Patent Office
Society 827–61.
Frakes, Michael D. and Melissa F. Wasserman. 2013. “Does Agency Funding Affect Decisionmaking?: An
Empirical Assessment of the PTO’s Granting Patterns”, 66 Vanderbilt Law Review 67–125.
Gans, Joshua S., et al. 2004. “Patent Renewal Fees and Self-Funding Patent Offices”, 4 Topics in Theoretical
Economics art. 6.
German Patent and Trademark Office. 2012. “PCT Applicant’s Guide”, www.wipo.int/pct/guide/en/gdvol2/
annexes/de.pdf.
Gilbert, Richard and Carl Shapiro. 1990. “Optimal Patent Length and Breadth”, 21 RAND Journal of
Economics 106–12.
Giummo, Jesse. 2014. “An Examination of the Intertemporal Returns of Patented Inventions”, 43 Research
Policy 1312–19.
Goldman, Dana P., et al., 2011. “The Benefits from Giving Makers of Conventional ‘Small Molecule’ Drugs
Longer Exclusivity Over Clinical Trial Data”, 30 Health Affairs 84–90.
Grabowski, Henry and Margaret Kyle. 2007. “Generic Competition and Market Exclusivity Periods in
Pharmaceuticals”, 27 Managerial and Decision Economics 1–12.
Grabowski, Henry G. and John M. Vernon. 1990. “A New Look at the Returns and Risks to Pharmaceutical
R&D”, 36 Management Science 804–21.

Patent duration 323
Grabowski, Henry G. and John M. Vernon. 2000a. “Effective Patent Life in Pharmaceuticals”, 19 International
Journal of Technology Management 98–120.
Grabowski, Henry G. and John M. Vernon. 2000b. “The Distribution of Sales Revenues from Pharmaceutical
Innovation”, 18 Pharmacoeconomics Supplement 21–32.
Green, Jerry R. and Suzanne Scotchmer. 1995. “On the Division of Profit in Sequential Innovation”, 26 RAND
Hemphill, C. Scott and Bhaven N. Sampat. 2012. “Evergreening, Patent Challenges, and Effective Market Life
in Pharmaceuticals”, 31 Journal of Health Economics 327–39.
Hutchins, Mike. 2003. “Extending the Monopoly: How ‘Secondary Patents’ Can Be Used to Delay or Prevent
Generic Competition Upon Expiry of the Basic Product Patent”, 1 Journal of Generic Medicine 57–71.
Jaffe, Adam B. and Josh Lerner. 2004. Innovation and Its Discontents: How Our Broken Patent System is
Endangering Innovation and Progress, and What to Do About It. Princeton, NJ: Princeton University Press.
Janis, Mark D. 1999. “Second Tier Patent Protection”, 40 Harvard International Law Journal 151–219.
Japan Patent Office. 2012. “Schedule of Fees”, www.jpo.go.jp/tetuzuki_e/ryoukin_e/ryokine.htm.
Johnson, Eric E. 2006. “Calibrating Patent Lifetimes”, 22 Santa Clara Computer and High Technology Law
Journal 269–314.
Kamien, Morton I. and Nancy L. Schwartz. 1974. “Patent Life and R & D Rivalry”, 64 The American Economic
Review 183–7.
Kapczynski, Amy, et al. 2012. “Polymorphs and Prodrugs and Salts (Oh My!): An Empirical Analysis of
‘Secondary’ Pharmaceutical Patents”, 7(12) PLoS ONE e49470.
Khoury, Amir H. 2010. “Differential Patent Terms and the Commercial Capacity of Innovation”, 18 Texas
Klemperer, Paul. 1990. “How Broad Should the Scope of Patent Protection Be?”, 21 RAND Journal of
Economics 113–30.
Lanjouw, Jean O. 1998a. “Patent Protection in the Shadow of Infringement: Simulation Estimations of Patent
Value”, 65 Review of Economic Studies 671–710.
Lanjouw, Jean O., et al. 1998b. “How to Count Patents and Value Intellectual Property: The Uses of Patent
Renewal and Application Data”, 46 Journal of Industrial Economics 405–32.
Lemley, Mark A. 1994. “An Empirical Study of the Twenty-Year Patent Term”, 22 AIPLA Quarterly Journal
369–424.
Lemley, Mark A. and Kimberly A. Moore. 2002. “Ending Abuse of Patent Continuations”, 84 Boston University
Levin, Richard C., et al. 1987. “Appropriating the Returns from Industrial Research and Development”,
Brookings Papers on Economic Activity 783–831.
Long, Clarisa. 2008. “Our Uniform Patent System”, 55 Federal Lawyer 44–50.
Love, Brian J. 2013. “An Empirical Study of Patent Litigation Timing: Could a Patent Term Reduction
Decimate Trolls Without Harming Innovators?”, 161 University of Pennsylvania Law Review 1309–59.
Magliocca, Gerard N. 2007. “Blackberries and Barnyards: Patent Trolls and the Perils of Innovation”, 82 Notre
Dame Law Review 1809–38.
Menell, Peter S. 1987. “Tailoring Legal Protection for Computer Software”, 39 Stanford Law Review 1329–72.
Moore, Gordon E. 1975. “Progress in Digital Integrated Electronics”, IEDM Technology Digest 11–13.
Moore, Kimberly A. 2005. “Worthless Patents”, 20 Berkeley Technology Law Journal 1521– 52.
Morgan, Steve, et al. 2011. “The Cost of Drug Development: A Systematic Review”, 100 Health Policy 4–17.
National Research Council. 2004. A Patent System for the 21st Century, www.nap.edu/catalog/10976/a-patent-sys
tem-for-the-21st-century.
Nordhaus, William D. 1972. “The Optimal Life of a Patent: Reply”, 62 The American Economic Review 428–31.
Oddi, A. Samuel. 1989. “Beyond Obviousness: Invention in the Twenty-First Century”, 38 American University
O’Donoghue, Ted, et al. 1998. “Patent Breadth, Patent Life, and the Pace of Technological Progress”, 7 Journal
of Economics and Management Strategy 1–32.
Olson, David S. 2017. “Removing the Troll from the Thicket: The Case for Enhancing Patent Maintenance
Fees in Relation to the Size of a Patent Owner’s Non-Practiced Patent Portfolio”, 68 Florida Law Review
519–69.
Pakes, Ariel and Mark Schankerman. 1984. “The Rate of Obsolescence of Patents, Research Gestation Lags,
and the Private Return to Research Resources”, in Zvi Griliches, ed., R & D, Patents, and Productivity.
Chicago, IL: University of Chicago Press 73–88.
Paley, Mark Aaron. 1996. “A Model Software Petite Patent Act”, 12 Santa Clara Computer & High Technology
Park, Walter G. 2010. “On Patenting Costs”, 2 World Intellectual Property Organization Journal 38–48.

Partnoy, Frank. 2001. “Finance and Patent Length”, University of San Diego School of Law, Law and Economics
Research Paper No. 19, http//ssrn.com/abstract=285144.
Phillips, John C. 1992. “Sui Generis Intellectual Property Protection for Computer Software”, 60 George
Washington Law Review 997–1041.
Quinn, Gene. 2015. “A Patent Conversation with Mark Cuban”, IPWatchdog, www.ipwatchdog.com/2015/11/10/a-
patent-conversation-with-mark-cuban/id=63090/.
Richards, John. 2010. “Utility Model Protection Throughout the World”, IPO Articles & Reports, Patents
Section, No. 4.
Richtel, Matt. March 11, 2000. “Chairman of Amazon Urges Reduction of Patent Terms”, New York Times C4.
Roin, Benjamin N. 2010. “Drug Patent Length”, Working Paper, www.law.berkeley.edu/files/bclt_IPSC2010_
Roin(1).pdf.
Roin, Benjamin N. 2014. “The Case for Tailoring Patent Awards Based on Time-to-Market”, 61 UCLA Law
Review 672–759.
Samuelson, Pamela, et al. 1994. “A Manifesto Concerning the Legal Protection of Computer Programs”, 94
Schacht, Wendy H. and John R. Thomas. 2002. “Pharmaceutical Patent Term Extensions: A Brief Explanation”,
Congressional Research Service Report for Congress.
Schankerman, Mark. 1998. “How Valuable is Patent Protection? Estimates by Technology Field”, 29 RAND
Schankerman, Mark and Ariel Pakes. 1986. “Estimates of the Value of Patent Rights in European Countries
During the Post-1950 Period”, 96 The Economic Journal 1052–76.
Scherer, F.M. 1972. “Nordhaus’ Theory of Optimal Patent Life: A Geometric Reinterpretation”, 62 The
Scotchmer, Suzanne. 1999. “On the Optimality of the Patent Renewal System”, 30 RAND Journal of Economics
181–96.
Stern, Richard H. 1993. “A Sui Generis Utility Model Law as an Alternative Legal Model for Protecting
Software”, 1 University of Baltimore Intellectual Property Law Journal 108–15.
Sukhatme, Neel U. 2014. “Regulatory Monopoly and Differential Pricing in the Market for Patents”, 71
Washington and Lee Law Review 1855–922.
Sukhatme, Neel U. and Judd N. L. Cramer. 2014. “How Much Do Patent Applicants Care About Patent Term?
Cross-Industry Differences in Term Sensitivity”, Working Paper, http://ssrn.com/abstract=2293245.
Tabarrock, Alexander. 2002. “Patent Theory and Patent Law”, 1 Contributions to Economic Analysis and Policy
art 9.
United Kingdom Intellectual Property Office. 2015. “Renewing Your Patent”, www.ipo.gov.uk/types/patent/p-
manage/p-renew.htm.
United States Office of Technology Assessment. 1981. “Patent Term Extension and the Pharmaceutical
Industry”, http://ota.fas.org/reports/8119.pdf.
United States Patent and Trademark Office. 2000. “Study of Alternative Fee Structures”, 65 Federal Register
58746-02-58748-02.
United States Patent and Trademark Office. 2013. “Performance and Accountability Report: Fiscal Year 2013”,
www.uspto.gov/about/stratplan/ar/USPTOFY2013PAR.pdf.
United States Patent and Trademark Office. 2015. “Fee Schedule”, www.uspto.gov/web/offices/ac/qs/ope/
fee010114.htm.
Valek, Michael. 2008. “Should Software Patents have Shorter Life Spans than Other Patents?”, CNET, http://
news.cnet.com/8301-13796_3-9917345-79.html.
Wagner, R. Polk. 2003. “(Mostly) Against Exceptionalism”, in F. Scott Kieff ed., Perspectives on Properties of
the Human Genome Project. Waltham, MA: Academic Press 367–82.
White, C. Michael. 1956. “Why a Seventeen Year Patent?”, 38 Journal of the Patent Office Society 839–59.
World Intellectual Property Organization, “Where Can Utility Models Be Acquired?”, www.wipo.int/sme/en/
ip_business/utility_models/where.htm.
Cases
Symbol Technologies, Inc. v. Lemelson Medical, Education and Research Foundation, L.P. (2002), 277 F.3d 1361.
37 C.F.R. 1.27.
7 U.S.C. § 2483

Patent duration 325
17 U.S.C. § 904
17 U.S.C. § 1305
21 U.S.C. § 355.
21 U.S.C. § 355a.
21 U.S.C. §§ 360aa-360ee.
35 U.S.C. § 41.
35 U.S.C. § 154.
35 U.S.C. § 156.
35 U.S.C. § 161.
35 U.S.C. § 173.
35 U.S.C. § 253.
42 U.S.C. § 262.
Agreement on Trade-Related Aspects of Intellectual Property Rights, 1869 U.N.T.S. 299 (1994).
American Inventors Protection Act, Pub. L. No. 106-113, 113 Stat. 1501 (1999).
Bayh–Dole Patent and Trademark Law Amendments Act, Pub. L. No. 96-517, 94 Stat. 3015 (1980).
Changes to Implement Micro Entity Status for Paying Patent Fees, 77 Fed. Reg. 75019 (2012).
English Statute of Monopolies of 1623, 21 Jac. 1, c. 3.
Hatch-Waxman Drug Price Competition and Patent Term Restoration Act, Pub. L. No. 98-417, 98 Stat. 1585
(1984).
Patent Act of 1790, 1 Stat. 109 (1790).
Uruguay Round Agreements Act, Pub. L. No. 103-465, 108 Stat. 4809 (1994).

14. Infringement
Lee Petherbridge* and Jason Rantanen**
Contents
I. Introduction
II. General Litigation Outcomes
A. Infringement Outcomes in the Pre-Federal Circuit Era
B. Infringement Outcomes in the Post-Federal Circuit Era
III. Outcomes Related to Case, and Other Variables
A. Infringement and Procedural Posture
B. Infringement and Subject Matter
C. Special Plaintiffs
IV. Appellate Treatment of Infringement Judgments
V. The Doctrine of Equivalents
A. Outcomes and the Doctrine of Equivalents
B. Developments in the Content of the Doctrine of Equivalents
C. Judge Dependence and the Doctrine of Equivalents
VI. Copying
VII. Conclusion
References
I. INTRODUCTION
Every patent is a “grant to the patentee, his heirs or assigns, of the right to exclude others
from making, using, offering for sale, or selling the invention throughout the United States
or importing the invention into the United States.” (35 U.S.C. § 154(a)).1 The action
for patent infringement2 provides the legal protection for a patentee’s right to exclude
** Professor of Law, Vachon Research Fellow, Loyola Law School, Loyola Marymount
University.
** Professor, Ferguson-Carlson Fellow in Law, and Director of the Innovation, Business, and
Law Program, University of Iowa Law School.
* 1 “[A]nd, if the invention is a process, of the right to exclude others from using, offering for
sale or selling throughout the United States, or importing into the United States, products made by
that process, referring to the specification for the particulars thereof.”
* 2 By an “action for patent infringement,” we refer to the liability determining aspects of a
suit involving such an action. We do not concern ourselves in this review with defenses and coun-
terclaims—for example, claims that the rights invoked are invalid—although we will occasionally
mention data concerning such defenses and counterclaims when necessary to provide appropriate
context to infringement observations reported in the studies we review.
326

Infringement 327
(35 U.S.C. § 271) and is conventionally viewed in economic terms.3 From an economic
perspective, the action for infringement exists for the purpose of supporting the incentive
structure imposed by the patent laws. This is so because: (1) it can be used to enforce rights
that allow innovators to recoup some of the investment they have made in providing an
innovation to the public, and (2) it at least loosely correlates what can be recouped to the
social benefits conferred by the innovation. The action for infringement can therefore
be thought of as existing to ensure that innovators get paid something for their social
contributions, and that the “something” is determined to some extent by the value of the
contribution made. It is easy to understand that, even if it operates imperfectly, the avail-
ability of the action could plausibly encourage the sort of risk-taking that occasionally
leads to valuable innovation.
For a patentee to succeed in an action for infringement, and thus vindicate its legally
protected interest, the patentee must prove that the “patented invention” is being made,
used, sold, offered for sale, or imported within or into the United States (35 U.S.C.
§ 271). While most cases, perhaps, focus on the question of whether the accused product
or process embodies the patented invention, disputes do arise over whether a patentee
can prove that it has been made, used, sold, offered for sale, or imported by an accused
tortfeasor (Transocean Offshore Deepwater Drilling, Inc. v. Maersk Contractors USA, Inc.,
617 F.3d 1296 (Fed. Cir. 2010).
In broad strokes—and leaving aside the doctrinal details surrounding the proving
of making, using, selling, offering for sale, or importing—doctrine requires a patentee
to follow a two-step process when seeking to prove that its right to exclude has been
infringed. First, patent claims must be construed. That process, known as claim construc-
tion, has been the subject of substantial study. Its central purpose is to translate the literal
text of a patent’s claims into a legally relevant context, such as that presented by a suit
claiming infringement. Second, the court or a factfinder will make an effort to determine
whether the properly construed claims “read on” to the accused device or process. The
process of “reading on” is an effort to determine whether each of the limitations defined
by the judicially construed claim language can be matched to elements (or features) of
the accused device or process.
When every limitation of a claim can be matched with an element in an accused device
or process, liability for infringement has been established. This form of infringement is
described as “literal”—as in, the text of the patent claim literally reads on to the accused
device or process. Importantly, however, it is not necessary that an infringer copy from
the patent in order to be found liable for literal infringement.
Complicating matters, liability for infringement is still possible when each and every
claim limitation cannot be literally matched with an element found in an accused device or
process. This is because infringement doctrine allows for a determination of liability when
the correspondence between the judicially construed claim language and the elements
of an accused device or process is not identical, but merely close enough. This form of
infringement is known variously as “equivalents” infringement, or “infringement under
the doctrine of equivalents.”
3
The studies we review either explicitly or implicitly rely on the economic theory of utility, or,
alternatively, are not particularly concerned with theory at all.

Both literal infringement and equivalents infringement are viewed as “direct” forms
of infringement. That is, they concern themselves with the liability of a tortfeasor that
makes, uses, sells, or imports devices or processes that infringe a patent. Like other tort
doctrines, however, patent infringement doctrine also contains features that allow for
“indirect,” or vicarious, liability. Patent law thus imposes liability on those who induce
or contribute to the direct infringement of a patent (35 U.S.C. § 271(b), (c); Limelight
Networks, Inc. v. Akamai Technologies, Inc., 134 S.Ct. 2111 (2014)).
Further complicating the law of infringement is a set of statutory provisions that
defines more narrowly drawn categories of infringing conduct. In large part, these
specialized forms of infringement address conduct that was at one time perceived as being
a loophole in existing infringement law, such as that addressed by 35 U.S.C. § 271(f); or
address specific policy goals, such as making the submission of an Abbreviated New Drug
Application (which itself does not constitute making, using, offering for sale, selling, or
importing into the United States a patented invention) a constructive act of infringement
(35 U.S.C. § 271(e)(2); Mossinghoff, 1999; Fisch et. al., 2004).
The doctrinal law surrounding claim construction, direct infringement, and indirect
infringement makes up much of the complexity of an action for infringement. Our review
avoids some of this complexity because claim construction receives separate treatment in
this book. For that reason, we mostly avoid this topic, discussing it only where necessary
to facilitate the discussion of ideas about infringement. In addition, only a single study
provides information concerning indirect infringement and it reports limited and mainly
summary information (Sichelman, 2010). Finally, we have located no studies systemati-
cally analyzing the statutory requirements of making, using, selling, offering for sale, or
importing a patented invention.
As we shall see, the body of available research that shows a systematic approach
to the issue of infringement is relatively small. More importantly, it is significantly
populated by studies that count litigation outcomes—for example, patentee success
in suits claiming infringement—and which do not discriminate along doctrinal lines
or seek to prove the utility consequences for patent policy. For this reason, the full
scope of our effort mainly reflects a review of the teachings of studies that have
systematically analyzed infringement in fairly general terms, and thus offer a limited
perspective from which to evaluate patent doctrine and policy. There is a perceptible
development, however, reflected in more recent studies, toward relating outcomes to
case and other real-world variables. This development offers an added level of descrip-
tive granularity to reported observations, and may in time come to support a more
robust theoretical assessment of patent law and policy. In any event, taken together,
the studies reviewed here suggest areas for future work that legal scholars may find
worth pursuing.
A few final points before we begin: a uniform limitation to interpreting the meaning of
the studies reviewed here is the risk of selection bias (Rantanen, 2016). Selection bias is a
form of statistical error that can occur when data selected for analysis is for some reason
not representative of the population a researcher seeks to analyze. It can sometimes
impact conclusions that might be drawn from collected data, but not always; and, as the
definition implies, its effects depend on the level of analysis and the nature of the claim
a researcher makes based on the data collected. In legal studies, the problem is nearly
ubiquitous—the basic concern being that in using written orders and opinions as data,

Infringement 329
there may be facets of the law and patterns of decisions incorrectly quantified because
they are for some reason less likely to appear in an analyzable writing.
It is difficult when planning a research project to know whether selection bias is likely
to cause a misrepresentation of the parameters of a population, and that predictive task is
even more difficult when reviewing the written work of others. It was obvious to us when
we agreed to write this review that selection bias presented the risk that the observations
reported in the studies we examine might not be representative of things as they exist
in the “real world” (although, of course, they might be). Nevertheless, we agreed to do
this review because even if some error bias exists, an analysis of the observations from
existing studies can encourage future work that improves on the validity, methodologies,
and data sources used. Note that because we do not know whether selection bias (or, for
that matter, measurement validity) has caused observational or inferential error, we mostly
proceed by addressing the reported observations in matter-of-fact terms, raising issues of
bias concern in only the most glaring of cases.
Another concern presented by observations reported in some of the studies we review
is reliability, which questions the consistency of the categorization of variables in the data
collection. For the studies we examine, much of what was collected seems to have been
relatively objective in nature—for example, was there a claim for infringement? How did it
come out? Other variables might be much more subjective in nature, such as characterizing
the subject matter of the patent suit or categorizing patent plaintiffs into special classes.
In each case, error can occur; although we suppose this less likely the more objective the
variable being measured, and not all of the studies we report used techniques to support a
claim to coding reliability. Since we lack the resources to redo all of the studies we discuss,
we proceed by addressing the reported observations in matter-of-fact terms, noting only
that those who plan to depend heavily on observations that cannot be confirmed as reli-
able should consider carefully whether the observations can be reproduced.
What follows is organized into five parts. First, we address the largest category of stud-
ies: those (mostly) counting general litigation outcomes from suits claiming infringement.
Second, we turn to a set of studies that relates general outcomes to other variables. Third,
we review studies reporting on the appellate treatment of infringement. We then turn
to what research shows concerning infringement under the doctrine of equivalents, and
finish with a special issue: the role of copying in infringement.
II. GENERAL LITIGATION OUTCOMES
Reports providing data about litigation outcomes on the issue of infringement can be
separated into two categories: outcomes on the question of infringement generally and
outcomes on infringement under the doctrine of equivalents. Reports about outcomes on
infringement generally often come from studies that examine infringement in the context
of other issues that arise in lawsuits concerning patent enforcement. These studies tend
not to distinguish between literal infringement and infringement under the doctrine
of equivalents, or between direct and indirect infringement. Exceptions include Risch
(2014a), which does separately report on literal and equivalents infringement; Sichelman
(2010), which reports separately on all four categories; and Allison and Lemley (2007) and
Petherbridge (2008, 2009, 2010), which report specifically on the doctrine of equivalents.

Another potentially informative dividing line is the creation of the Federal Circuit. The
establishment of the court is often described as a significant landmark in the political
development of patent law (Dreyfuss, 1989; Wagner and Petherbridge, 2004). Congress
certainly intended for the Federal Circuit to influence the patent law, and more commen-
tators than can reasonably be acknowledged have claimed, or repeated the claim, that it
has. We therefore begin by reviewing studies that report infringement litigation outcomes
from disputes preceding the creation of the Federal Circuit, and follow that with studies
that report infringement outcomes after the establishment of the Court.
A. Infringement Outcomes in the Pre-Federal Circuit Era
Perhaps the most prominent early report of patent infringement outcomes is Federico
(1956). Drawing on a mix of United States Patent Quarterly (USPQ) reported decisions,
as well as a subset of unpublished decisions for which outcome information could be
gathered from published decisions, Federico presents a three-tiered approach for coding
outcomes that offers some information about infringement. It focuses not on cases, but
on patents, and categorizes outcomes as follows: patents that are valid and infringed
(includes patents as long as any claims were infringed and not held invalid); patents that
are invalid (includes patents if any claims were held invalid and no claims were held valid
and infringed); and patents that are not infringed (includes the remaining share—that is,
patents in which the claims were held not infringed and not invalid, and patents that were
held not infringed where invalidity was not at issue).
Using data gathered this way, Federico (1956) reports that at the district courts, for
the period 1948–54, 30.3 percent of adjudicated patents were valid and infringed, 53.5
percent were invalid, and the additional 16.2 percent fell into the not-infringed category.
This work was followed by Koenig (1980), who, addressing the period 1953–77, reports
that 32.5 percent of patents were adjudicated valid and infringed at the district courts,
53.2 percent invalid, 6.3 percent valid and not infringed, and 8 percent not infringed with
validity not decided.
Subsequent work re-examining the period, again on a per-patent basis, reports that
during the pre-Federal Circuit era (1953–82), district courts concluded that a patent was
valid and infringed 31 percent of the time, invalid 56 percent of the time, and not infringed
13 percent of the time (Henry and Turner, 2006).
Taken together, these complementary observations suggest that from the middle of
the twentieth century to the time the Federal Circuit was created, judgments of liability
for patent infringement in district courts were realistic case outcomes. Patentees proved
a violation of a legally protected interest about one-third of the time. Moreover, given
the coding used in these reports, liability-causing acts were probably even more common.
The reason for saying this is that the coding rubrics used facially allocate at least some
instances of infringement to other categories when the patents involved had relevant
claims invalidated.
All of these studies rely on the USPQ for data, and we feel compelled to point out that
this data source presents special concerns of potential selection bias (Rantanen, 2016).
The reason is that while being a periodical that contains, among other types of intellectual
property cases, patent opinions issuing from the district and appellate courts, the USPQ
is not a periodical that contains all such opinions. Rather, it contains only those opinions

Infringement 331
Table 14.1 Outcomes for patents subjected to appellate review (1925–54) (Federico)
Valid and infringed Invalid Not infringed

Year Number Percent Number Percent Number Percent
1925–29 245 37.7 218 33.4 188 29.9
1930–34 223 31.1 274 38.0 222 30.9
1935–39 199 24.8 409 50.9 195 24.3
1940–44 92 15.3 364 60.5 146 24.2
1945–49 53 20.2 169 64.5 40 15.3
1950–54 61 18.2 204 60.7 71 21.1
selected by the editors for publication because, for example, the editors deem them to be
“potentially precedential or to include noteworthy fact patterns.” (Henry et. al., 2016).
This suggests that cases that are less noteworthy, or less likely to represent an important
precedent, are less likely to be in the database. A consequence of this may be that the
USPQ emphasizes judgments of invalidity, and situations where the lower tribunal is not
affirmed. It also likely systematically excludes non-noteworthy decisions and summary
affirmances. We do not know which way this cuts for outcomes—which is the main
problem—but one way it might cut is to emphasize invalidity relative to infringement. If
that is the case, it provides another reason to suspect that the infringement rate might have
been somewhat higher than the studies just discussed suggest.4
Some of the same studies offer a glimpse into appellate behavior before the creation
of the Federal Circuit. Federico (1956) reviewed all patent litigation appellate opinions
published in the USPQ for the period 1925–54, and reported numbers and shares of valid
and infringed, invalid, and not infringed patents on an annual and five-year basis. The
five-year summaries are provided in Table 14.1.
In addition, Federico (1956) reports that patent holders fared much worse in the
appellate courts than the district courts: of the 145 patents held valid and infringed by
district courts and reviewed during the period 1948–54, appellate courts “held 70 valid
and infringed and reversed the district courts in connection with 75 patents, holding them
invalid or not infringed;” whereas of “the 283 patents in which the district court had held
against the patentee there were reversals in only 6 instances…the court of appeals reversed
the district courts in only 2.1 percent of the instances of holdings against the patentee.”
Applying a similar methodology, Koenig (1980) reports decision shares at the courts of
appeals from 1953–1978. The five-year summaries are provided in Table 14.2.
Similar types of data are also reported by Dann (1960) for the period 1948–57,
Dearborn (1964) for the period 1953–63 (Koenig, 1980), and Henry and Turner (2006) for
the regional circuits after 1953 (essentially the period 1953–83). For that period, the Henry
and Turner study reports that 1068 (55 percent) patent cases resulted in an outcome of
invalid at the courts of appeal, 232 (12 percent) in an outcome of not infringed, and 627
(33 percent) in an outcome of valid and infringed.
4
We note that Henry and Turner (2016) do address this issue to some extent, through a set
of robustness checks.

Table 14.2 Outcomes for patents subjected to appellate litigation (1953–78) (Koenig)
Valid Not infringed Invalid

validity not decided
Infringed Not infringed
Year Number Percent Number Percent Number Percent Number Percent
1953–57 69 22.9 11 3.7 32 10.6 189 62.8
1958–62 88 32.0 17 6.2 22 8.0 148 53.8
1963–67 95 25.7 13 3.5 37 10.0 224 60.8
1968–72 81 28.2 13 4.5 19 6.6 174 60.7
1973–77 58 27.0 9 4.2 12 5.6 136 63.2
1978 8 44.4 0 0.0 0 0.0 10 55.6
In a study seeking to quantify selection effects on adjudications in patent disputes, Marco

(2004) reports patent holder win/loss rates on validity and infringement at all levels of
the judicial system for the period between 1970–97, reporting that patent owners won on
the issue of infringement in 59.6 percent (214 out of 359) of adjudications on that issue.
B. Infringement Outcomes in the Post-Federal Circuit Era
There are surprisingly few sources reporting general district court outcome data for
the post-Federal Circuit period. Henry and Turner (2006) report that from 1982–2002,
district courts reached a judgment that a patent was valid and infringed 37 percent of the
time, invalid 28 percent of the time, and valid and not infringed 35 percent of the time
(Henry and Turner, 2006). The remaining studies that generally report on the post-Federal
Circuit period use measurements that are likely to be heavily influenced by district court
judgments, and for that reason we report them here.
Allison et. al. (2014, 2015) count infringement outcomes at any level of the judicial system
(i.e., trial or appellate) as long as they occurred within a specified timeframe. Observations
of this sort can be dominated by trial level decisions (Allison and Lemley, 2007) for obvious
reasons, and the likely prominence of district court contributions to Allison et. al.’s reported
outcomes is the reason we include them here. For the cohort of cases filed in 2008 or 2009,
patent owners won on the issue of infringement 36.1 percent of the time (or in 197 out of
545 cases), and prevailed overall 25.8 percent of the time (164 out of 636 cases)—a rate
comfortably within the range established by earlier reports (Allison et al., 2015).
Of these studies, Allison et al. (2015) come the closest to reporting a “true” rate of
infringement—that is, the rate at which a patent is actually infringed, regardless of
whether it is valid or invalid. This rate is difficult to obtain due to the way patent litigation
operates: in order to prevail, a patent owner must win both validity and infringement,
whereas an accused infringer need only win on one issue or the other (Lemley, 2012). The
consequence is that a decision of invalidity produces no determination on infringement
and a decision of non-infringement typically produces no determination on invalidity.5
5
We say “typically” because, at least after the Supreme Court’s decision in Cardinal Chemical
Co. v. Morton Int’l, Inc., 508 U.S. 83 (1993), the Federal Circuit was required to address the issue of

Infringement 333
Thus, it is difficult, if not impossible, to ascertain which patents that were held invalid
would also have been held to be infringed and vice versa.
In a recently published study, Henry and Turner (2016) address this problem by
calculating ranges for infringement and validity determinations from the late 1920s to
2006. The methodology used does not allow for a particularly easy comparison with other
studies because it calculates bounded ranges for the “true” rates of infringement and
validity rather than reporting outcomes alone, and revolves around the determination of
breakpoints in the rates of infringement and invalidity (i.e., points in time when the rate
experienced a significant change). Nevertheless, this seems the most appropriate place in
which to describe the study.
Henry and Turner (2016) postulate two models for calculating upper and lower ranges
for the rates of infringement and validity during a given quarter of a year.6 Within these
ranges lies the “true” rate of invalidity and infringement for patents for a given year. While
not allowing for assessment of overall patentee “wins,” the approach employed in this study
may be more valuable for assessing changes in the reasons for patentee losses over time.
Under the bounded-ranges approach in the context of district courts plus appellate
courts (i.e., a dataset similar to that described above for Allison et. al.), Henry and Turner
(2016) report that when only patents that were found not invalid are considered, the “true”
rate of infringement was between 50.2–75.4 percent for 1927–51, 67.9–85.5 percent for
1951–90, and 38.2–55.9 percent for the period after 1990.7 (For reference, the rate at which
patents are held valid ranged from 43.4–58.2 percent for 1927–39, 22.6–40.1 percent for
1939–83, and 55.7–74.3 percent for the period after 1983.)
Taken together, these results suggest that, as a general matter, the observed rate of
infringement judgments at district courts may have remained relatively steady over the
roughly 60 years from 1948–2009. Patentees are proving a violation of a legally protected
interest in the right to exclude in around one-third of cases that are decided in ways meas-
urable by existing studies. In addition, Henry and Turner (2016) provide evidence that
the rate of patentee success on the issue of infringement might have decreased somewhat
during the period in which the Federal Circuit has been the main appellate court for patent
cases, even as success on validity has risen.
We encourage caution when interpreting these findings, however, particularly those
invalidity if it was raised by the accused infringer even if the court held the patent not infringed. In
these circumstances, a patent could be held both not infringed and invalid.
6
One limitation of Henry and Turner’s (2016) model’s rates is that they assume that
determinations of validity and infringement are independent, or at least inversely associated.
While there is theoretical basis for this assumption, available empirical evidence suggests
otherwise, at least at the district court (Moore, 2000): that judges and juries tend to decide in
the same direction when resolving both validity and infringement. If this is so, then one would
expect a higher rate of infringement when the population consists of patents whose validity was
undecided and patents for which the decision maker affirmatively rejected an invalidity argu-
ment, and a higher rate of validity when the population consists of patents whose validity was
undecided and patents for which the decision maker affirmatively rejected a non-infringement
argument.
7
The Henry and Turner (2016) study also reports district court-only data using this methodol-
ogy. The study reports a district court infringement rate of 58.4 percent for the period 1927–50,
71.2 percent for 1950–91 and 40.7 percent after 1991 when validity is senior.

that suggest that patentees have been less successful in proving infringement in the post-
Federal Circuit era, as these are the sorts of conclusions that, as the authors note, might
be especially influenced by changes in the legal standard or in the underlying population
rather than changes in the courts’ determinations as between patent owners and infring-
ers. For example, perhaps the most direct interpretation is that somehow the Federal
Circuit has directed district courts to constrict patentees’ legally protected interests. And,
indeed, as we shall see when it comes to infringement under the doctrine of equivalents,
there is good reason to think this may have happened to some extent. On the other hand,
perhaps something correlating to the creation of the Federal Circuit has encouraged
patentees to become more audacious litigants, in the sense that they became more likely
to bring claims nearer to the periphery of their legally protected interests. Or perhaps par-
ties are resolving cases more often on infringement than on validity for strategic reasons
(Ford, 2013). If so, that fact could prompt the appearance that courts are being harsher
to patent owners on infringement, even though the courts might still be applying a set of
standards not much different from (or perhaps even more generous than) those applied
in pre-Federal Circuit times.
These observations may be taken as suggesting other interpretations too, as well as
raising theoretical questions such as: what amount of empirical reassurance do innovators
need—in the form of other innovators successfully enforcing the right to exclude—in
order to be (optimally) stimulated by the patent system to the risk-taking that sometimes
leads to valuable innovation? On the evidence in these studies, no interpretations are
clear, and theoretical questions central to the economic validity of the patent system
are unanswerable. The value of these studies is that they offer the possibility of empiri-
cally grounding new perspectives and the possibility of new framings that might, with
additional work, illuminate better basic theoretical and doctrinal aspects of patent law.
III. OUTCOMES RELATED TO CASE, AND OTHER

VARIABLES
There is a perceptible trend in more recent studies toward relating outcomes to case
and other real-world variables. This development at a minimum offers an added level of
descriptive granularity to observations, and such approaches may in time come to support
a greater theoretical evaluation of patent law and policy. Here, we review observations in
this category, starting with some concerning the relationship between infringement and
posture.
A. Infringement and Procedural Posture
How patentees fare on infringement can depend on who is deciding the case (Moore,
2000). Observing district court judgments from patent infringement trials occurring
between 1983–99, Moore reports that of the 1359 decisions on infringement made by
these trials, 65 percent favored patentees and 35 percent favored accused infringers. The
study also suggests that patentees do significantly better in these trials if the fact finder
is a jury as opposed to a judge. Juries found infringement 71 percent of the time in 706
decisions; judges found infringement 59 percent of the time in 653 decisions.

Infringement 335
Moore also observes “herding,” a well-documented characteristic of legal decision

making in which judges and other fact finders tend to decide all issues in favor of the same
party. In this case, juries and judges tend to find patents either infringed and not invalid or
not infringed and invalid. Moore reports that juries did this 86 percent of the time in cases
where both infringement and validity were adjudicated, while judges did it 74 percent of
the time. The same study observes that being a plaintiff might in its own right be a benefit
in patent trials. When patentees were plaintiffs, they won 68 percent of trial decisions,
but that rate of success dropped to 52 percent when the case was a declaratory judgment
action filed by an accused infringer (success on validity dropped from 71 percent to 50
percent when the case was a declaratory judgment action).
The conclusion that a patentee’s legally protected interests have not been invaded by
an accused infringer is apparently much more suited to summary disposition than is
the alternative judgment. For the cohort of cases filed in 2008 or 2009, accused infring-
ers won 54 percent (257 of 473 cases) of their summary judgment motions alleging
non-infringement, a number that rises to 62 percent (316 of 509 cases) when including
stipulated judgments of non-infringement. Patentees, in contrast, won on 32 percent of
their motions for summary judgment of infringement, and had many fewer (128) motions
decided (Allison et. al., 2014).
The reason for this isn’t entirely clear, although one is tempted to speculate that it has
to do with a cost-benefit calculus that makes it economic for patent owners to bring suits
in the face of this sort of summary loss. Another possibility might be that because non-
infringement is essentially a determination of exclusion, it might be easier to argue than
infringement, which places the burden on a patent owner to convince a fact finder that
an accused device or process is included within the limitations defined by patent claims.
But if that is so, it would appear to be so only for summary judgments, as Moore (2000)
indicates that trials are more likely to go a patentee’s way, even when decided by a judge.
These observations might be reconciled, however, by the observation that a judge has to
allow a claim to go to trial and perhaps judges usually only do that when they think the
plaintiff has a pretty good case.
B. Infringement and Subject Matter
Using the method of counting any outcome at any level mentioned earlier, Allison et
al. (2015) observe that there is wide variation of patentee win rates on infringement by
technology—specifically, from 15 percent for optics (n=20), 19.7 percent for software
(n=193), 28.8 percent for electrical (n=59), 29.6 percent for biotech (n=27), and 43.2
percent for mechanical (n=148), to 68.4 percent for chemistry (n=98). The researchers
also examined infringement findings per industry and found that these too exhibit a broad
range (from 16.7 percent infringement findings for communications up to 68 percent for
pharmaceuticals).
Why this should be so isn’t entirely clear, and Allison et. al. (2015) do not claim to have
a definitive answer, leaving the reasons and theories that might be underlying the observa-
tions to future researchers. It is tempting to speculate, however, that infringement here is
in some way influenced by information within the four corners of the patent (Bessen and
Meurer, 2008). An oft-voiced idea is that patents in certain areas, most notably chemistry,
are usually considered to be comparatively (e.g., to software patents) clear in defining the

scope of the patented subject matter. Cases involving patents directed to such inventions
might turn out better for patentees, at least in part for the reason that it is easier to explain
to a judge or jury how, for example, a defendant’s chemical has the same formula as that
in the patent claim.
That said, perhaps the law itself plays a role in the observed variation, as pharmaceuti-
cal cases often involve chemical patents that arise under a special procedural process in
which the accused infringer must argue that it plans to make or use a chemical that will
have the same effect as the patented chemical when introduced into the body (Freilich,
2013). Infringement in many of these cases is often nearly conceded.
Considerations of this nature may explain Allison et. al.’s (2015) observations. But
there are still other reasons why infringement outcomes may depend on subject matter.
For example, perhaps juries have an especially generous perception of the social contribu-
tions of pharmaceutical companies, and are already inculcated with a perception that
patents might have a causal connection to that contribution. Or, perhaps, the clarity
of patents has the effect of discouraging both vague claims of infringement as well as
potential infringers who are able to determine for themselves that certain acts will likely
be found by courts to intrude on a patentee’s legally protected interests.
Observations of this sort, when reliable and valid, suggest valuable avenues for future
work. For example, some of these observations suggest the possibility that defining the
right to exclude in ways that make liability and validity more easily predicted by patent
owners and competitors could adjust the benefit and cost possibilities of the patent
system in ways that might end up creating more value for the public. If that interpretation
is correct, future work might build upon observations such as these by studying ways to
more usefully define the right to exclude conferred by the patent.
C. Special Plaintiffs
A policy dispute has recently developed around whether certain types of patent owners
should be able to enforce their patents. Some commentators argue that so-called non-
practicing entities (NPEs), or “patent trolls,” should be limited in their ability to assert the
rights that attend their patents (Risch, 2015; Freilich, 2016; Kesan, 2018). This dispute has
stimulated researchers to investigate how NPEs fare in patent infringement suits.
Across two studies, Michael Risch (2012) reports that NPEs lose on infringement
at a much higher rate than non-NPEs. He observes that of the 46 cases in his study of
NPEs “with merits rulings (forty of which directly ruled on infringement), only two
cases resulted in a finding of literal infringement, and no case found infringement by the
doctrine of equivalents.” In a subsequent study, Risch (2014) finds that when infringement
was determined on the merits for the NPE and non-NPE sets used in the study, non-NPEs
won on this issue much more often than NPEs: whereas NPEs obtained only three find-
ings of infringement out of the 119 adjudicated cases (and no findings of infringement by
equivalents), non-NPEs obtained a finding of literal infringement 46 times and a finding
of infringement by equivalents five times out of the 92 adjudicated cases.
Shrestha (2010) presents contrasting findings, observing that of the 29 cases in which
there was a judgment on the merits between 2000–08 in lawsuits by NPEs, patent owners
won seven, and seven other cases had a finding of no infringement (the remaining 15 cases
resulted in a win by the accused infringer on the basis of invalidity or unenforceability).

Infringement 337
In an unpublished manuscript, Freilich (2016) examines whether NPEs tend to win

on independent or dependent patent claims. Freilich finds that when NPEs win, they
tend to do so using the narrower dependent claims without needing to rely on broader
independent claims.
Like other studies relating outcomes to other variables, these studies open more
questions than they close. What, for example, is the theoretical significance of patent
enforcement by entities that do not make or use embodiments of the patents they own?
What is the meaning of observed differences between NPEs and non-NPEs when it comes
to successfully proving an intrusion into a legally protected interest? Are NPEs subject
to some sort of discrimination, and, if so, should the law care? Are NPEs systematically
bringing more audacious and speculative claims of infringement? If so, what should we
make of it and what, if anything, is a correct policy response?
While all of these studies need to be supported by future work before any policy
action is warranted, it should be clear that these sorts of studies may have implications
for policy prescriptions. If, for example, legislators wish to limit the social impact of
NPEs, one way to do so might be to restrict continuation practice in ways that make
it more difficult for NPEs to have issued claims that are specifically tailored to market
developments (Freilich, 2016). Alternatively, perhaps legislators may wish to reconsider
that aspect of the patent law that allows patentees to claim a legally protected interest
against prior uses, as has been done in connection with certain subject matter (35 U.S.C.
§ 273).
IV. APPELLATE TREATMENT OF INFRINGEMENT

JUDGMENTS
Dunner (1995) examined 1307 Federal Circuit opinions and orders including precedential
and non-precedential decisions (and Rule 36 summary affirmances) issued between
October 1, 1982, and March 15, 1994. The study reports that in appeals arising from the
district courts, the Court of Federal Claims, and the International Trade Commission, the
Federal Circuit reached a disposition of infringement 52 percent of the time and one of
no infringement 40 percent of the time, with the remaining 7 percent vacated. In reaching
these conclusions, it affirmed district court findings of infringement 85 percent of the
time and of no infringement 77 percent of the time.
Moore (2000) finds that about 40 percent of the issues appealed from patent trials
involve infringement, and observes that across the period studied (1983–99) the Federal
Circuit affirmed jury decisions on infringement 77 percent of the time and judge decisions
on infringement 82 percent of the time. (This was comparable with other issues—validity
was affirmed 78 percent of the time overall; 76 percent of the time—but lower than others,
viz. willfulness affirmed 94 percent of the time for juries and 80 percent of the time for
judges).
Drawing on data collected by PatStats.org, a project of the University of Houston Law
Center, Sichelman (2010) reports that the Federal Circuit reversed district courts on the
issues of literal direct infringement and infringement under the doctrine of equivalents 15
percent of the time, contributory infringement 16 percent of the time, and inducement of
infringement 13 percent of the time for the period 2000–07, all of which lie on the lower

half of the issues that the Federal Circuit reviewed.8 The study caveats these reversal rates
as absolute metrics, however, due to a concern that they may exclude summary affirm-
ances for the 2000–04 period.
Kesan and Ball (2011) report a study of patent cases filed between 1995–2003 and fol-
lowed through to resolution. It observes that of the appeals of the district court decisions
in their dataset, the issue of infringement made up 237 out of the 809 total observations.
Using a “somewrong” standard, the study reports that the Federal Circuit reversed 27
percent of the time on the issue of infringement; using an “allwrong” standard, it reports
that the Federal Circuit reversed on infringement 20 percent of the time.9
Henry and Turner (2006)10 examine both the rate at which decisions of infringement
were appealed and the rate at which those decisions were affirmed. During the pre-Federal
Circuit period, decisions of not invalid and infringed were appealed 56 percent of the
time and affirmed 60 percent of the time. Decisions of not infringed but not invalid were
appealed 48 percent of the time and affirmed 69 percent of the time. And decisions of
invalidity were appealed 55 percent of the time and affirmed 85 percent of the time.
After the Federal Circuit took over, appeals of decisions of not invalid and infringed
were appealed 61 percent of the time and affirmed 72 percent11 of the time. Decisions of
not infringed but not invalid were appealed 60 percent of the time and affirmed 69 percent
of the time. And decisions of invalidity were appealed 69 percent of the time and affirmed
57 percent of the time.
V. THE DOCTRINE OF EQUIVALENTS
Arguably, research focused on infringement under the doctrine of equivalents has pro-
vided somewhat greater insight into the doctrinal law than has other research concerning
the action for patent infringement. Here, we turn to a review of the studies that present
those data.
8
Values inferred from Figure 1 of Sichelman (2010).
9
Note that there is some degree of uncertainty here because there were 137 rulings that could
not be classified, some of which could be Rule 36 affirmances where the issue was uncertain from
the briefs, and another 168 cases that had not been resolved as of their study end date.
10
One of the authors (Rantanen) has substantial concerns about placing too much weight on
the appeal and affirmance rates reported in Henry (2006), due to the nature of that study’s dataset.
Specifically, the researchers drew on the USPQ for their set of district and appellate decisions, with
the additional modification of tracking the case history of the reported district court opinions
through Westlaw (to the extent possible). However, as discussed above, the USPQ is not a complete
set of all patent litigation decisions, as it does not include at least some non-precedential opinions
or any summary affirmances under Federal Circuit Rule 36 (which the court began issuing in
1989 (Dunner, 1995)). This produces two study design effects when examining appeals. First, it
potentially selects district court opinions that are more significant and thus potentially more likely
to be appealed. In addition, it potentially results in the systematic exclusion of at least some Rule 36
summary affirmances, an exclusion that will bias the apparent reversal rate upwards (Moore, 2005).
11
Again, Rantanen expresses caution based on the apparent exclusion of summary affirm-
ances from the dataset.

Infringement 339
A. Outcomes and the Doctrine of Equivalents
The general picture from written opinions—the only sort represented in existing
research—is that patentees are successful between roughly one in four and one in 20
times when litigating issues involving the doctrine of equivalents. Whether patentees are
closer to the one in four or closer to the one in 20 rate of success depends on: (1) what one
counts as patentee success, (2) whether one measures district courts separately from the
Federal Circuit, and (3) what period one examines. Research shows that patentee success
of any sort was highest in the early 1990s—as far back as observations go—and that the
rate trended steadily downward from that point until at least 2007. Patentees have also
tended to be modestly more successful in the district courts than in the Federal Circuit.
Allison and Lemley (2007) report patentee success under the doctrine of equivalents
for a period spanning May 1999, through August 2005. Combining 217 district court and
196 Federal Circuit “decisions” found in Westlaw for the period, patentees were observed
to be successful on the doctrine of equivalents about 24 percent of the time. Because
the researchers defined patentee success as any successful outcome for a patentee on the
issue of the doctrine of equivalents, the 24 percent rate of “patentee success” includes
things such as surviving an accused infringer’s motion for summary judgment of no
infringement under the doctrine of equivalents, or having an adverse judgment of no
infringement under the doctrine of equivalents vacated on appeal. To account for the fact
that patentees may have gone on to lose at least some of these cases on the theory that
there was no infringement under the doctrine of equivalents, the researchers examined
how often patentees were likely to ultimately win cases when the doctrine of equivalents
was the theory of infringement. They found that patentees ultimately win fewer than 10
percent (29 of 413 decisions) of cases involving the doctrine of equivalents.
Petherbridge (2010) examines patentee success on the doctrine of equivalents using
878 Federal Circuit analyses12 of the doctrine of equivalents authored from January,
1992 through May, 2007. Although he uses a different methodology to define data entries
and measure patentee success, the observed rates of patentee success on the doctrine of
equivalents are similar to those reported by Allison and Lemley (2007).13 Measuring
patentee success as a Federal Circuit analysis that expresses the judgment that an accused
infringer is liable for equivalents infringement, Petherbridge reports a downward trend
in overall patentee success. For the period from 1992 to roughly 1998, approximately 18
percent of Federal Circuit analyses of the doctrine of equivalents express the judgment
that an accused infringer is liable for infringement under the theory. That rate of success
declines, however, as one moves toward 2007; and by the end of the period studied, only
about 5 percent of Federal Circuit analyses express such a judgment.
At the same time, however, Federal Circuit analyses became much less dispositive of
appeals, apparently due in part to the influence of conflicting methodological approaches
to claim construction used by Federal Circuit judges. To reduce the influence of claim
12
An analysis is an individual decision of the court on the question of infringement under the
doctrine of equivalents. There are therefore more analyses than both cases or opinions, although
the differences are not large, as most opinions offer only a single analysis.
13
Petherbridge reports graphs showing a 30-analysis moving average of decisions.

construction-based remands on observations, and offer a somewhat more absolute

measure of ultimate patentee success on the doctrine of equivalents, Petherbridge also
examines patentee success in a subset of analyses in which the Federal Circuit’s analysis
evinced a dispositive outcome. From this perspective, approximately 20 percent of Federal
Circuit analyses of the doctrine of equivalents express the judgment that an accused
infringer is liable for equivalents infringement across the period of 1992 to roughly 1998.
As one moves toward 2007, during which time claim construction-based remands become
much more common, that rate of success trends down to under 10 percent.
Addressing part of this same period, Allison and Lemley (2007) report that patentees
won 23.3 percent of the time at the Federal Circuit during the period covering the 18
months preceding the Federal Circuit’s decision in Markman v. Westview Instruments, 52
F.3d 967 (Fed. Cir. 1995). It is not clear whether the reported wins are dispositive, as in
winning the case, or whether it is “patentee success”, as Allison and Lemley more broadly
use that term. It represents, in any event, a similar rate to that reported in Petherbridge
(2010) (roughly 20 percent from 1992–98), so taken together, the studies that have looked
at written opinions appear to find similar levels of patentee success across overlapping
periods of time. Another feature of the seemingly poor results patentees achieve with the
doctrine of equivalents is an increase in the adverse-to-patentee disposition of the issue
on summary judgment. Allison and Lemley (2007) report that over two-thirds of the
419 decisions in their data arise from motions for summary judgment of no equivalents
infringement.
There is also a positive correlation between an increase in the rate of incoming appeals
of summary judgment of no equivalents infringement and the rate at which the Federal
Circuit affirms those appeals (Petherbridge, 2010). Between 1992–2007, the rate of
Federal Circuit equivalents analyses addressing summary judgments of non-infringement
under the doctrine of equivalents trends upward from just under 50 percent to over 75
percent.14 As the rate of incoming summary judgments of non-infringement increases,
the rate at which those judgments are affirmed correspondingly increases, from roughly
20 percent to 55 percent.
Taken together, these results suggest two significant features of doctrine of equivalents
jurisprudence: first, it appears to have become increasingly amenable to disposition on
summary judgment; and, second, it has become increasingly hostile to patentee success.
B. Developments in the Content of the Doctrine of Equivalents
The observations just mentioned suggest the possibility of a change in the doctrinal
content of the doctrine of equivalents, and researchers have pursued evidence of doctrinal
changes that could have the capacity to explain the apparent decline of equivalents as a
viable theory of liability. (Allison and Lemley, 2007; Petherbridge, 2008, 2010; Schwartz,
2011). The evidence brought to bear suggests that at least some of the reason for
14
Of significant note, cases appealed to the Federal Circuit appear to broadly reflect the
outcome characteristics of the district court judgments reported in Allison and Lemley. This
suggests that the cases patentees decide to appeal are not an excessively biased sample of district
court judgments.

Infringement 341
s eemingly poor patentee success on the doctrine of equivalents has to do with a Supreme
Court-encouraged assault on the doctrine of equivalents that integrates procedure and
legal rules.
The procedure part of the explanation, speculated by Allison and Lemley (2007), is
that the rise of Markman (1995) hearings introduced a procedural framework to patent
infringement cases that ended up killing the doctrine of equivalents. The basic idea is as
follows:
● By making claim construction a question of law, the Markman decision caused

judges to construe claims at pre-trial “Markman” hearings.
● A decision on claim construction is often dispositive of the question of literal
infringement, so having a Markman hearing increases the likelihood that a trial
judge will grant summary judgment of literal infringement.
● Patentees usually claim literal infringement and infringement under the doctrine of
equivalents.
● To have a final and therefore appealable judgment, the equivalents claim has to be
decided (interlocutory appeals are not generally available for claim construction).
● In cases where judges have summarily determined no literal infringement, the most
efficient and straightforward path to case disposal or a review of the claim construc-
tion is by also summarily deciding the equivalents claim.
● Judges therefore have a strong incentive to also decide the equivalents claim on
summary judgement.
The rules part of the explanation (Adams, 2006; Petherbridge, 2010) is that the Supreme
Court rather explicitly encouraged the Federal Circuit to develop rules that can be used
to effect summary judgment of non-infringement when it comes to the doctrine of
equivalents,15 and that the Federal Circuit followed that guidance by developing and
reinvigorating a host of legal limitations to the use of the doctrine of equivalents—all
questions of law that could therefore be operated as part of the claim construction
inquiry. These developments provided the tools judges needed to write orders and appel-
late decisions summarily rejecting claims of infringement by equivalents.
In substance, the legal limitations are a list of rules that are used to deny patentees
access to the doctrine of equivalents.16 The most well known include:
● amendment-based prosecution history estoppel;

● argument-based prosecution history estoppel;
● the all elements rule;
● prohibition on a range of equivalents that would capture prior art;
15
The central piece of case law support for this is the Supreme Court’s Warner-Jenkinson
Co. v. Hilton Davis Chemical Co., 520 U.S. 17 (1997) opinion (especially n. 8), but complemen-
tary instruction can be taken from the Court’s Markman v. Westview Instruments, Inc., 517
U.S. 370 (1996) and Festo Corp. v. Shoketsu Kinzoku Kogyo Kabushiki Co., 535 U.S. 722 (2002)
decisions.
16
There are no reported doctrines that operate in the other direction—that is, to expand
access to the doctrine of equivalents.

● prohibition on a range of equivalents that would capture subject matter disclosed,

but not claimed in a patent specification—and is thus dedicated to the public;
● specification disclaimer or disavowal of a range of equivalents; and
● limitations of the doctrine in some instances to after arising technologies.
In addition, it has been shown that patent law offers two infringement standards govern-
ing the equivalents comparison of properly construed claims to an accused device or
process: the insubstantial differences test and the function-way-result test
Neither of the standards provides for much structured analysis and therefore might not
be expected to limit infringement analyses. They are probably more usefully understood as
symbolizing the nature of the factual comparison that needs to be undertaken (assuming
a court ever reaches that stage in the analysis). In addition, they are generally analyzed
after any relevant legal limitations because a conclusion on any of the legal limitations
may eliminate the need to have a jury—or other factfinder—consider infringement in view
of these standards.
Some observations that support a role for procedure in the decline of the doctrine of
equivalents include the following:
● A decline in patentee success on the doctrine of equivalents correlates with a large

increase in the appearance of claim construction in connection with equivalents
analyses (Petherbridge, 2008; Schwartz, 2011).
● A decline in patentee success on the doctrine of equivalents correlates with a large
increase in appeals addressing summary judgment of no equivalents infringement
(Allison and Lemley, 2007; Petherbridge, 2010).
● Patentee success on the doctrine of equivalents is very sensitive to judicial treatment
of claim construction in equivalents analyses (Petherbridge, 2008).
● In a model statistically controlling for analyses that apply legal limitations, infringe-
ment standards, and the presence of a claim construction analyses, appeals from
summary judgments of no equivalents infringement predict patentee failure on the
issue of the doctrine of equivalents (Petherbridge, 2010).
This last observation also suggests that patentees are losing because of the application
of legal rules—and not just because the underlying model offers evidence for that inter-
pretation. After all, courts have to offer some explanation for why patentees’ equivalents
claims fail. Arguing that there is no genuine issue of fact on the intensely factual questions
presented by the application of an infringement standard clearly seems less desirable than
relying on a legal rule that a judge may interpret as prohibiting altogether access to the
range of equivalents a patentee seeks.
Some additional observations that support the idea that legal limitations have played a
role in the decline of the doctrine of equivalents17 include the following:
17
Somewhat surprisingly, studies have affirmatively rejected a special role for the doctrine
laid down in the Festo case concerning prosecution history estoppel (Allison and Lemley, 2007;
Petherbridge, 2010), which is not to say it presents no problems for patentees, or the patent system
(Lichtman, 2004).

Infringement 343
● The Federal Circuit has addressed the large increase in appeals from summary
judgment of no equivalents infringement by affirming many of the judgments
(Petherbridge, 2010).
● In a model statistically controlling for procedural posture and the presence of a
claim construction analysis, the presence of an analysis of a legal limitation strongly
predicts patentee failure in a period during which claim construction became a
prominent feature of Federal Circuit equivalents analyses (1999–2007), but not in
an earlier period (1992–99) (Petherbridge, 2010).18
● The analysis in Federal Circuit opinions of legal limitations to the doctrine of
equivalents predicts Federal Circuit reversal of lower-court judgments of equiva-
lents infringement (Petherbridge, 2010).
A few concluding thoughts: first, it seems unlikely that claim construction driven proce-
dural changes and legal rules are the entirety of the story of the decline of the doctrine
of equivalents. Another element may be simple judicial hostility to the doctrine. It is
hard to examine the Federal Circuit’s work with the doctrine—including the existence
of the procedural developments and the direction of the development of legal rules so
far highlighted by studies—without forming the impression that many of the court’s
judges are not fans of the doctrine. Just why that might be is not clear, and will need to
be the subject of future work, but one place to begin might be with the Federal Circuit’s
mandate itself. A more uniform and predictable patent law may be no place for a robust
doctrine of equivalents; and for judges who are intent on following Congress’s guidance
on the broader goals for patent law, the doctrine of equivalents might have to be strictly
limited.
Second, the evidence supporting the just described mechanism underlying the decline
in patentee success on the doctrine of equivalents suggests that claim construction (with
associated doctrines) has become a dominating lever of the judicial treatment of the
doctrine of equivalents. Due to the centrality of claim construction to all major patent
doctrines, what has happened with the doctrine of equivalents may be a model for what
might occur or be occurring with other major patent doctrines (Petherbridge, 2008).
C. Judge Dependence and the Doctrine of Equivalents
The doctrine of equivalents is like some other patent doctrines in that outcomes can
depend on which judges hear the case (Wagner and Petherbridge 2004; Petherbridge et
al., 2011). Panel participation data shows that Federal Circuit judges reach significantly
different outcomes on the question of infringement by equivalents (Petherbridge, 2009).
Judicial differences do not appear to be exclusively a function of when judges were
appointed, and are therefore not entirely explainable in terms of broader changes in the
court’s view of the doctrine.
18
Specifically, that there is an inverse correlation between the rate at which the Federal Circuit
modifies claim constructions in equivalents analyses and the rate at which those analyses are
affirmed. In addition, in models, Federal Circuit analyses that do not modify claim construction
predict affirmances on the issue of the doctrine of equivalents—an outcome that is usually adverse
to patentee success on the doctrine.

Judges Rader, Lourie, Clevenger, Schall, and Bryson, for example, were all
appointed between 1990–94, and authored opinions across a 1992–2007 study period
(Petherbridge, 2009). Significant differences in patentee success both on the equivalents
appeal itself and on the ultimate disposition of equivalents claims have been observed
that depend on the panel participation of these judges. The same study also looks at
the 10 most active judicial authors between January 1, 2001 and December 31, 2005
and finds that equivalents outcomes differ significantly across these judicial authors
during the period.
Other judicial divisions of relevance to the doctrine of equivalents also support the
conclusion of judge dependency. Empirical work has identified a set of Federal Circuit
judges that exhibits a distinct methodological approach to claim construction (Wagner
and Petherbridge, 2004; 2012). These judges, known in the literature as “proceduralists,”
approach claim construction in a way that gives primary weight to the general, com-
monly understood meaning that patent claim language would have had to ordinarily
skilled artisans at the time of invention (Wagner and Petherbridge, 2016). The procedural
approach tends to prioritize, inter alia, the notice function of claiming. A strong view of
the notice function of claiming is, of course, antithetical to the broad application of the
doctrine of equivalents.
Other empirically identified approaches to claim construction may be less likely to
come into conflict with the doctrine of equivalents because they do not seem to emphasize
the notice function of patent claims as much as the procedural approach. For example,
Wagner and Petherbridge (2004) identify so-called “holistic” judges. These judges exhibit,
in a sense, a methodological commitment to claim construction analysis. But it differs
from the procedural approach in that it tends to disregard commonly held art-specific
understandings of the meaning of claim terms. Instead, it favors a patent idiosyncratic
interpretation of claim language based (mainly) on the technical teachings of the speci-
fication or prosecution history. Finally, “swing” judges are judges who do not appear to
have much of a commitment to methodological approach and can be found using both
procedural and holistic approaches more often than other judges.
Patentee success—both on the equivalents appeal itself and on ultimate disposition of
equivalents claims—depends on the panel participation of judges associating with these
distinct approaches to claim construction analysis (Petherbridge, 2009).
When researchers examined opinion authorship, and its connection to legal limita-
tions and patentee success, they observed that the effects seen for panel participation
were largely amplified. Proceduralists are the least likely to author opinions favoring
patentees. Swing judges—judges who appear not to have a firm commitment to either of
the two empirically identified methodological approaches to claim construction—are the
most friendly to patentees in their opinions; and holistic judges—who, as noted above,
appear committed to a patent idiosyncratic interpretation of claim language—are, like
proceduralists, relatively unfriendly to patentees on the issue of infringement under the
doctrine of equivalents.
Not all groupings of judges show significantly different treatment of appeals concern-
ing the doctrine of equivalents under all conditions. For example, there is no evidence that
judges with patent backgrounds are more or less friendly to the doctrine of equivalents
than their counterparts when panel participation is examined (Petherbridge, 2009).
However, when authorship is examined, judges with patent backgrounds tended to

Infringement 345
author more opinions favoring patentees on the doctrine of equivalents, in some cases
significantly more. Judges without patent backgrounds, when authoring opinions, tended
to be less generous to patentees when it came to the doctrine of equivalents and, in some
cases, were significantly less so.
VI. COPYING
Copying is not an element that needs to be met to establish infringement liability, although
it can be relevant to the issue of willfulness in infringement. Nevertheless, a narrative
equating infringement to copying has always been present. Referring to the policy under-
lying equivalents infringement the Supreme Court had this to say:
to permit imitation of a patented invention which does not copy every literal detail would be
to convert the protection of the patent grant into a hollow and useless thing…. leav[ing] room
for—indeed encouraging—the unscrupulous copyist to make unimportant and insubstantial
changes.... Graver Tank & Mfg. Co. v. Linde Air Prods. Co., 339 U.S. 605 (1950).
This sort of rhetoric makes it seem as though perhaps patent infringement involves a
“wrong” that correlates with a culpable mental state, and in furtherance of this thought
Graver Tank later describes a copier as a “pirate.” But even if the language the court chose
is hyperbolic, it is still of great interest whether infringers are in fact copying patented
inventions. If they are, then perhaps infringers are benefiting from the technical teachings
patentees provide. If not, and infringers are innocently stumbling into costly infringement
disputes, then perhaps the patent system has serious inefficiencies that adversely impact
innovation and deserve closer attention.
The role of copying in connection with infringing acts has not been exhaustively stud-
ied. Only one study has examined the degree to which patent infringement suits actually
involve allegations of copying (Cotropia and Lemley, 2009). From a sample of 200 patent
infringement complaints filed between January 1, 2000, and May 1, 2007, collected from
the District of Delaware and the Eastern District of Texas, it observes that only 10.9
percent “alleged that the defendant had copied the invention, either from the patent or
from the plaintiff’s commercial product.” In addition, allegations of copying can depend
on subject matter, with ranges from “a low of 0% in optics and in semiconductors to a
high of 65% in pharmaceuticals.” The authors also examined 1871 infringement opinions
from Westlaw for the period December 31, 2005–March 1, 2008, and found that only 129
(6.9 percent) mentioned an allegation of copying.
Cotropia and Lemley (2009) is a nice example of a paper that uses research to highlight
an issue that could, with future work, have broad policy implications. Its central observa-
tion that copying is rarely alleged suggests the possibility that the copying of patents
or patented products is relatively rare. That said, the study by Cotropia and Lemley is
limited to the extent that it was conducted for a period when pleading requirements were
substantially looser than they are at present. Furthermore, there is no obligation for
patent owners to allege copying, at least for purposes of direct infringement.
If it is true that the copying of patents or patented products is relatively rare, it may
follow that much infringement is “innocent,” in that competitors may usually bring to
market goods developed without any assistance from a patentee that later seeks to use

the law to obtain additional rents, or even an injunction. If this is typically what happens,
then important aspects of economic theory are implicated. An important theoretical
component of the economic validity of the patent system as currently constituted is
founded on the economic benefits supposed to follow from the broad dissemination of
knowledge that is thought to come from granting patents. That knowledge is supposed to
reduce the likelihood that future competitors will need to “reinvent the wheel,” improve
competition by illuminating paths for future innovation, and properly allocate capital by
showing what has been accomplished and enclosed with law by others. It is also supposed
to protect competition from being lost to innocent infringement, or to the incentives
created by the risk of it.
Indeed, if it could be shown empirically that these theoretical expectations for the
current patent system are largely invalid, it would be a very big deal. It might well require
nothing less than a retheorizing of the patent system.
So, while Cotropia and Lemley (2009) may not have shown that the copying of patents
and embodiments of patented inventions does not occur at a level many might require
to justify revamping the entire patent system, this sort of research not only raises in a
substantial way new questions, but inspires thinking about how more and better data
about copying might be gathered so that those who are concerned with patent law and
policy can be more certain whether it aligns with the theory that justifies it.
VII. CONCLUSION
While the new insights into infringement offered by recent research helpfully improve our
understanding, they also highlight how little is really known and how much more needs
to be done in order for legal scholars to properly understand the nature and role of this
important feature of patent law.
To begin with, research has had very little to say about literal infringement. This is
something of a surprise, and perhaps the reason for it is the sense that most disputes
concerning literal infringement do not present genuine issues of fact once claim construc-
tion has been decided.
While sensible—especially when one considers how influential claim construction
appears to be to the judicial treatment of the doctrine of equivalents—it still seems
that there should be infringement issues that claim construction alone cannot resolve.
For example, there is the question of whether the defendant’s acts actually violated
§ 271—whether it made, used, offered for sale, sold, or imported an infringing product, for
example. Halo Electronics, Inc. v. Pulse Electronics, Inc., F.3d 1371 (Fed. Cir. 2014). There
are also cases involving factual questions of whether the accused product or method meets
the limitation of the claims. Catalina Lighting, Inc. v. Lamps Plus, Inc., 295 F.3d 1277
(Fed. Cir. 2002). In addition, Kesan and Ball (2008) offer the tantalizing finding that over
one-quarter of the 809 cases examined involved issues of infringement “based on grounds
unrelated to claim construction” (issues of infringement involving claim construction, by
comparison, had 236 observations). Taken together, there could be a substantial number
of infringement issues left unresolved by claim construction that have not received much
attention from researchers. We think this area deserves at least some investigation in the
future.

Infringement 347
Research also has had next to nothing to say about indirect infringement, and foreign
infringement. These seem to be fresh pastures for future research.
Finally, much of what we do know is expressed in terms of fairly general outcome
counts, although some studies have begun to offer information about infringement out-
comes and cases and related them to real-world variables. Studies by Allison, for example,
have reported information about subject matter/industry; while Risch, inter alia, has
reported evidence about the impact of patent owner identity on infringement outcomes.
Clearly, however, there is much more to know.
REFERENCES
Adams, Charles W. 2006. “The Doctrine of Equivalents: Becoming a Derelict on the Waters of Patent Law”,
84 Nebraska Law Review 1113–57.
Allison, John R., and Mark A. Lemley. 2007. “The (Unnoticed) Demise of the Doctrine of Equivalents”, 59
Allison, John R., Mark A. Lemley, and David L. Schwartz. 2014. “Understanding the Realities of Modern
Allison, John R., Mark A. Lemley, and David L. Schwartz. 2015. “Our Divided Patent System”, 82 University
Allison, John R., Mark A. Lemley, and Joshua Walker. 2011. “Patent Quality and Settlement Among Repeat
Patent Litigants”, 99 Georgetown Law Journal. 677–712.
Barney, James R., and Charles T. Collins-Chase. 2011. “An Empirical Analysis of District Court Claim
Construction Decisions, January to December 2009”, 2011 Stanford Technology Law Review 1–22.
Innovators at Risk. Princeton: Princeton University Press.
Coolley, Ronald B. 1989. “What the Federal Circuit Has Done and How Often: Statistical Study of the CAFC
Patent Decisions—1982 to 1988”, 71 Journal of the Patent and Trademark Office Society 385–98.
Cotropia, Christopher A., and Mark A. Lemley. 2009. “Copying in Patent Law”, 87 North Carolina Law Review
1421–66.
Dann, C. Marshall. 1960. “Floor Discussion of Mr. Cooch’s Paper”, Dynamics of the Patent System 54.
Dearborn, Richard J., and R. Bradlee Boal. 1964. “Adjudications by Circuits and Arts Involved”, The
Encyclopedia of Patent Practice and Invention Management 22.
Dreyfuss, Rochelle C. 1989. “The Federal Circuit: A Case Study in Specialized Courts”, 64 New York University
Law Review 1–76.
Dunner, Donald R., J. Michael Jakes, and Jeffrey D. Karceski. 1995. “A Statistical Look at the Federal Circuit’s
Patent Decisions: 1982–1994”, 5 Federal Circuit Bar Journal 151–80.
Federico, P. J. 1956. “Adjudicated Patents”, 1948–54:38 Journal of the Patent and Trademark Office Society 233–49.
Fisch, Alan M., and Brett H. Allen. 2004. “The Application of Domestic Patent Law to Exported Software: 35
U.S.C. § 271(f)”, 25 University of Pennsylvania Journal of International Law 557–91.
Ford, Roger Allen. 2013. “Patent Invalidity Versus Noninfringement”, 99 Cornell Law Review 71–128.
Freilich, Janet. 2013. “The Paradox of Legal Equivalents and Scientific Equivalence: Reconciling Patent Law’s
Doctrine of Equivalents with the FDA’s Bioequivalence Requirement”, 66 Southern Methodist University
Freilich, Janet. 2016. Pushing Patent Boundaries: An Empirical Assessment of How Patent Trolls and Other
Winning Litigants Use Patent Scope (unpublished manuscript).
Henry, Matthew D., and John L. Turner. 2006. “The Court of Appeals for the Federal Circuit’s Impact on Patent
Litigation”, 35 Journal of Legal Studies 85–117.
Henry, Matthew D., and John L. Turner. 2016. “Across Five Eras: Patent Enforcement in the United States
1929–2006”, 13 Journal of Empirical Legal Studies 454–86.
Henry, Matthew D., Thomas P. McGahee, and John L. Turner. 2013. “Dynamics of Patent Precedent and
Enforcement: An Introduction to The UGA Patent Litigation Datafile”, University of Georgia Working
Paper, available at http://ssrn.com/abstract=2228103.
Kesan, Jay P. 2018. “Patent Trolls”, in Peter S. Menell and David L. Schwartz, eds., Research Handbook on the
Economics of Intellectual Property, Vol. II. Cheltenham: Edward Elgar Publishing.
Kesan, Jay P., and Gwendolyn G. Ball. 2006. “How Are Patent Cases Resolved? An Empirical Examination of
the Adjudication and Settlement of Patent Disputes”, 84 Washington University Law Review 237–312.

Kesan, Jay P., and Gwendolyn G. Ball. 2011. “Judicial Experience and the Efficiency and Accuracy of Patent
of Law and Technology 393–467.
Koenig, Gloria K. 1980. Patent Invalidity: A Statistical and Substantive Analysis. New York: Clark Boardman.
Lemley, Mark A. 2012. “The Fractioning of Patent Law”, in Shyamkrishna Balganesh, ed., Intellectual Property
and the Common Law, Cambridge University Press 504–13.
Lichtman, Douglas. 2004. “Rethinking Prosecution History Estoppel”, 71 University of Chicago Law Review
151–82.
Marco, Alan C. 2004. “The Selection Effects (and Lack Thereof) in Patent Litigation: Evidence from Trials”, 4
Topics in Economic Analysis & Policy 1–45.
Moore, Kimberly A. 2000. “Judges, Juries, and Patent Cases—An Empirical Peek Inside the Black Box”, 99
Moore, Kimberly A. 2004. “Empirical Statistics on Willful Patent Infringement”, 14 Federal Circuit Bar Journal
227–40.
Moore, Kimberly A. 2005. “Markman Eight Years Later: Is Claim Construction More Predictable?”, 9 Lewis
and Clark Law Review 231–47.
Mossinghoff, Gerald J. 1999. “Overview of the Hatch-Waxman Act and its Impact on the Drug Development
Process”, 54 Food & Drug Law Journal 187–94.
Paul, Darcy. 2003. “The Judicial Doctrine of Equivalents”, 17 Harvard Journal of Law and Technology 248–82.
Petherbridge, Lee. 2008. “The Claim Construction Effect”, 15 Michigan Telecommunication and Technology
Petherbridge, Lee. 2009. “Patent Law Uniformity?”, 22 Harvard Journal of Law and Technology 421–73.
Petherbridge, Lee. 2010. “On the Decline of the Doctrine of Equivalents”, 31 Cardozo Law Review 1371–405.
Rantanen, Jason. 2016. “Empirical Analyses of Judicial Opinions: Methodology, Metrics, and the Federal
Circuit”, 49 Connecticut Law Review 227–91.
Risch, Michael. 2012. “Patent Troll Myths”, 42 Seton Hall Law Review 457–99.
Risch, Michael. 2015a. “A Generation of Patent Litigation: Outcomes and Patent Quality”, 52 San Diego Law
Review 67–132.
Risch, Michael. 2015b. “The Layered Patent System”, 101 Iowa Law Review 1535–579.
Schwartz, David L. 2010. “Pre-Markman Reversal Rates”, 43 Loyola of Los Angeles Law Review 1074–107.
Schwartz, David L. 2011. “Explaining the Demise of the Doctrine of Equivalents”, 26 Berkeley Technology
Seaman, Christopher B. 2012. “Willful Patent Infringement and Enhanced Damages After In Re Seagate: An
Empirical Study”, 97 Iowa Law Review 417–71.
Shrestha, Sannu K. 2010. “Trolls or Market Makers? An Empirical Analysis of Nonpracticing Entities”, 110
Sichelman, Ted. 2010. “Myths of (Un)Certainty at the Federal Circuit”, 43 Loyola of Los Angeles Law Review
1161–94.
Wagner, Polk, and Lee Petherbridge. 2004. “Is the Federal Circuit Succeeding? An Empirical Assessment of
Judicial Performance”, 152 Pennsylvania Law Review 1105–80.
Wagner, Polk, and Lee Petherbridge. 2012. “Did Phillips Change Anything? Empirical Analysis of the Federal
Circuit’s Claim Construction Jurisprudence”, in Shyamkrishna Balganesh, ed., Intellectual Property and the
Common Law, Cambridge University Press 123–50.
Wagner, Polk, and Lee Petherbridge. 2018. “Teva and the Process of Claim Construction”, 70 Florida Law
Review 379.
Case List
Cardinal Chemical Co. v. Morton Int’l, Inc. (1993), 508 U.S. 83.
Catalina Lighting, Inc. v. Lamps Plus, Inc. (2002), 295 F.3d 1277.
Festo Corp. v. Shoketsu Kinzoku Kogyo Kabushiki Co. (2002), 535 U.S. 722.
Graver Tank & Mfg. Co. v. Linde Air Prods. Co. (1950), 339 U.S. 605.
Halo Electronics, Inc. v. Pulse Electronics, Inc. (2014), 769 F.3d 1371.
Limelight Networks, Inc. v. Akamai Technologies, Inc. (2014), 134 S.Ct. 2111.
Markman v. Westview Instruments, Inc. (1995), 52 F.3d 967.
Transocean Offshore Deepwater Drilling, Inc. v. Maersk Contractors USA, Inc. (2010), 617 F.3d 1296.
Warner Jenkinson Co., Inc. v. Hilton Davis Chemical Co. (1997), 520 U.S. 17.

Infringement 349
35 U.S.C. § 154.
35 U.S.C. § 271.
35 U.S.C. § 273.

15. Presumption of validity
Christopher B. Seaman* 19
Contents
I. Introduction
II. Historical Development of the Presumption of Validity
A. Statutory Text
B. Case Law before i4i
C. Microsoft Corp. v. i4i Limited Partnership
III. Theory Regarding the Presumption of Validity
A. Justifications for the Presumption
B. Critiques of the Presumption
IV. Empirical Studies on the Presumption of Validity
A. Invalidity Rates in Patent Litigation
B. Standard of Proof to Overcome the Presumption
1. Observational studies
2. Experimental studies
V. Implications and Areas for Future Research
VI. Conclusion
References
I. INTRODUCTION
By statute, every patent issued by the U.S. Patent and Trademark Office (USPTO) is
entitled to a presumption of validity in litigation (35 U.S.C. § 282). Although Congress
placed the burden of establishing invalidity on the challenger, it failed to explicitly specify
the particular standard of proof necessary to overcome this presumption. This issue
was the subject of a Supreme Court decision, Microsoft Corp. v. i4i Limited Partnership,
564 U.S. 91 (2011), which held a patent’s invalidity must be established by clear and
convincing evidence, even when the USPTO did not consider some of the prior art raised
in the invalidity challenge during prosecution of the patent-in-suit. However, the Court
simultaneously held a jury may be instructed that such unconsidered prior art may “carry
more weight and go further toward sustaining” the challenger’s burden of persuasion to
demonstrate invalidity.
Several empirical studies have attempted to assess the impact of the presumption of
validity in Microsoft v. i4i’s wake (Chatlynne, 2010; Chatlynne et al., 2011; Schwartz
and Seaman, 2013; Bock, 2015). These studies generally support the conclusion that
* Associate Professor of Law, Washington and Lee University School of Law.
350

Presumption of validity 351
the clear-and-convincing evidence standard required to overcome the presumption can

have a material impact on invalidity decisions, at least on the margins. In addition, jury
instructions regarding the impact of prior art not previously considered by the USPTO
may have a statistically significant impact on invalidity outcomes.
The balance of this chapter is organized as follows. Section II recounts the development
of the presumption of validity, including the Court’s decision in Microsoft v. i4i. Section
III presents the theoretical justifications for the presumption of validity, as well as several
critiques of the presumption. Section IV assesses the existing empirical literature on the
presumption of validity. Section V considers the implications of these empirical studies,
as well as unresolved issues and areas for future research. Section VI concludes.
II. HISTORICAL DEVELOPMENT OF THE PRESUMPTION OF

VALIDITY
A. Statutory Text
The presumption of validity is a longstanding feature of U.S. patent law. Since at least
the nineteenth century, the party challenging a patent’s validity—usually an accused
infringer—has had the burden of proof on this issue.1 Early Supreme Court decisions
recognized this basic principle as well, concluding that “the burden to prove [invalidity]
is on the [accused infringer], not only because they make the charge, but because the
presumption arising from letters patent is the other way”. Agawam Woolen Co. v. Jordan,
74 U.S. 583 (1868).2
In the 1952 Patent Act, Congress expressly codified the presumption of validity,
providing that “[a] patent shall be presumed valid” and that “[t]he burden of establishing
invalidity of a patent or any claim thereof shall rest on the party asserting such invalidity”
(35 U.S.C. § 282). The Act’s legislative history explained that this provision “introduce[d]
a declaration of the presumption of validity, which is now a statement made by courts in
decisions, but has had no expression in the statute” (H.R. Rep. No. 82-1923, 1952; S. Rep.
No. 82-1979, 1952). Similarly, P.J. Federico’s influential commentary on the 1952 Patent
Act stated that:
section 282 declares that a patent shall be presumed valid and that the burden of establishing
invalidity of a patent shall rest on a party asserting it. That a patent is presumed valid was the
law prior to the new statute, but it was not expressed in the old statute. The statement of the
presumption in the statute should give it greater dignity and effectiveness (Federico, 1993).
1
For example, the 1946 amendments to the Patent Act provided that in an infringement
action, a party “may plead” and “prove on trial” the invalidity of the patent as a defense (35 U.S.C.
§ 69, 1946). Older versions of the Patent Act used similar language (Patent Act, § 61 (1870); Patent
Act, § 15 (1836); Patent Act, § 6 (1793)).
2
See also Coffin v. Ogden (1873) (concluding that “[t]he burden of proof rests upon” the
accused infringer to demonstrate that “the patentee was not the original and first inventor of the
things patented, or of a substantial and material part claimed to be new”).

B. Case Law before i4i
Although the Patent Act establishes a presumption of validity, there is no textual require-
ment for a particular evidentiary standard to overcome this presumption. Indeed, prior
to 1952, “the case law was far from consistent—even contradictory” about the strength of
the presumption. Am. Hoist & Derrick Co. v. Sowa & Sons, Inc., 725 F.2d 1350 (Fed. Cir.
1984). Several nineteenth-century Supreme Court decisions suggested that invalidity of
an issued patent had to be established beyond a reasonable doubt; Cantrell v. Wallick, 117
U.S. 689 (1886); Coffin v. Ogden, 85 U.S. 120 (1873); while later cases in some lower federal
courts concluded that a lower standard applied. Munising Paper Co. v. Am. Sulphite Pulp
Co., 228 F. 700 (6th Cir. 1915); Potthoff v. Hanson & Van Winkle Co., 174 F. 983 (3d Cir.
1909).
In Radio Corporation of America v. Radio Engineering Labs, Inc., 293 U.S. 1 (1934)
(RCA), the Supreme Court held that a heightened standard of proof was required to
overcome the presumption of validity. The issue in RCA was the validity of the two
patents-in-suit; the defendant conceded infringement, but argued both patents were
invalid because “they were issued to a patentee who was not the first inventor.” Affirming
the district court’s decision that the patents were not invalid, the Court explained that “a
patent regularly issued . . . is presumed to be valid until the presumption has been over-
come by convincing evidence of error.” While the Court acknowledged that the strength
“of that presumption has found varying expression in this and other courts,” it concluded
that the party asserting invalidity ultimately “bears a heavy burden of persuasion, and
fails unless his evidence has more than a dubious preponderance.” In other words, the
Court explained, “the presumption of validity shall prevail . . . unless the countervailing
evidence is clear and satisfactory.”
Following RCA, lower federal courts generally required a challenger to demonstrate
invalidity by “clear and convincing evidence,” at least when the challenge was based on
the prior art previously considered by the USPTO. Tenneco Chems., Inc. v. William T.
Burnett & Co., 691 F.2d 658 (4th Cir. 1982); Carson Mfg. Co. v. Carsonite Int’l Corp., 686
F.2d 665 (9th Cir. 1981).3 But if the examiner had not considered the allegedly invalidat-
ing prior art, courts often concluded that the presumption was “weakened” or “largely
dissipated”. Jacuzzi Bros., Inc. v. Berkeley Pump Co., 191 F.2d 632 (9th Cir. 1951); Gillette
Safety Razor Co. v. Cliff Weil Cigar Co., 107 F.2d 105 (4th Cir. 1939); H. Schindler & Co.
v. C. Saladino & Sons, 81 F.2d 649 (1st Cir. 1936); Butler Mfg. Co. v. Enterprise Cleaning
Co., 81 F.2d 711 (8th Cir. 1936). As one court explained, the presumption of validity
should be reduced “when pertinent prior art is shown not to have been considered by the
Patent Office” because:
the Examiner’s expertise may have been applied to an incomplete set of data[,] and there can be
no certainty that he would have arrived at the same conclusion in the face of the evidence and
3
But see Dickstein v. Seventy Corp., 522 F.2d 1294 (6th Cir. 1975) (concluding that “a
preponderance of evidence is sufficient to establish invalidity” in the “usual” patent case”); Rains
v. Niaqua, Inc., 406 F.2d 275 (2d Cir. 1969) (“Reasonable doubt on the issue of validity must
be resolved in favor of the patent holder, but in the usual case a preponderance of the evidence
determines the issue”).

argument presented to the court. Chicago Rawhide Mfg. Co. v. Crane Packing Co., 523 F.2d 452
(7th Cir. 1975).
At least two Circuits went even further, holding that the proper standard of proof in such
situations was by a preponderance of the evidence. Tenneco Chems., Inc. v. William T.
Burnett & Co., (1982); Carson (1981).
Since 1982, the U.S. Court of Appeals for the Federal Circuit—which has exclusive
jurisdiction over appeals from claims arising under the Patent Act—has consistently
required “clear and convincing evidence” to overcome the presumption of validity.4 It
did not alter this standard even when the prior art at issue in the invalidity challenge was
not before the USPTO during prosecution. Kahn v. Gen. Motors Corp., 135 F.3d 1472
(Fed. Cir. 1998); Connell v. Sears, Roebuck & Co., 722 F.2d 1542 (Fed. Cir. 1983). However,
the Federal Circuit acknowledged that it may be “easier to overcome the presumption” if
the examiner “fail[ed] to consider the most relevant prior art”. Trans-World Mfg. Corp. v.
Al Nyman & Sons, Inc., 750 F.2d 1552 (Fed. Cir. 1984).
In a 2007 decision, the Supreme Court appeared to open the door to a lower standard
of proof in invalidity challenges when the relevant prior art was not considered by the
USPTO. In KSR International Co. v. Teleflex, Inc., 550 U.S. 398 (2007), the Court unani-
mously reversed the Federal Circuit’s reliance on the so-called “teaching, suggestion, or
motivation” test for determining obviousness under § 103. Near the end of the opinion,
the Court declined to address whether the patentee’s failure to disclose a key prior art
patent “voids the presumption of validity given to issued patents,” and instead concluded
that the claim at issue was “obvious despite the presumption.” But the Court then noted
(apparently in dictum) that “the rationale underlying the presumption—that the PTO, in
its expertise, has approved the claim—seems much diminished here.”
C. Microsoft Corp. v. i4i Limited Partnership
Four years after KSR, the U.S. Supreme Court directly confronted in Microsoft Corp. v.
i4i Limited Partnership (2011) the unresolved question of the standard of proof for over-
coming the presumption of validity. In the lower courts the accused infringer, Microsoft,
unsuccessfully challenged the validity of plaintiff i4i’s patent, alleging that i4i had sold
software embodying the invention more than one year before it applied for a patent (the
“on-sale bar”). At trial, Microsoft argued that it should be required to prove invalidity
by only a preponderance of the evidence because evidence of the prior art sale of i4i’s
4
See, for example, Ultra-Tex Surfaces, Inc. v. Hill Bros. Chem. Co., 204 F.3d 1360 (Fed. Cir.
2000) (“An accused infringer alleging that a claim is invalid must overcome the statutory presump-
tion of validity that attaches to an issued patent . . . by proving invalidity by facts supported by
clear and convincing evidence”); Greenwood v. Hattori Seiko Co., 900 F.2d 238 (Fed. Cir. 1990)
(“Under 35 U.S.C. § 282, one attacking the validity of a patent must present clear and convincing
evidence establishing facts which lead to the legal conclusion that the patent is invalid”); Am. Hoist
& Derrick Co. v. Sowa & Sons, Inc. (1984) (“[Section] 282 creates a presumption that a patent is
valid and imposes the burden of proving invalidity on the attacker. That burden is constant and
never changes and is to convince the court of invalidity by clear evidence”); see also Microsoft v. i4i
(2011) (explaining that “[i]n the nearly 20 years since American Hoist, the Federal Circuit has never
wavered in [its] interpretation of § 282”).

software had not been presented to the USPTO during prosecution of the patent-in-suit.
The district court rejected Microsoft’s argument, holding that Microsoft had the burden
of proving invalidity by clear and convincing evidence, which it failed to do (i4i Limited
Partnership v. Microsoft Corp., 670 F. Supp. 2d 568 (E.D. Tex 2009), and the Federal
Circuit affirmed. i4i Limited Partnership v. Microsoft Corp., 598 F.3d 831 (Fed. Cir. 2010).
The Court granted certiorari on the issue of the standard of proof for Microsoft’s invalid-
ity defense. Microsoft Corp. v. i4i Limited Partnership, 562 U.S. 1060 (2010).
At the Supreme Court, Microsoft advanced two arguments in support of reversal (Brief
for Petitioner Microsoft Corp., 2011). First, it asserted that since § 282 was silent regard-
ing the standard of proof for finding a patent invalid, an accused infringer need only prove
a patent claim invalid by a preponderance of the evidence, which is the ordinary burden
of proof in civil cases. Second, Microsoft—citing KSR—argued in the alternative that
the preponderance of the evidence standard was required when the party challenging the
patent’s validity relied on prior art not previously considered by the USPTO.
The Court unanimously rejected both of Microsoft’s proposals. First, analyzing § 282,
the Court explained that while the statute “by its express terms . . . establishes a presump-
tion of patent validity, and it provides that a challenger must overcome that presumption
to prevail on an invalidity defense,” it “includes no express articulation of the standard of
proof ” (Microsoft Corp. v. i4i Limited Partnership, 2011). Citing RCA, the Court reasoned
the presumption of validity had acquired “a settled meaning in the common law” that “a
defendant raising an invalidity challenge bore a heavy burden of persuasion, requiring
proof of the defense by clear and convincing evidence.” Thus, when Congress enacted
§ 282 to provide that an issued patent is “presumed valid,” the Court concluded that this
presumption “encompassed not only an allocation of the burden of proof but also an
imposition of a heightened standard of proof.”
As to Microsoft’s second argument, the Court held that the standard of proof did not
vary depending on whether the USPTO had previously considered the prior art raised in
the invalidity challenge. It explained that “[o]ur pre-1952 cases never adopted or endorsed
the kind of fluctuating standard of proof that Microsoft envisions,” and that nothing
in § 282 “suggests that Congress meant to . . . enact a standard of proof that would rise
and fall with the facts of each case” (Microsoft Corp. v. i4i Limited Partnership, 2011).
However, the Court also recognized that “new evidence supporting an invalidity defense
may carry more weight in an infringement action than evidence previously considered by
the PTO,” and thus “the challenger’s burden to persuade . . . o[n] its invalidity defense may
be easier to sustain.” Consequently, “the jury may be instructed to consider that it has
heard evidence that the PTO had no opportunity to evaluate before granting the patent,”
although the Court declined to endorse any particular formulation for such an instruction.
III. THEORY REGARDING THE PRESUMPTION OF VALIDITY
A. Justifications for the Presumption
Several theoretical justifications have been offered in support of the presumption of

validity. The most commonly asserted justification is deference to the USPTO’s expertise
in examining patents (Lichtman and Lemley, 2007; Devlin, 2008). As one scholar notes:

the PTO is a specialized administrative agency, comprised of individuals with scientific expertise
in their respective areas of competence and charged, amongst other things, with determining
whether a given application meets the requirements of patentability. Given the sophistication
of this body … it surely makes sense that patents be presumptively deemed valid in subsequent
challenges (Devlin, 2008).
The Federal Circuit has rationalized the presumption in similar terms, asserting that:
deference . . . is due to a qualified government agency presumed to have properly done its job,
which includes one or more examiners who are assumed to have some expertise in interpreting
the references and to be familiar from their work with the level of skill in the art and whose duty
it is to issue only valid patents (Am. Hoist & Derrick Co. v. Sowa & Sons, Inc., 1984).
As a result, the presumption “forces courts to defer to the expertise of the PTO, thereby
avoiding redundant and possibly inferior second looks” on the issue of validity (Lichtman
and Lemley, 2007).
Another proffered justification is the promotion of stable patent rights.5 Bringing a
patented technology to market is a risky enterprise that requires substantial investment in
labor and capital (Kieff, 2001; Lichtman and Lemley, 2007). The presumption of validity
may help “reduce[] the risk associated with those investments” because “[a] patent holder
whose patent benefits from a presumption knows that, if his development and com-
mercialization efforts turn out to be successful, he will likely have a valid patent that will
empower him to exploit that success” (Lichtman and Lemley, 2007). This presumption
is particularly valuable in fields with high research, development, and commercialization
costs, such as pharmaceuticals, where successfully bringing a patented product to market
may cost hundreds of millions of dollars (DiMasi et al., 2003; Lichtman and Lemley,
2007). It also comports with the rationale for requiring a clear and convincing standard
of proof in other areas of the law, as a heightened standard is often required when
particularly important interests are at stake. Santosky v. Kramer, 455 U.S. 745 (1982);
Addington v. Texas, 441 U.S. 418 (1979).
Third, a strong presumption of validity reflects the asymmetric stakes in patent invalid-
ity disputes, as a decision rendering a claim invalid has preclusive effect in future litigation,
while a decision that a claim is not invalid does not bar subsequent validity challenges by
others (Sumida, 2011; Ford, 2013). In other words, a patentee need only lose once on inva-
lidity to terminate the patent right. Thus, a rebuttable presumption of validity is likely to
decrease the risk that a court will not enforce a valid patent (Ayres and Klemperer, 1999).
Finally, some scholars have attempted to justify the presumption in part based on its
“expressive” function—in other words, the signaling effect both inside and outside the
patent community of stating that patents are valid unless clearly demonstrated otherwise
(Janis, 2004). Indeed, Federico’s commentary on the 1952 Patent Act provides some sup-
port for this claim, as he acknowledged that the presumption’s codification in § 282 served
to “give it greater dignity and effectiveness” (Federico, 1993). Weakening or eliminating
this presumption, as some have advocated, would send an expressive message as well—
“that patents are less secure and the patent system deserves less respect” (Janis, 2004).
5
See, for example, Magnivision, Inc. v. Bonneau Co., 115 F.3d 956 (Fed. Cir. 1997) (“The
purpose of the presumption of validity . . . is to contribute stability to the grant of patent rights”).

B. Critiques of the Presumption
Several critiques of the presumption of validity have also been articulated. One such
critique is that well-known constraints on the USPTO’s ability to rigorously examine
all patent applications undermine the rationale that its decisions should be accorded
substantial deference in court. This includes the ex parte nature of patent prosecution,
which deprives the examiner of the benefits of an opposing viewpoint when reviewing
an application in most cases (Lemley, 2001).6 In addition, there is no requirement that
a patent applicant must search for relevant prior art (Lemley, 2001; Kesan, 2002), and
examiners typically lack access to certain types of potentially relevant prior art, such as
a prior “public use” or sale of the invention, unless this information is known to and
disclosed by the applicant (Leslie, 2011; Taylor, 2011). Information asymmetries may
exist between the inventor(s) and the USPTO as well, as the applicant often knows more
about the relevant field(s) of technology than the examiner reviewing the application
(Taylor, 2011; Seymore, 2015). Moreover, the burden of establishing that a patent should
not be granted lies with the USPTO, as a patent application is deemed to presumptively
comply with the requirements for patentability upon filing (Lemley, 2001; Seymore,
2013; Seymore, 2015). Taken together, critics of the presumption of validity contend
these limitations result in an “information-poor process” of examination that permits the
granting of weak patents (Lichtman and Lemley, 2007).
Some scholars have also argued that deference to the USPTO’s decisions regarding
patentability is unwarranted because resource constraints mean that examiners are not
able to adequately vet all applications. On average, patent examiners are estimated to
spend less than 20 hours per application (Lemley, 2001), meaning that a full-time exam-
iner would be expected to examine over 100 patent applications annually. At the same
time, the number of patent applications filed continues to increase, with over 600,000
patent applications filed in 2017—more than double the amount filed in 2000 (USPTO,
2018a). And despite additional revenue-setting authority granted to the USPTO by the
Leahy-Smith America Invents Act (AIA), the pendency of patent examination continues
to be lengthy, with an average total time of 23.7 months as of January 2019 (USPTO,
2019).7 These limitations may cause patent examiners to engage in a less rigorous review
of pending applications. For example, one recent empirical study found that “as examin-
ers are given less time to review applications, the less prior art they cite, the less likely they
are to make time-consuming prior art rejections, and the more likely they are to grant
patents” (Frakes and Wasserman, 2017). In contrast, patent infringement defendants
have a strong incentive to extensively search for relevant prior art and prepare invalidity
arguments (Lemley, 2001).
Another critique of the presumption of validity is that it imposes undesirable private
and social costs. Under the deferential “clear and convincing evidence” standard, at least
some patents of dubious validity will be upheld in court (Lemley and Shapiro, 2005). As
6
One exception created by the America Invents Act permits third parties to submit patents,
published patent applications, or other printed publications of potential relevance to the examiner
of a patent application, along with a concise description of the alleged relevance of each document
submitted (35 U.S.C. § 122(a), 2012).
7
This figure excludes Requests for Continued Examination.

Mark Lemley has explained, jurors in patent cases “are notoriously reluctant to second-
guess patent examiners, not only because they assume examiners know more than they
do but because they believe examiners spend much more time examining any given patent
than they actually do” (Lemley, 2001). The presumption may also cause accused infringers
to agree to a license “rather than fight even a bad patent in court” (Lemley, 2001). When
combined with the presumption of validity, the mere existence of a patent—even one of
dubious validity—may cause potential competitors to not enter a market, thus permitting
the patentee to charge supra-competitive prices (Lemley, 2001; Leslie, 2006).
Based on these critiques, numerous academics, government agencies, and others have
advocated reducing the standard of proof to overcome the presumption of validity, at
least when the invalidity claim relies on prior art not considered by the USPTO (Lemley,
2001; Federal Trade Commission, 2003; Lichtman and Lemley, 2007; Devlin, 2008;
Daniel, 2008; Alsup, 2009; Taylor, 2011; Ford, 2013). Others contend that altering the
clear-and-convincing standard may adversely affect innovation by increasing the uncer-
tainty about a patent’s enforceability (Kesan and Banik, 2000; Janis, 2004; Chatlynne,
2009; Morrisett, 2012).
IV. EMPIRICAL STUDIES ON THE PRESUMPTION OF

VALIDITY
This section describes the state of empirical studies on the presumption of validity. It first
covers empirical scholarship on invalidity findings in patent cases. Next, it discusses the
impact of the presumption of validity and the standard of proof required to overcome it.
A. Invalidity Rates in Patent Litigation
The validity of the patent(s)-in-suit is never decided in the vast majority of patent cases,
as most disputes are resolved before a final decision on the merits (Kesan and Ball, 2006).
When a merits decision is reached, however, several empirical studies have reported that
patents are found invalid at a relatively high rate considering the strong presumption of
validity.
Looking at Federal Circuit decisions from 1982–94, Dunner et al. (1995) reported that
32 percent of asserted patents were found invalid as anticipated under § 102. The same
study found the Federal Circuit held 50 percent of patents invalid as obvious under § 103.
In a study of final validity decisions in patent cases between 1989–96 published in U.S.
Patents Quarterly, Allison and Lemley (1998) found that nearly half (46 percent) of the
299 patents at issue were adjudged invalid.8 The authors also found that in cases where
patents were held invalid, the invalidity challenger disproportionately relied on prior art
not considered by the USPTO.9 In a recent follow-up to this study, Allison, Lemley and
8
The most common basis for invalidity was obviousness under § 103 (42 percent of invalid
patents), followed by § 102 non-prior art (31 percent of invalid patents) (Allison and Lemley, 1998).
9
Specifically, Allison and Lemley (1998) found that the probability of invalidity based on
unconsidered prior art was 41 percent, while the probability of invalidity based on previously
considered prior art was under 30 percent.

Schwartz (2014) evaluated all substantive decisions rendered in patent cases filed in 2008
and 2009 and found that accused infringers won 42.4 percent of all invalidity challenges.10
B. Standard of Proof to Overcome the Presumption
Prior to the i4i decision, scholars and others debated whether the standard of proof
required to overcome the presumption of validity had a significant impact on invalidity
decisions.11 Lichtman and Lemley (2007) contended that “[w]hile we can’t prove that
presumptions matter, we believe that they likely do, at least at the margins.” Similarly,
Alsup (2009) asserted that the clear-and-convincing evidence creates “a huge advantage
for the patent holder.” In contrast, Janis (2004) argued that “one possible outcome of the
proposed change to preponderance standard for overcoming the presumption of validity”
is that it might cause “little difference in the outcomes of cases” as “judges will reach the
same result that they would have under the old standard.” More recently, the Supreme
Court observed in Commil USA, LLC v. Cisco Systems, Inc., 135 S. Ct. 1920 (2015), that
the “clear and convincing standard” sets a “high bar” for establishing invalidity.
Empirical studies have attempted to assess this issue in two ways: observationally and
experimentally. Observational studies use data from actual events, while experimental
studies use controlled experiments on human subjects to answer research and policy
questions (Croson, 2002). For example, an experiment attempting to evaluate the impact
of jury instructions on litigation outcomes can present subjects with information about
a hypothetical dispute, randomly give them one of several different jury instructions, and
then ask each subject to reach a decision as if they were a juror in the case (Schwartz and
Seaman, 2013; Bock, 2015). Both types of studies regarding the presumption of validity
are discussed in more detail below.
1. Observational studies

In one observational study, Chatlynne (2010) analyzed 45 invalidity challenges (in 33
cases) at the Federal Circuit between April 2008 and June 2009 to evaluate whether the
invalidity determination was dependent on the standard of proof. After reviewing the
procedural posture and outcome, the author concluded that 16 percent of invalidity
challenges (seven of 45) were potentially “evidentiary-standard dependent,” meaning
they could have resulted in a different decision under a preponderance standard.12 As a
10
Allison, Lemley and Schwartz (2014) report that the most common bases for finding a
patent invalid are obviousness under § 103, prior art under § 102, and indefiniteness. However, the
study also notes that invalidity determinations based on patentable subject matter under § 101 have
increased since the prior study.
11
Some cases also support the claim that the presumption of validity matters. See, for example,
Aventis Pharma Deutschland GmbH v. Lupin Ltd., 2006 WL 2008962 (E.D. Va. 2006) (“If the
preponderance of the evidence standard was the standard to judge this case, the Court might agree
with [the accused infringer], but . . . that is not the standard to be applied here. . . . A patent is
presumed valid and invalidity must be shown by clear and convincing evidence”).
12
Chatlynne (2010) concluded that any cases where a finding of invalidity was upheld after
trial, or where summary judgment was granted, were “evidentiary-standard independent” because
these decisions would have come out of the same way under a (less deferential) preponderance
standard. The remaining decisions were considered “evidentiary-standard dependent.”

result, the author contended that this “relatively small percentage” of cases “suggests that
the evidentiary standard is not, in itself, a substantial driver in invalidity determinations.”
In a second observational study, Chatlynne, Kenny and Watkins (2011) compared a first
set of invalidity decisions in lower federal courts from 1966 through 1982 with a second
set of invalidity decisions by the Federal Circuit from 2008–10. The authors of the study
asserted that courts used a preponderance standard when the party challenging validity
introduced evidence not previously considered by the USPTO in the first set of cases,
while the second set of cases occurred when a clear and convincing evidence standard
applied at all times. The study failed to find a statistically significant difference in invalid-
ity outcomes between the two time periods (32 percent versus 34 percent, respectively),13
and the authors concluded that the standard of proof “may not be as significant a driver
in invalidity outcomes as some have suggested.”
There are several reasons, however, to view the findings of these observational studies
with some caution. First, both studies involve small sample sizes. Chatlynne (2010) relied
on 45 observations, which the author conceded was “too small to draw firm conclusions.”
Chatlynne, Kenny and Watkins (2011) relied on a total of 108 observations from the
combined time periods, but the authors did not conduct a power analysis to determine
whether their inability to reject the null hypothesis (i.e., that there was no material differ-
ence in invalidity outcomes between the two time periods) represented a Type II error.14
Second, the court decisions relied on in these observational studies are subject to the
well-known problem of selection effects. Specifically, court decisions do not represent a
random sample of overall disputes because the litigating parties will consider informa-
tion pertinent to the likelihood of success on a claim—such as changes to the relevant
standard of proof—and adjust their strategy accordingly, such as resolving a dispute by
settlement when it appears the chances of success in court are low (Priest and Klein, 1984;
Clermont and Eisenberg, 1992). Third, and perhaps most significantly, there have been
several major changes in patent litigation between the two time periods (1966–82 and
2008–10) studied in Chatlynne, Kenny and Watkins (2011). These changes include the
requirement that a judge interpret a patent’s claims (Markman v. Westview Instruments,
Inc., 517 U.S. 370 (1996)), the shift from judges to juries as the primary factfinder in most
patent trials,15 and the increased importance of summary judgment in resolving invalidity
claims (Schwartz and Seaman, 2013; Allison et al., 2014). The authors conceded that they
did not attempt to control for these and other changes.16
13
However, Chatlynne, Kenny and Watkins (2011) did report that they found a statistically
significant difference in the first set of decisions for invalidity decisions that appeared to apply a
preponderance standard (32 percent invalidity rate, eight of 25 cases) compared to decisions that
appeared to apply a clear and convincing evidence standard (13 percent invalidity rate, six of 45
cases).
14
Power analysis is used to calculate the minimum sample size required to detect an effect of
a particular size (Cohen 1992).
15
One study found that juries tried less than 3 percent of patent cases from 1968–1970, com-
pared to almost 60 percent of patent cases from 1997–1999 (Moore, 2000).
16
Chatlynne, Kenny and Watkins (2011) state:
The comparison herein of data from periods separated by more than twenty-five years
surely implicates additional factors such as changes in applicable practice, procedure, search

2. Experimental studies

Two recent experimental studies have attempted to assess the impact of the presumption
of validity and the standard of proof to overcome it. Schwartz and Seaman (2013)
conducted a survey experiment where subjects (N = 500) were asked to decide invalidity
under three different evidentiary standards. Subjects were recruited online17 and provided
with a short fact pattern regarding an invalidity claim adapted from a real-life patent
lawsuit.18 In this fact pattern, the party challenging the patent’s validity introduced and
relied upon prior art which was not considered by the USPTO during examination of
the patent-in-suit. Each subject then received one of three instructions regarding the
evidentiary standard required to overcome the presumption of validity: (1) clear and
convincing evidence; (2) clear and convincing evidence, with an additional instruction of
the type articulated in i4i stating that the burden of proof is “more easily satisfied when,
as in this case, the prior art on which the claim of [invalidity] is based was not considered
by the Examiner”; and (3) preponderance of the evidence. The study found a statistically
significant difference between subjects given the first (clear and convincing) instruction
(27.1 percent invalid) and the third (preponderance) instruction (38.3 percent invalid).
In addition, the study found no statistically significant difference between subjects given
the second (clear and convincing with additional i4i-type language) instruction (43.6
percent invalid) and the third (preponderance) instruction. The authors contended that
these results “suggest that if the Court in i4i had switched from a clear and convincing
standard . . . to a preponderance standard . . . it may have resulted, certeris paribus, in
more patents being found invalid by juries.” In addition, the authors asserted that in
light of the fact that “subjects found patents invalid at rates indistinguishable from the
preponderance standard when . . . an i4i-type instruction was given,” “the importance of
[such] instructions may have been underappreciated by both courts and parties, as such
an instruction may effectively reduce the standard of proof, even if it formally remains
clear and convincing evidence.”19
A second experiment by Bock (2015) sought to test the impact of instructing the jury on
the presumption of validity, but without varying the standard of proof. In the experiment,
the subjects (N = 1745) were given a fact pattern adapted from Schwartz and Seaman
(2013).20 Some subjects were then randomly assigned an instruction regarding the pre-
sumption of validity, while others did not receive this instruction.21 All subjects, however,
were instructed that invalidity had to be proven by clear and convincing evidence. The
technology and much of the surrounding body of legal rules . . . These factors could impact this
investigation in ways the authors could neither determine nor account for.
17
The authors used Amazon Mechanical Turk for this purpose. See Amazon Mechanical
Turk, www.mturk.com.
18
The fact pattern for this experiment was based on Callaway Golf Co. v. Acushnet Co., 576
F.3d 1331 (Fed. Cir. 2009).
19
In a second experiment reported in the same paper, Schwartz and Seaman (2013) found
that different wordings of an i4i-type instruction regarding unconsidered prior art did not result in
statistically significant differences.
20
One difference in the fact pattern provided in Bock (2015) is that the USPTO previously
considered all prior art raised in the subsequent invalidity challenge in court.
21
In an additional condition, some subjects were instructed regarding resource limitations on
the USPTO’s ability to fully examine patents and the problem of poor patent quality. Bock (2015)

study found a statistically significant difference between the control (no instruction on
presumption of validity) condition (31.7 percent invalid) and treatment (instruction on
presumption of validity) condition (24.7 percent invalid). The author concluded that the
“results of the experiment . . . largely confirm the conventional view of litigants that the
presumption instruction may have a substantial impact on the jury’s decision on invalidity
issues.”
These experimental studies also have several potential limitations, as their authors
acknowledge. The most significant limitation is ecological validity—in other words,
whether the experiment approximates real-life conditions enough that the results are gen-
eralizable (Breau and Brook, 2007). In both experiments, subjects were given a short (one
to two pages) written fact pattern, a jury instruction or two, and then were asked several
questions regarding their decision on invalidity. Real-life patent trials are significantly
longer and more complicated, often involving multiple complex issues (e.g., infringe-
ment and remedies) in addition to invalidity. In addition, jurors in patent cases hear live
testimony from witnesses and arguments from the parties’ attorneys over an extended
period of time and receive instructions from a judge, rather than reviewing a written fact
pattern. Finally, jurors in real-life patent cases are required to deliberate as a group and
reach unanimity regarding their decision, a feature which is absent from both experiments.
V. IMPLICATIONS AND AREAS FOR FUTURE RESEARCH
The existing empirical scholarship suggests that the presumption of validity and the
standard of proof to overcome it matter in patent litigation, at least in close cases
(Schwartz and Seaman, 2013; Bock, 2015). In addition, invalidity challengers which
offer “new” prior art—that is, evidence not previously considered by the USPTO during
prosecution of the patent—combined with a jury instruction of the type authorized by
i4i may be more likely to prevail on invalidity than those which do not (Schwartz and
Seaman, 2013). These findings have several significant implications for participants in
patent litigation and the patent system.
One issue that litigants may face in an invalidity challenge is whether any alleged
“new” prior art is merely duplicative or cumulative of the information considered by the
examiner during prosecution. As the Federal Circuit has explained, “[t]he issue of . . .
cumulativeness of non[-] considered prior art relative to considered art often arises in the
context of validity challenges.” AbbVie Deutschland GmbH & Co., KG v. Janssen Biotech,
Inc., 759 F.3d 1285 (Fed. Cir. 2014). After i4i, courts have placed the burden on the party
asserting invalidity to demonstrate that the alleged “new” prior art is more relevant than
the previously considered prior art. If this burden is not met, then the trial court may
decline to grant an i4i-type jury instruction (AbbVie, 2014).
In addition, the particular wording of i4i-type jury instructions regarding the impact
of “new” prior art may be relevant. Several model jury instructions on this issue have
found that this instruction resulted in a statistically significant difference in outcomes (38.5 percent
invalid) compared to subjects who did not receive such an instruction (31.7 percent invalid).

been promulgated, including by the Northern District of California22 and the American
Intellectual Property Law Association.23 Courts have articulated alternative instructions
as well.24 These instructions could be experimentally tested to determine whether particu-
lar wording has an impact on invalidity decisions.
Third, to avoid the potential adverse impact of an i4i-type instruction, patent applicants
may have an incentive to more extensively search for, disclose, and explain the importance
of relevant prior art during prosecution. If the applicant successfully overcomes these
references, the patentee will have a stronger prosecution record to rely on when the patent
is subsequently asserted in litigation (Alsup, 2009).
Fourth, the absence of a heightened presumption of validity in the various forms of post-
issuance administrative proceedings at the USPTO, such as ex parte re-examination25 and
22
See Model Patent Jury Instructions for the Northern District of California (2015) (“During
this case, the [alleged infringer] has submitted prior art that was not considered by the United
States Patent and Trademark Office (PTO) during the prosecution of the [ ] patent. The [alleged
infringer] contends that such prior art invalidates certain claims of the [ ] patent. In deciding the
issue of invalidity, you may take into account the fact that the prior art was not considered by the
PTO when it issued the [ ] patent. Prior art that differs from the prior art considered by the PTO
may carry more weight than the prior art that was considered and may make the [alleged infringer’s]
burden of showing that it is highly probable that a patent claim is invalid easier to sustain”).
23
See American Intellectual Property Law Ass’n (2014) (“Regardless of whether the prior art
was considered by the USPTO Examiner during the prosecution of the application which resulted
in the issued patent, [the Defendant] must prove that it is highly probable that each asserted claim
is invalid. When, however, a party challenging the validity of a patent relies on prior art that was
considered by the Examiner, that party’s ability to satisfy its highly probable evidence burden may
be more difficult. On the other hand, when a party challenging the validity of a patent presents
evidence that was not considered by the Examiner, such new evidence may be given more weight
and may make it easier to satisfy that party’s highly probable evidence burden”). However, the
most recent version of the AIPLA’s model patent jury instructions has omitted language suggesting
that the burden of establishing invalidity may be more easily met with unconsidered prior art. See
American Intellectual Property Law Ass’n (2017) (“Regardless whether or not a particular prior
art reference[s] [[was] [were]] considered by the Patent Examiner during the prosecution of the
application which matured into the [abbreviated patent number] patent, [the Defendant] must prove
that it is highly probable that the challenged claim[s] [[is] [are]] invalid. This burden of proof on [the
Defendant] never changes regardless whether or not the Patent Examiner considered the reference”).
24
For instance, the Federal Circuit found no reversible error in AbbVie (2014), when the
district court gave the following instruction:
You should consider whether that additional information would have been “material” to the PTO’s
decision to grant the patents. Information is “material” if there is a substantial likelihood that a rea-
sonable patent examiner would consider it important in deciding whether to allow the application
to issue as a patent . . . If the PTO did not have all the material information before it when it made
its decision as to a particular claim, [the challenger]’s burden may be easier to meet. That is particu-
larly true if the additional information was not only material, but would have carried significant
weight had it been considered by the PTO. But if the additional information was not material, or it
would not have carried significant weight, [the challenger]’s burden may be more difficult to meet.
Id. at 1294.
25
See In re Swanson, 540 F.3d 1368 (Fed. Cir. 2008) (“In PTO examinations and reexamina-
tions, the standard of proof—a preponderance of evidence—is substantially lower than in a civil
case [and] there is no presumption of validity”).

inter partes review,26 may cause an infringement defendant to prefer to contest invalidity in
these proceedings, rather than in court where the clear and convincing evidence standard
still applies (Klimczak, 2012). This was particularly true in light of the Patent Trial and
Appeal Board’s former rule that a claim subject to inter partes review generally “shall be
given its broadest reasonable construction in light of the specification,” which may sweep
in more prior art than a (potentially) narrower claim construction by a district court.27
Indeed, this practice already appears to be common; a recent empirical study found that
over 75 percent of patents in inter partes review were also involved in co-pending litigation
(Love and Ambwani, 2014).
A final significant but unresolved issue is whether the presumption of validity applies
in challenges to patent subject-matter eligibility. This is a particularly salient issue in the
wake of Alice Corp. v. CLS Bank International, 134 S. Ct. 2357 (2014), which led to a wave
of decisions invalidating issued patents under § 101 (Tran, 2015; Gugliuzza and Lemley,
2017). Lower courts as well as some Federal Circuit judges have split on this issue.28 In
the author’s opinion, however, the presumption of validity should be inapplicable to chal-
lenges regarding patent eligibility under § 101. It is well established that patent eligibility
is a question of law.29 In contrast, as explained in Justice Breyer’s concurring opinion
in i4i, the clear-and-convincing evidence standard “applies to questions of fact and not
to questions of law” (Microsoft Corp. v. i4i Limited Partnership, 2011). As a result, this
heightened standard of proof should be inapplicable to claims of invalidity under § 101.
VI. CONCLUSION
In sum, the empirical literature suggests that the statutory presumption of validity and
the clear-and-convincing evidence standard to overcome it, as required by the Supreme
26
See 35 U.S.C. § 316(e) (“In an inter partes review instituted under this chapter, the petitioner
shall have the burden of proving a proposition of unpatentability by a preponderance of the evi-
dence.”); see also Cooper v. Lee, 86 F. Supp. 3d 480 (E.D. Va. 2015) (“Both ex parte reexamination
and inter partes review do not give an issued patent a presumption of validity . . .”).
27
The USPTO modified this rule in 2018, replacing the broadest reasonable interpretation
standard in inter partes review proceedings with the same claim construction standard that is used
to construe the claim in a civil action in a federal district court (USPTO 2018b).
28
Compare Ultramercial, 772 F.3d 709 (Fed. Cir. 2014) (Mayer, J., contending that the pre-
sumption of validity does not apply to challenges to patent-eligible subject matter under § 101),
and OpenTV, Inc. v. Apple, Inc., 2015 WL 1535328 (W.D. Tex. 2015) (holding that “[t]he issue of
invalidity under Section 101 presents a question of law” and that “no . . . presumption of eligibility
applies on the section 101 calculus (internal quotations omitted)), with BASCOM Global Internet
Servs., Inc. v. AT&T Mobility LLC, 107 F. Supp. 3d 639 (N.D. Tex. 2015) (“An issued patent is
entitled to a presumption of validity, and a party arguing that a patent claims ineligible subject
matter under § 101 must prove as much by clear and convincing evidence”), and Exergen Corp. v.
Brooklands Inc., 125 F. Supp. 2d 307 (D. Mass. 2015) (finding that the “heightened standard” of
proof applies to § 101 challenges).
29
See, for example, OIP Techs., Inc. v. Amazon.com, Inc., 788 F.3d 1359 (Fed. Cir. 2015)
(“Patent eligibility under 35 U.S.C. § 101 is an issue of law reviewed de novo”); CyberSource Corp.
v. Retail Decisions, Inc., 654 F.3d 1366 (Fed. Cir. 2011) (“Issues of patent-eligible subject matter are
questions of law and reviewed without deference”).

Court in i4i, have had a meaningful impact on invalidity decisions in patent litigation, at
least in marginal cases. In addition, invalidity challenges based on prior art not previously
considered by the USPTO are more likely to be successful.
REFERENCES
Allison, John R., Mark A. Lemley and David L. Schwartz. 2014. “Understanding the Realities of Modern
Alsup, William. 2009. “A District Judge’s Proposal for Patent Reform: Revisiting the Clear and Convincing
Standard and Calibrating Deference to the Strength of the Examination”, 24 Berkeley Technology Law
Journal 1647–55.
Amazon Mechanical Turk, www.mturk.com.
American Intellectual Property Law Ass’n, AIPLA Model Patent Jury Instructions (2014), www.aipla.org/
committees/committee_pages/Patent-Litigation/Committee%20Documents/2014%20Model%20Jury%20I
nstructions%20Draft.docx.
American Intellectual Property Law Ass’n, AIPLA Model Patent Jury Instructions (2017), www.aipla.org/
learningcenter/library/books/other-pubs/Documents/2017-3-15%20AIPLA%20Model%20Patent%20Jury%20
Instructions.docx.
Ayres, Ian and Paul Klemperer. 1999. “Limiting Patentees’ Market Power Without Reducing Innovation Incentives:
The Perverse Benefits of Uncertainty and Non-Injunctive Remedies”, 97 Michigan Law Review 985–1033.
Bock, Jeremy W. 2015. “Does the Presumption of Validity Matter? An Experimental Assessment”, 49 University
of Richmond Law Review 417–63.
Breau, David L. and Brian Brook. 2007. “‘Mock’ Mock Juries: A Field Experiment on the Ecological Validity
of Jury Simulations”, 31 Law and Psychology Review 77–92.
Brief for Petitioner Microsoft Corp. 2011. Microsoft Corp. v. i4i Limited Partnership, No. 10-290. Available at
2011 WL 288890.
Chatlynne, Etan S. 2009. “The Burden of Establishing Patent Invalidity: Maintaining a Heightened Evidentiary
Standard Despite Increasing ‘Verbal Variances’”, 31 Cardozo Law Review 297–326.
Chatlynne, Etan S. 2010. “Investigating Patent Law’s Presumption of Validity—An Empirical Analysis”, 2010
Patently-O Patent Law Journal 37–55.
Chatlynne, Etan S., Stephen Kenny and Lucas Watkins. 2011. “Investigating Patent Law’s Presumption of
Validity, Part II: An Empirical Analysis of How Unconsidered Evidence and Evidentiary Standards Affect
Jury Verdicts”, 2011 Cardozo Law Review de novo 46–63.
Clermont, Kevin M. and Theodore Eisenberg. 1992. “Trial by Jury or Judge: Transcending Empiricism”, 77
Cornell Law Review 1124–74.
Cohen, Jacob. 1992. “Statistical Power Analysis”, 1 Current Directions in Psychological Science 98–101.
Croson, Rachel. 2002. “Why and How to Experiment: Methodologies from Experimental Economics”, 2002
University of Illinois Law Review 921–45.
Daniel, B.D. 2008. “Heightened Standards of Proof in Patent Infringement Litigation: A Critique”, 36 AIPLA
Quarterly Journal 369–417.
Devlin, Alan. 2008. “Revisiting the Presumption of Patent Validity”, 37 Southwestern University Law Review
323–69.
DiMasi, Joseph A. et al. 2003. “The Price of Innovation: New Estimates of Drug Development Costs”, 22
Journal of Health Economics 151–81.
Dunner, Donald R., J. Michael Jakes and Jeffrey D. Karceski. 1995. “A Statistical Look at the Federal Circuit’s
Federal Trade Commission. 2003. To Promote Innovation: The Proper Balance of Competition and Patent
Law and Policy. Available at www.ftc.gov/sites/default/files/documents/reports/promote-innovation-proper-
balance-competition-and-patent-law-and-policy/innovationrpt.pdf.
Federico, P.J. 1993. “Commentary on the New Patent Act”, 75 Journal of the Patent & Trademark Office Society
161–231.
Ford, Roger A. 2013. “Patent Invalidity Versus Noninfringement”, 99 Cornell Law Review 71–128.
Frakes, Michael D. and Melissa F. Wasserman. 2017. “Is the Time Allocated to Review Patent Applications
Inducing Examiners to Grant Invalid Patents?: Evidence from Microlevel Application Data”, 99 Review of
Economics and Statistics 550–63.

Gugliuzza, Paul R. and Mark A. Lemley. 2017. “Can the Court Change the Law by Saying Nothing?”, Available
at https://papers.ssrn.com/abstract_id=3015459.
Janis, Mark D. 2004. “Reforming Patent Validity Litigation: The ‘Dubious Preponderance’”, 19 Berkeley
Kesan, Jay P. 2002. “Carrots and Sticks to Create a Better Patent System”, 17 Berkeley Technology Law Journal
763–97.
Kesan, Jay P. and Gwendolyn G. Ball. 2006. “How Are Patent Cases Resolved? An Empirical Examination of
the Adjudication and Settlement of Patent Disputes”, 84 Washington University Law Review 237–312.
Kesan, Jay P. and Mark Banik. 2000. “Patents as Incomplete Contracts: Aligning R&D Investment with
Incentives to Disclose Prior Art”, 2 Washington University Journal of Law and Policy 23–54.
Kieff, Scott. 2001. “Property Rights and Property Rules for Commercializing Inventions”, 85 Minnesota Law
Review 697–754.
Klimczak, Ryan. 2012. “i4i and the Presumption of Validity: Limited Concerns over the Insulation of Weak
Patents”, 27 Berkeley Technology Law Journal 299–323.
Lemley, Mark A. 2001. “Rational Ignorance at the Patent Office”, 95 Northwestern University Law Review
1495–532.
Lemley, Mark A. and Carl Shapiro. 2005. “Probabilistic Patents”, 19 Journal of Economic Perspectives 75–98.
Leslie, Christopher R. 2006. “The Anticompetitive Effects of Unenforced Invalid Patents”, 91 Minnesota Law
Review 101–83.
Leslie, Christopher R. 2011. “Antitrust, Inequitable Conduct, and Intent to Deceive the Patent Office”, 1 UC
Irvine Law Review 323–56.
Lichtman, Doug and Mark A. Lemley. 2007. “Rethinking Patent Law’s Presumption of Validity”, 60 Stanford
Law Review 45–72.
Love, Brian J. and Shawn Ambwani. 2014. “Inter Partes Review: An Early Look at the Numbers”, 81 University
of Chicago Law Review Dialogue 93–107.
Model Patent Jury Instructions for the Northern District of California. 2015. Available at www.cand.uscourts.
gov/filelibrary/5/Model_Patent_Jury_Instructions_7.16.15.pdf.
Moore, Kimberly A. 2000. “Judges, Juries, and Patent Cases—An Empirical Peek Inside the Black Box”, 99
Morrissett, John A. 2012. “I 4 An i: Why Changing the Standard for Overcoming the Presumption of Patent
Validity Will Cause More Harm Than Good”, 18 Richmond Journal of Law and Technology 7.
Phipps, Charles E. 2000. “The Presumption of Administrative Correctness: The Proper Basis for the Clear and
Convincing Evidence Standard”, 10 Federal Circuit Bar Journal 143–62.
Priest, George L. and Benjamin Klein. 1984. “The Selection of Disputes for Litigation”, 13 Journal of Legal
Studies 1–55.
Schwartz, David L. and Christopher B. Seaman. 2013. “Standards of Proof in Civil Litigation: An Experiment
from Patent Law”, 26 Harvard Journal of Law and Technology 429–80.
Seymore, Sean B. 2013. “The Presumption of Patentability”, 97 Minnesota Law Review 990–1046.
Seymore, Sean B. 2016. “Patent Asymmetries”, 49 UC Davis Law Review 963–1015.
Sumida, Matthew K.K. 2011. “Defendant Class Actions and Patent Infringement Litigation”, 58 UCLA Law
Review 843–87.
Taylor, David O. 2011. “Clear but Unconvincing: The Federal Circuit’s Invalidity Standard”, 21 Fordham
Intellectual Property, Media & Entertainment Law Journal 293–363.
Tran, Jasper L. 2015. “Software Patents: A One-Year Review of Alice v. CLS Bank”, 97 Journal of the Patent
and Trademark Office Society 532–50.
U.S. Patent and Trademark Office. 2018a. FY 2018 Performance and Accountability Report, 179, www.uspto.
gov/sites/default/files/documents/USPTOFY18PAR.pdf.
U.S. Patent and Trademark Office. 2018b. Changes to the Claim Construction Standard for Interpreting Claims
in Trial Proceedings Before the Patent Trial and Appeal Board, 83 Federal Register 51340-59.
U.S. Patent and Trademark Office. 2019. Data Visualization Center, www.uspto.gov/corda/dashboards/patents/
main.dashxml?CTNAVID=1004.
Cases
AbbVie Deutschland GmbH & Co. v. Janssen Biotech, Inc. (2014), 759 F.3d 1285.
Addington v. Texas (1979), 441 U.S. 418.
Agawam Woolen Co. v. Jordan (1868), 74 U.S. 583.
Alice Corp. v. CLS Bank Int’l (2014), 134 S. Ct. 2357.
Am. Hoist & Derrick Co. v. Sowa & Sons, Inc. (1984), 725 F.2d 1350.
Aventis Pharma Deutschland GmbH v. Lupin Ltd. (2006), 2006 WL 2008962.

BASCOM Global Internet Servs., Inc. v. AT&T Mobility LLC (2015), 107 F. Supp. 3d 639.
Butler Mfg. Co. v. Enterprise Cleaning Co. (1936), 81 F.2d 711.
Callaway Golf Co. v. Acushnet Co. (2009), 576 F.3d 1331.
Cantrell v. Wallick (1886), 117 U.S. 689.
Carson Mfg. Co. v. Carsonite Int’l Corp. (1981), 686 F.2d 665.
Chicago Rawhide Mfg. Co. v. Crane Packing Co. (1975), 523 F.2d 452.
Coffin v. Ogden (1873), 85 U.S. 120.
Commil USA, LLC v. Cisco Systems, Inc. (2015), 135 S. Ct. 1920.
Connell v. Sears, Roebuck & Co. (1983), 722 F.2d 1542.
Cooper v. Lee (2015), 86 F. Supp. 3d 480.
CyberSource Corp. v. Retail Decisions, Inc. (2011), 654 F.3d 1366.
Dickstein v. Seventy Corp. (1975), 522 F.2d 1294.
Exergen Corp. v. Brooklands Inc. (2015), 125 F. Supp. 2d 307.
Gillette Safety Razor Co. v. Cliff Weil Cigar Co. (1939), 107 F.2d 105.
Greenwood v. Hattori Seiko Co. (1990), 900 F.2d 238.
H. Schindler & Co. v. C. Saladino & Sons (1936), 81 F.2d 649.
i4i Limited Partnership v. Microsoft Corp. (2009), 670 F. Supp. 2d 568.
i4i Limited Partnership v. Microsoft Corp. (2010), 598 F.3d 831.
In re Swanson (2008), 540 F.3d 1368.
Jacuzzi Bros., Inc. v. Berkeley Pump Co. (1951), 191 F.2d 632.
Kahn v. Gen. Motors Corp. (1998), 135 F.3d 1472.
KSR Int’l Co. v. Teleflex, Inc. (2007), 550 U.S. 398.
Magnivision, Inc. v. Bonneau Co. (1997), 115 F.3d 956.
Microsoft Corp. v. i4i Limited P’ship (2010), 562 U.S. 1060.
Microsoft Corp. v. i4i Limited P’ship (2011), 564 U.S. 91.
Munising Paper Co. v. Am. Sulphite Pulp Co. (1915), 228 F. 700.
OIP Techs., Inc. v. Amazon.com, Inc. (2015), 788 F.3d 1359.
OpenTV, Inc. v. Apple, Inc. (2015), 2015 WL 1535328.
Potthoff v. Hanson & Van Winkle Co. (1909), 174 F. 983.
Radio Corp. of Am. v. Radio Eng’g Labs, Inc. (1934), 293 U.S. 1.
Rains v. Niaqua, Inc. (1969), 406 F.2d 275.
Santosky v. Kramer (1982), 455 U.S. 745.
Tenneco Chems., Inc. v. William T. Burnett & Co. (1982), 691 F.2d 658.
Trans-World Mfg. Corp. v. Al Nyman & Sons, Inc. (1984), 750 F.2d 1552.
Ultramercial, Inc. v. Hulu, LLC (2014), 772 F.3d 709.
Ultra-Tex Surfaces, Inc. v. Hill Bros. Chem. Co. (2000), 204 F.3d 1360.
35 U.S.C. § 69 (1946).
35 U.S.C. § 122(a) (2012).
35 U.S.C. § 282 (2012).
35 U.S.C. § 316(e) (2012).
H.R. Rep. No. 82-1923 (1952).
Patent Act, § 6 (1793).
Patent Act, § 15 (1836).
Patent Act, § 61 (1870).
S. Rep. No. 82-1979 (1952).

16. Inequitable conduct and patent misuse
Lee Petherbridge* and Jason Rantanen**
30
Contents
I. Introduction
II. Inequitable Conduct
A. Doctrinal Architecture
1. Missing doctrinal components
2. Different doctrinal standards
3. Intent to deceive
4. Materiality
B. Theory and Purpose
C. The Prevalence of Inequitable Conduct Claims
1. The frequency of written opinions addressing inequitable conduct
2. The frequency of inequitable conduct claims in pleadings
D. Outcomes of Inequitable Conduct Claims
1. District courts
2. The Federal Circuit
E. Multi-layer Studies
F. Concluding Thoughts: Inequitable Conduct
III. Patent Misuse
A. The Number of Misuse Opinions and Outcomes
B. Forms of Misuse
C. What Explains the Observed Decline in Patent Misuse Challenges?
D. Concluding Thoughts: Patent Misuse
References
I. INTRODUCTION
Most conceptions of the patent law envision a system of rules that balances private
rights against public interests such as promoting innovation, removing impediments to
competition, and making new and useful information broadly available. The patent law
doctrines of inequitable conduct and patent misuse play a role in creating and enforcing
this balance. The former bars the enforcement of patents that were obtained through
deceptive conduct during patent prosecution, while the latter bars the enforcement of
** Professor of Law, Vachon Research Fellow, Loyola Law School, Loyola Marymount University.
** Professor, Ferguson-Carlson Fellow in Law, and Director of the Innovation, Business, and
Law Program, University of Iowa Law School.
367

patents when a patentee exploits them in one of a number ways that courts have declared
to be socially harmful.
Here, we review recent empirical insights into both patent doctrines, starting with
inequitable conduct and then moving to patent misuse.
II. INEQUITABLE CONDUCT
The foundation of the inequitable conduct doctrine is the historic equitable doctrine of
unclean hands, which requires that “he who comes into equity must come with clean
hands.” Precision Instrument Mfg. Co. v. Auto. Maint. Mach. Co., 324 U.S. 806 (1945).
Patentees come in equity when they seek the traditional remedy for patent infringement,
injunctive relief, and a patentee’s hands may be unclean if a patentee has, to its advantage,
deceived the public in the process of securing a patent. Broadly speaking, the doctrinal
rubric asks whether a patent applicant either failed to disclose information material to
patentability or made misrepresentations to the Patent Office that were material to patent-
ability, and if so, whether the non-disclosures or misrepresentations were made with an
intent to deceive the Patent Office into allowing patent claims to issue. Therasense, Inc.
v. Becton Dickinson & Co., 649 F.3d 1276 Fed. Cir. 2011) (en banc). If materiality and
intent to deceive are established, a trial judge is supposed to determine, on balance of the
evidence and other equities that may be involved, whether the patents should be declared
unenforceable (Therasense, 2011).
The doctrine of inequitable conduct has never been comprehensively theorized, but
in policing the relationship between a patent applicant and the Patent Office, it seeks to
serve at least the following purposes: encouraging efficient administration of the patent
granting process; encouraging fairness in the application of the patent laws between
competitors; encouraging patent quality; adjusting patent value to account for inefficien-
cies in administrative process and systemic inefficiencies flowing from applicant strategic
behavior; protecting the reputation of the patent bar; and protecting the reputation of
the patent system (Petherbridge et. al., 2011; Cotter, 2011; Rantanen, 2012; Petherbridge,
2013; Chiang, 2013; Duffy, 2013).
Although inequitable conduct is litigated with some frequency and is commonly
discussed by practitioners and academics, the doctrine itself has not been the subject of
much empirical research. That said, empirical research into inequitable conduct has made
progress over the last 25 years, emphasizing, so far, three sources of data: Federal Circuit
opinions, district court orders, and party pleadings.
Because Federal Circuit opinions mainly express the review of legal challenges to a
district court’s application of the law of inequitable conduct, they likely present the
optimal dataset for examining the articulated content of the inequitable conduct doctrine
itself. Nevertheless, like any choice of data source, the use of Federal Circuit opinions
entails limitations. District court orders have not yet been much used to study inequitable
conduct, but what little research is available suggests that more investigation is warranted.
However, these, too, come with their own set of limitations, some of which are mentioned
later in this chapter. Pleadings have been more often examined to study inequitable
conduct, but have presented researchers with substantial pitfalls. These issues are explored
in more detail below.

Inequitable conduct and patent misuse 369
A uniform limitation to interpreting the meaning of the studies reviewed here is the
risk of selection bias. Selection bias is a form of statistical error that can occur when data
selected for analysis is for some reason not representative of the population a researcher
seeks to analyze. It can sometimes impact conclusions that might be drawn from collected
data, but not always; and as the definition implies, its effects depend on the level of analy-
sis and nature of the claim a researcher makes based on the data collected. In legal studies
the problem is nearly ubiquitous—the basic concern being that in using written orders
and opinions as data, there may be facets of the law and patterns of decisions incorrectly
quantified because they are for some reason less likely to appear in an analyzable writing.
As readers consider the studies reviewed here, including those exploring patent misuse,
they are cautioned to keep this issue in mind.
From the review of existing studies it seems clear that in order to develop the most
complete understanding of the inequitable conduct doctrine, future work will need not
only to examine Federal Circuit opinions, but also to draw more than current work does
from district court decisions, patents and prosecution histories, and, probably, from other
features of cases that do not find expression in written opinions.
Nevertheless, as noted above, the last 25 years have seen significant advances in the
understanding of inequitable conduct, with the most significant coming quite recently.
Content analysis has offered some insights into the doctrinal architecture of inequitable
conduct, and patents have been used to evaluate whether the inequitable conduct doctrine
might serve its theoretical purposes. In addition, much has been learned about how fre-
quently parties to patent cases raise the issue of inequitable conduct, as well as how courts
address the issue when it is presented to them. We review below the studies providing new
information about these topics.
A. Doctrinal Architecture
While legal scholars have described the historical development of inequitable conduct
(McGowan, 2010), and offered arguments about the doctrinal significance of notable
cases (Rantanen, 2012) or sets of cases (Dolak, 2010), it is only very recently that research
has begun to paint a general picture of the architecture of the inequitable conduct doc-
trine itself. That nascent picture of the formal legal structure of inequitable conduct comes
from a single comprehensive content analysis of all Federal Circuit inequitable conduct
decisions, from the creation of the court through May 27, 2010 (Petherbridge et al., 2011).
1. Missing doctrinal components

Petherbridge et al. confirm the presence of widely recognized aspects of the legal frame-
work for analyzing inequitable conduct—in particular, that materiality and intent to
deceive are significant doctrinal features—but also reveal, surprisingly, that there is little
to no law analyzing the equitable weighing that traditional doctrine holds is supposed to
follow the factual assessments of materiality and intent.
The Federal Circuit hardly ever talks about equitable balancing and virtually never
concludes that lower tribunals misbalanced the equities. Across the 361 inequitable
conduct analyses the court performed between 1983–2010, only 8.33 percent evince any
review of the required equitable balancing. The court never reversed a lower tribunal’s
finding that inequitable conduct did not occur for failure to properly balance; and out of

151 analyses reviewing lower tribunal judgments that inequitable conduct had occurred,
only two (1.32 percent) were vacated due to the tribunal’s failure to properly balance. In
both cases, the perceived flaw was not that the lower tribunal misapplied the balancing
as much as that the lower tribunal had not done the balancing at all. It thus appears that
while the inequitable conduct doctrine may demand that a lower tribunal express that it
has performed the required balancing, there is little in the doctrine or jurisprudence to
guide a lower tribunal’s exercise of that discretion.
This remarkable observation suggests a couple of avenues for future study. The first,
and perhaps most obvious, is whether legal scholars might improve on what the courts
have accomplished. It is difficult to believe that the many hundreds of inequitable conduct
cases have not given rise to a body of equitable doctrine that articulates a connection
between the doctrine and the equities implicated by the policies it serves. So perhaps
a more traditional, case-interpretive approach might begin to show that the equitable
discretion authorized by the inequitable conduct doctrine is something more than a trial
judge’s whim. A second avenue, and one that complements the first, is whether something
resembling equitable balancing might be found, unarticulated, in the analyses the Federal
Circuit uses to regularly attack lower tribunals’ factual findings of materiality and intent
to deceive. In either event, if they exist, bringing the doctrinal features of inequitable
conduct’s equitable balancing into light might lead to a more useful and efficient doctrine;
and if no doctrinal features can be resolved, then scholars and jurists can confront the
question of whether such features should be developed and what they should look like.
Another traditionally mentioned feature of the doctrine that may also play a much
smaller role than imagined is cumulativeness. Traditional statements of the inequitable
conduct doctrine hold that information cannot be material to patentability—and thus
be capable of sustaining the finding of materiality that is prerequisite to a determination
of inequitable conduct—if allegedly material information is cumulative to information
already of record in a case. Larson Mfg. Co. of South Dakota, Inc., 559 F.3d 1317 (Fed
Cir. 2009).
This substantive bar to finding inequitable conduct is, however, discussed in only about
15 percent of Federal Circuit analyses of inequitable conduct, and is argued successfully
by a patentee in only 5.8 percent (Petherbridge et. al., 2011). The observation that it is so
rarely discussed, and even less often helpful to a patentee, has several possible explana-
tions that may be worthy of future study. Perhaps, for example, the law of cumulativeness
is so very clear that the issue is rarely disputed and for that reason only rarely appears
in Federal Circuit opinions; alternatively, it may be the case that for some reason the
law around cumulativeness is either too costly to apply, or, perhaps more likely, usually
irrelevant to most claims of inequitable conduct. In any event, the observation throws
the significance of cumulativeness as a rule prohibiting inequitable conduct findings into
serious question.
2. Different doctrinal standards

A second feature of the doctrinal architecture suggested by Petherbridge et al. is that
different layers of the judicial system appear to require different degrees of impropriety
to reach a finding of inequitable conduct. In particular, the research suggests that the
Federal Circuit applies a standard for finding inequitable conduct that is stricter than that
applied by the tribunals it reviews.

Although the inequitable conduct doctrine is fact based and (supposedly) dependent
on a district judge’s exercise of equitable discretion, a decisional structure that should
uniformly lead to a high degree of appellate deference to lower tribunal decision making,
Federal Circuit doctrine treats lower tribunal inequitable conduct judgments very differ-
ently. When the appealed judgment is that no inequitable conduct occurred, the Federal
Circuit affirms approximately 91 percent of the time. But when the appealed judgment is
that inequitable conduct did occur, the Federal Circuit affirms only about 46 percent of
the time. These differences are very strongly significant. Cast in terms of reversal rates, the
Federal Circuit reverses roughly 54 percent of judgments of inequitable conduct, while
reversing only roughly 9 percent of judgments of no inequitable conduct.
The same pattern of differential treatment is reported by other studies (Dunner, 1995;
1
Mammen, 2009), and suggests that the Federal Circuit’s judgment-dependent treatment
of lower tribunal inequitable conduct judgments—hostile toward one outcome and
friendly toward another—has been a regular feature of modern patent law.
Some explanation for this differential treatment may be found in Federal Circuit judges
holding different understandings of the doctrine. For example, patentees seem to be
extraordinarily successful (as high as 100 percent) on appeals involving inequitable con-
duct when certain judges author opinions, and much less successful (as low as 40 percent)
when other judges author (Petherbridge et. al., 2011). Moreover, when only cases vacating
or reversing lower tribunal findings of inequitable conduct are examined, patentees are
again sometimes extraordinarily successful (i.e., 100 percent) and sometimes much less
so (i.e., 20 percent) in opinions authored by different judges (Petherbridge et. al., 2011).
Taken together these studies suggest that, while Federal Circuit inequitable conduct
doctrine includes a standard for finding inequitable conduct that is stricter than that
applied by at least some of the tribunals it reviews, that stricter standard may not be the
only one present in Federal Circuit law. Although Federal Circuit judges are bound by
the law imposed by the decisions of prior panels, that law may not constrain judges to
the same standard of decision. Some judges, it seems, think the standard for inequitable
45.70 % Affirmed
54.30 % Reversed or Vacated
Figure 16.1 Frequency of incoming judgments of inequitable conduct affirmed and not

affirmed (1983–2010) (n = 151)1
1
From Petherbridge et al., (2011, Fig. 3). Figure 16.1 shows the percentage of Federal Circuit
analyses in which lower tribunals’ judgments that patentees have not engaged in inequitable con-
duct are affirmed (cf. Mammen, 2009, p. 1332).

9.27
% Affirmed
% Reversed or Vacated
90.73
Figure 16.2 Frequency of incoming judgments of no inequitable conduct affirmed and

not affirmed (1983–2010) (n = 205)2
conduct is quite high (in one case, even impossible), while other judges are clearly more
2
willing to accept findings of inequitable conduct and might be largely content with (1)
ensuring that the district court applied the proper process to reach its decision and (2)
applying to it the relevant clear error and abuse of discretion forms of review.
The 54 percent reversal rate of findings of inequitable conduct, noted above, might, in
view of all of this, simply reflect some lower tribunals applying one of the proper stand-
ards of decision, but getting reviewed by a panel of Federal Circuit judges that prefers a
higher, and also proper in view of the cases, standard.
3. Intent to deceive

The Federal Circuit inequitable conduct doctrine appears to mainly express its stricter
standard through the doctrinal element of intent to deceive, the main evidence for this
being that patentees are much more likely to win for the reason that the lower tribunal
erred in finding intent to deceive than for any other reason.
In 41.33 percent of Federal Circuit inequitable conduct analyses, the sole reason given
for patentee success was that the challenger had not established intent to deceive. By
contrast, significantly fewer analyses—only 16.97 percent—gave failure to establish mate-
riality as the sole reason for patentee success, while 23.62 percent of analyses indicated
that challengers had failed to establish intent to deceive and materiality (Petherbridge et.
al., 2011).
Furthermore, in the 82 inequitable conduct analyses authored by the Federal Circuit
that reversed or vacated a trial judge’s determination that inequitable conduct had
occurred, 50 percent were rejected solely because of error in the lower tribunal’s finding of
intent to deceive. A significantly smaller number—18.29 percent—were rejected solely due
to error in the finding of materiality. When analyses that found error in both materiality
and intent to deceive are included (overall, 23.17 percent), nearly three-quarters (73.17
percent) of all rejected lower tribunal judgments of inequitable conduct involved error in
the lower tribunal’s finding of intent to deceive (Petherbridge et. al., 2011).
2
From Petherbridge et al. (2011, Fig. 4). Figure 16.2 shows the percentage of Federal Circuit
analyses in which lower tribunals’ judgments that patentees have not engaged in inequitable con-
duct are affirmed (cf. Mammen, 2009, p. 1332).

Similarly, a study examining a comparatively smaller period surrounding the Federal

Circuit’s notable decision in Therasense v. Becton Dickinson observes that when a district
court concluded that inequitable conduct did not occur, 86–90 percent of the time it gave
as a reason the lack of intent to deceive. By contrast, lack of materiality was given as a
reason only 51–57 percent of the time (Swanson, 2014).3
Taken together, the studies suggest that most inequitable conduct appeals tend to
precipitate into contests over one aspect of the analysis: intent to deceive.
4. Materiality
While the materiality aspect of inequitable conduct is less likely to be important in Federal
Circuit analyses, research has nevertheless provided some interesting insights about this
element of the doctrine.
We have already mentioned the weakness of cumulativeness as a prophylactic to find-
ings of materiality, and thus, to judgments of inequitable conduct. Of perhaps greater
interest, however, are the various forms of materiality expressed in the inequitable conduct
doctrine. While most analyses (62 percent) address only evidence concerning information
an applicant omitted—for example, the failure to disclose a piece of prior art—a surpris-
ing number (14 percent) address only evidence concerning some act an applicant took,
such as disclosing a misleading affidavit. Another 13 percent of analyses address evidence
involving omissions and acts, supposedly undertaken by a patent applicant with intent to
deceive the Patent Office (Petherbridge et. al., 2011).
Research has also categorized more specific forms of behaviors and actions considered
material to patentability. The frequency with which they are observed is shown in Table
16.1.
The ratios of the type of alleged misconduct to outcomes expressed in the table provide
4
little indication that the inequitable conduct doctrine differentiates between different sorts
of material misconduct. With one exception, patentee success appears equally likely no
matter the act or omission. The exception is procedural materiality, acts such as paying
Table 16.1 Categories of alleged material misconduct and patentee outcome4
Specific conduct Patentee success Patentee failure Total

Non-disclosure art 161 58 219
Non-disclosure data 27 4 31
False affidavit 27 13 40
Inventorship 19 4 23
Mischaracterized 21 6 27
Procedural 5 7 12
Other 45 18 63
3
These percentages do not add up to 100 percent because sometimes a decision failed to find
either intent or materiality.
4
From Petherbridge et al. (2011). Because some analyses involved multiple types of alleged
material misconduct, the totals indicated on this chart exceed the number of analyses n=361
collected.

a small entity fee when a large one is required, or inequitably seeking a petition to make
special—which could have the effect of causing a patent to issue more quickly than it
might have if it had to take its place in the examination queue. These cases are, however,
relatively rare.
B. Theory and Purpose
To test the theoretical underpinnings of inequitable conduct, researchers have applied a

novel methodology: using patents to evaluate whether the inequitable conduct doctrine
advances its theoretical goals (Petherbridge et. al., 2013). Briefly, researchers built several
custom datasets that permitted the comparison of patents involved in inequitable conduct
appeals, and most especially those found unenforceable after appeal, with a variety of
other patents. By comparing the characteristics of patents found unenforceable with
those of patents that were unlitigated, litigated, invalid, invalid due to obviousness, and
invalid for lack of sufficient disclosure, researchers were able to isolate the characteristics
of patents found to be unenforceable for inequitable conduct.
As noted earlier, the general theory of inequitable conduct is that it serves to police
the relationship between applicants and the Patent Office, and provides incentives for
applicants to take care to fully disclose known information and to avoid misrepresenta-
tions. Petherbridge and his colleagues therefore predicted that if the inequitable conduct
doctrine works well, patents held unenforceable should be distinguishable from other
patents according to features that could associate with patentees taking unreasonable—in
view of the theory of the law—risks during patent prosecution.
The study theorized, and tested for, three patent characteristics that should correlate
with risky prosecution behavior: the number of parent applications or a patent’s pendency
in prosecution; the number of listed prior art references; and the number of patent claims.
The authors predicted that patents found unenforceable should have more parents, longer
pendency, fewer listed prior art references, and more claims than other patents. The
observations show, inter alia, that when compared to patents held invalid for obviousness
and for lack of sufficient disclosure, unenforceable patents have significantly more parent
applications/longer pendency; significantly fewer prior art references; and significantly
more claims.
These observations suggest that the inequitable conduct doctrine may in practice align
relatively well with its underlying theory. Even so, the researchers did address several
alternative explanations for the observations, and noted that future work was necessary
to confirm the study’s interpretation of the data. In addition, the researchers emphasized
that the larger question of whether the inequitable conduct doctrine is, on balance, effi-
cient, remains outstanding—a conclusion also reached by a contemporaneous economic
analysis (Cotter, 2011).
In the vein of efficiency, Nolan-Stevaux (2005) and Brown (2008) offer some provoca-
tive, if indecisive, findings. Nolan-Stevaux (2005) reports that for the period 1997–2004,
district courts found that of the 18 patents in the study where a court concluded that
inequitable conduct was present, 16 (89 percent) were also held invalid; in contrast, of
the 74 patents where no inequitable conduct was found and validity was also ruled upon,
only 18 (25 percent) were held invalid. Brown’s subsequent study, addressing cases in the
period 2006–07, observes that of the 14 instances in which inequitable conduct was found

and validity ruled upon, the patent was also found invalid in all 14; in contrast, of the 42
instances in which no inequitable conduct was found, but validity was ruled upon, in only
ten (24 percent) was the patent also found valid.
The theoretical significance of these observations is, of course, ambiguous. If the
within-patent correlation of findings of inequitable conduct with findings that at least
some claims are invalid is due to the fact that material information in the possession of
applicants was withheld from the Patent Office, or misrepresented during the prosecution
of the involved patents, one can propose that it supports a conclusion like the one reached
by Petherbridge et. al., (2013)—that inequitable conduct may in practice align well with
its underlying theory.
Alternatively, although not exclusively, one can look at the correlation and propose that
it might be evidence that inequitable conduct is an inefficient component of patent law,
because it may be mostly redundant to invalidity. That interpretation is, unfortunately,
no more certain than the “it aligns well with its theory” interpretation, because invalidity
doctrine will never operate perfectly; and even if invalidity law might be used in cases
where patents were obtained by inequitable conduct, that would not mean that the incen-
tives provided by inequitable conduct do not add more than they cost to the patent system.
Both of these possible implications are further dampened in strength because the stud-
ies do not reveal whether the evidence used to invalidate the patents in the cases examined
is the same evidence used to argue for unenforceability. If the evidence used for invalidity
is different from that used for unenforceability, the meaning of the within-patent correla-
tion would seem to become much more complex.
The observations are also difficult to interpret from a theoretical perspective because
of the narrow picture they present. We do not, for example, have data for the set of
cases in which inequitable conduct was found, but invalidity was not adjudicated (or the
reverse). In addition, invalidity and unenforceability need not travel together in time. A
patent found not unenforceable today can have claims found invalid years from now, and
claims found not invalid today may be found unenforceable in the future. Also vexing is
the concern of “herding,” a well-documented characteristic of judicial decision making in
which judges tend to decide all issues in the same direction. Applied here, a judge might
conclude that someone who behaves inequitably toward the public also deserves to have
patent claims invalidated—a conclusion the same judge might not reach in the absence
of inequitable behavior.
Taken together, these observations, while not theoretically decisive, are nevertheless
stimulating and will hopefully encourage a more complete study.
C. The Prevalence of Inequitable Conduct Claims
The inequitable conduct doctrine has long functioned under a pall of sharp and repeated
criticism (e.g., Lynch, 1988; Mammen, 2009; Dolak, 2010; Wasserman, 2008; Goldman,
1993). Central to this criticism is the assertion that allegations of inequitable conduct are
overused and abused, so much so that the doctrine is damaging the patent system.
In 1984, for example, Judge Giles Rich observed that “‘Fraud in the PTO’ has been
overplayed, is appearing in nearly every patent suit, and is cluttering up the patent
system.” Kimberly-Clark Corp. v. Johnson & Johnson, 745 F.2d 1437 (Fed. Cir. 1984). Not
long after, in 1988, Judge Nichols coined the now famous “plague” reference by writing:

“the habit of charging inequitable conduct in almost every major patent case has become
an absolute plague”. Burlington Industries v. Dayco Corp., 849 F.2d 1418 (Fed. Cir. 1988).
The idea that claims of inequitable conduct are a “plague” upon the patent system
has captured the imaginations of researchers like few other ideas in patent law. Almost
certainly for that reason, the idea has drawn more attention from researchers than any
other issue involving the doctrine. Unfortunately, the idea itself is somewhat vague,
personally normative, and hyperbolic. Perhaps for those reasons, and perhaps also due to
a dearth of theorizing around inequitable conduct, researchers’ efforts to find evidence for
“the plague” has produced some of the most ambivalent and problematic research there
is concerning inequitable conduct.
Broadly speaking, research exploring the idea that claims of inequitable conduct are a
plague upon the patent system looked for evidence in three different forms: the frequency
of written opinions addressing inequitable conduct; the frequency of inequitable conduct
claims in pleadings; and the outcomes of inequitable conduct claims. Because we do not
think the studies reveal much about whether there is a “plague” of inequitable conduct
claims, we will instead emphasize the more concrete contributions that studies framed
around the idea of a plague have made to knowledge about inequitable conduct.
1. The frequency of written opinions addressing inequitable conduct

In the first study to take an original look at district court opinions, Nolan-Steveaux (2005)
reports 244 orders issued between January 1, 1995 and December 31, 2004 “in which
allegations of inequitable conduct were decided in either summary judgment motions or
following a bench trial”—an average of 24.4 written opinions per year. By comparison,
Nolan-Steveaux observes that approximately 2500 patent cases were terminated in 2002
alone.
Reporting raw numbers of written opinions deciding inequitable conduct and compar-
ing the number of written opinions deciding inequitable conduct to all terminated cases
are both useful ways of gauging the frequency of written opinions deciding inequitable
conduct claims. Nolan-Stevaux’s observations thus provide an important perspective—
one that indicates that inequitable conduct resolutions likely represent a relatively small
percentage of terminations.
However, because cases often settle before a judgment on the merits of all issues, the
category of terminations is different from the category of terminations on the merits.
Moreover, a denial of a motion for summary judgment on inequitable conduct, or a find-
ing of no inequitable conduct following a bench trial, would not necessarily even result
in a termination because other issues can remain in patent cases. Accordingly, another
useful perspective is one in which the denominator is not case terminations, but rather
some measure of judicial attention to substantive issues of patent law.
Mack (2006) attempts this approach, seeking to compare the number of published
patent opinions that decide inequitable conduct to the total number of patent opinions
published for the years 2000–04. He reports that “an inequitable conduct adjudication
appeared in 16% to 35% of all reported patent opinions,” depending on the year exam-
ined. We suspect this result is inaccurate, however, for at least the reason that it appears
to be based on a methodological flaw that may produce an inflated number of opinions
deciding inequitable conduct.
Mack draws upon data obtained from PatStats.org, a project of the University of

Houston Law Center, in an attempt to examine the ratio of inequitable conduct opinions
to all opinions.5 A valuable dataset, PatStats has characteristics that make it incompat-
ible with Mack’s methodology. To begin with, according to a May 22, 2005, snapshot
of PatStats.org from the Internet Archive (as well as today’s PatStats webpage), data is
provided on a per-patent and sometimes per-claim basis, and thus “It is therefore not valid
to conclude from these data that patentees won or lost a certain number of ‘cases,’ or that
the accused infringers did so” (PatStats).
Mack does draw such conclusions. This is problematic because a single case that
resolved the enforceability of five patents could, it appears, be counted as five inequitable
conduct cases by Mack’s methodology. Depending on what one wishes to know, that may
be a fine way of quantifying information about inequitable conduct. However, Mack
appears to compare the raw PatStats count of inequitable conduct adjudications to
patent cases.6 Comparing “cases” to individual inequitable conduct determinations within
cases, however, would produce an inflated rate of “cases” addressing inequitable conduct.
Confusing matters further, PatStats captures both district court and appellate decisions
and the data reported in Mack does not distinguish between them. For these reasons, we
are hesitant to rely on the 16–35 percent range reported in that study as an accurate report
of the percentage of decisions with an inequitable conduct adjudication.
Using a similar methodology, Brown (2008) compares the number of opinions involv-
ing inequitable conduct to the overall number of patent opinions. Although Brown
acknowledges the per-patent limitation of the patent data, and reports both per-patent
and per-case data for the years 2006 and 2007, Brown repeated the data from Mack for
the years 2000–04 and provides per-patent data for 2005. For these reasons, we are of the
view that Brown’s observations for 2006 and 2007 are likely to be the most reliable—viz.
in 2006, 12 percent (43 out of 359) of all (trial and appellate) patent opinions involved an
issue of inequitable conduct, and in 2007, 15 percent (65 out of 439) involved this type
of issue.7
Finally, some other studies report different measures of the frequency of written
opinions of inequitable conduct. Searching the relevant Federal Circuit databases in
both LEXIS and Westlaw, and limiting by hand and computer script the search results
to opinions that “evince an analysis of a claim of inequitable conduct,” Petherbridge et.
al. (2011) find 338 Federal Circuit opinions analyzing inequitable conduct over a roughly
28-year period spanning the creation of the Federal Circuit to May 27, 2010. Mammen
(2009) reports finding 321 Federal Circuit opinions through 2008. Accordingly, the
Federal Circuit authors, on average, roughly 12 opinions a year that review a judgment
5
Mack notes that “All statistics were obtained from the University of Houston Law Center,
U.S. Patent Litigation Statistics, http://www.patstats.org.”
6
Mack equates “cases” with “reported patent decisions, including those affirmations by the
Federal Circuit under Rule 36 with no published opinion and those decided by the International
Trade Decision and the Court of Federal Claims.” We cannot provide further analysis of this data
as we have been unable to find “case” data for 2000–04 in the archived versions of PatStats.
7
Supporting the concern about the patent/case issue, Brown reports that for 2007, PatStats
reported 95 instances when inequitable conduct was decided on a per-patent basis in 65 cases. We
also note that for 2005–07, the “cases” are identified on PatStats. However, we would reiterate the
cautions given on the PatStats site itself when relying on its data for per-case analyses.

on inequitable conduct. Assuming the court summarily affirmed some judgments under
the authority of Federal Circuit Rule 36 during these periods, the court likely reviewed,
on average, slightly more than 12 appeals each year.
2. The frequency of inequitable conduct claims in pleadings

The studies set out above suggest that decisions involving issues of inequitable conduct are
present in only a very small fraction of patent cases. Nevertheless, a common move has
been to reason that “[b]ecause the vast majority of patent cases settle before trial or resolve
during summary judgment on issues other than inequitable conduct, it can be inferred that
the percent of patent cases in which a litigant plead[s] inequitable conduct is substantially
higher” (Mack, 2006). Researchers have, accordingly, reasoned that even if published
decisions do not reveal a “plague,” one may still be lurking in the morass of pleadings.
Mammen (2009) explores the extent to which inequitable conduct is pled by examining
the frequency with which responsive pleadings in district court proceedings mention
the term “inequitable conduct.” It reports a startling result: answers including the term
rocketed from 110 in 2000 to 1157 in 2008. When compared to the number of new patent
cases filed, Mammen concludes that the rate at which inequitable conduct was alleged in
answers rose from 4 percent in 2000 to 40 percent in 2008.
This report can no longer be taken as reliable because the databases examined by
Mammen contained substantially fewer searchable pleadings for the early years of the
study than for the later years of the study (Rantanen, 2013). The result is a high likelihood
that Mammen’s measurements significantly underestimate the frequency of patent case
pleadings containing “inequitable conduct” in the earlier periods studied. When those
underestimated measures were compared to the number of new patent cases filed, taken
from a separate data source, unreasonably low rates of inequitable conduct pleadings per
case were inferred for the earlier observation periods.
A methodological improvement that focuses on normalizing for the number of cases
for which searchable answers are available has been made available (Rantanen, 2013).
Observations taken with it show that the yearly rate at which patent cases with an answer
had an answer that included the term “inequitable conduct” range from approximately 26
percent in 2000 to 41 percent in 2008. Table 16.2 uses a version of the Rantanen improve-
ment to update to 2015 the percentage of cases with answers containing “inequitable
conduct.” It shows a range of 21–40 percent of cases, as well as a modest decline starting
around 2012. At present, references to “inequitable conduct” continue to show up in
about 20 percent of cases in which an answer is filed in a given year.
Rantanen’s improvement provides a more accurate account of cases with answers
including the term “inequitable conduct,” although some caveats are warranted. To begin
with, the improved method uses answers to find cases, and therefore may not accurately
reflect the number of overall patent cases (as some complaints are never answered). In
addition, this form of measurement misses complaints that might contain allegations
of inequitable conduct, such as those in a declaratory judgment action. Also, relying, as
Mammen and Rantanen do, on the term “inequitable conduct” to isolate relevant plead-
ings does not ensure that inequitable conduct is being pled. The term might appear in a
pleading in a descriptive or background capacity, or inequitable conduct might be pled
using terms other than “inequitable conduct.” Finally, and perhaps most importantly for
those who would infer a “plague” of inequitable conduct claims from Rantanen’s results,

Table 16.2 Cases with answers and answers containing “inequitable conduct”8
Year Baseline cases with Cases with an answer Percentage of cases

searchable answers containing “inequitable with answers containing
conduct” “inequitable conduct”
2000 236 61 26%
2001 436 138 32%
2002 744 218 29%
2003 1022 324 32%
2004 1445 463 32%
2005 1592 547 34%
2006 1814 605 33%
2007 1957 773 39%
2008 2162 869 40%
2009 2062 748 36%
2010 2162 769 36%
2011 2514 708 28%
2012 3507 740 21%
2013 3947 863 22%
2014 3404 761 22%
2015 (through 1008 203 20%
5/1/2015)
the presence of the term “inequitable conduct” in an answer does not mean that the issue
will be material to the litigation or receive any substantial attention.
The Federal Circuit is widely believed to have reacted to the claim that inequitable
8
conduct is overused and abused by patent defendants by attempting to limit the reach of
the doctrine. In Exergen Corp. v. Wal-Mart Stores, Inc., 575 F. 3d 1312 (Fed. Cir. 2008), a
panel of the court attempted to raise the pleading standard to something akin to common
law fraud—believed by the panel to be stricter than the then-current pleading standard for
inequitable conduct. Shortly thereafter, in Therasense, the en banc court sought, inter alia,
to make it more difficult for lower tribunals to find materiality by attempting to make the
standard more stringent (Rantanen and Petherbridge, 2012). Although the decisions are
quite recent, researchers have attempted to understand their impact on the willingness of
defendants to plead inequitable conduct.
8
To obtain this data, the following searches were performed on the Lex Machina database
for documents filed during the relevant year: (1) Search term (Deny OR Denied OR Denies) with
the Patent case type and answer document type selected; (2) Search term “inequitable conduct”
AND (Deny OR Denied OR Denies) with the Patent case type and answer document type selected.
Results were then limited to cases containing a document matching the search algorithm. Due to
the way the Lex Machina search function currently operates, search results are not automatically
limited to only those documents which are actually text searchable, even when the term “patent”
is included as per the original methodology. To obtain a baseline number of searchable answers,
permutations of “deny” were included (as these terms appear in virtually all answers). This issue
essentially disappears as of 2008, as search results obtained using this method are nearly the same
(within 1–2 percent) as those obtained without any keyword limitations.

In an attempt to understand the influence of Exergen and Therasense on inequitable

conduct pleading, Swanson (2014) uses a variant of the pleading counting methodologies
described above to examine pleading rates during three periods: before Exergen, between
Exergen and Therasense, and after Therasense.
For the periods, and with data from Lex Machina, Swanson compares the number
of cases with pleadings containing “inequitable conduct” to the number of “cases in
which a pleading was filed.” While a reasonable approach to normalizing for a baseline
number of cases (provided that the pleadings are text-searchable, which Rantanen
(2013) indicates is not a serious issue in the relevant database by 2008), it appears
Swanson miscalculates the number of “cases in which a pleading was filed” (Swanson,
pers. comm.). For the period May 26, 2011 to April 10, 2013, Swanson reports 17 561
total patent cases with pleadings. This seems to be too high a number of cases—in the
order of just over 8500 a year. This calculation error casts uncertainty on Swanson’s
main empirical finding, that Exergen caused a reduction in pleadings, but Therasense
did not.
After consulting with Swanson, one of us9 re-ran Swanson’s measurements, correcting
for the measurement error, and found 10 742 cases with pleadings filed during the period
studied, with 1413 of those cases containing a pleading with the term “inequitable
conduct.” The resulting rate of cases with pleadings containing “inequitable conduct”
is thus 13 percent, a rate notably lower than that derived by the Rantanen improvement
on Mammen (roughly 21 percent for the period). The most likely explanation for the dif-
ference is that both methods count in different ways. The Swanson method, for example,
picks up cases with complaints that are unanswered (at all or during the periods), while
the Rantanen improvement on Mammen does not. To the extent this occurs, it should
inflate the denominator and produce a rate of inequitable conduct pleading smaller than
that calculated by Rantanen. We do not take a position on one approach being better than
the other. Both represent ways of measuring the extent that inequitable conduct is pled,
and both perspectives should be of interest to researchers.
When the correct measurements are taken the Rantanen improvement of Mammen
(only cases with answers), and the corrected Swanson (cases with any pleading filed), yield
the following rates of inequitable conduct pleading.
One final point for researchers considering the meaningfulness of these data. The
period around and following Therasense has seen a notable upward trend in patent cases
filed. To the extent measurements were taken at a point too early to expect answers for
many cases (because complaints are unlikely to include the term “inequitable conduct”),
the corrected Swanson measurement might produce a somewhat inconsistent fraction
across the periods studied.
D. Outcomes of Inequitable Conduct Claims
1. District courts

Perhaps the most significant study of district court inequitable conduct outcomes,
Nolan-Stevaux (2005) observes that, overall, 25 percent of district court orders deciding
9
Rantanen.

Table 16.3 Cases with pleadings and answers containing inequitable conduct around
Exergen and Therasense10
Year Pre-Exergen Post-Exergen, Post-Therasense

pre-Therasense
Cases with a pleading containing 1375 1397 1430
“inequitable conduct”
Total patent cases with a pleading 5287 6163 10,834
Proportion pleaded 26% 23% 13%
Cases with an answer containing 1237 1270 1355
“inequitable conduct”
Total patent cases with an answer 3196 3622 5860
Proportion with answers 39% 35% 23%
containing “inequitable conduct”
inequitable conduct are decided against the patentee. More specifically, district courts
declined to grant summary judgment of inequitable conduct 50 percent of the time,
10
granted summary judgment of no inequitable conduct 42 percent of the time, and granted
summary judgment of inequitable conduct 8 percent of the time. In decisions following
trial, district courts found inequitable conduct 30 percent of the time and no inequitable
conduct 70 percent of the time.
The 25 percent loss rate for patentees in district court orders observed by Nolan-Stevaux
may be a fairly persistent rate, as an earlier study, by Dunner (1988), finds similarly, that
between 1977–87, “inequitable conduct charges were rejected at the district court level
about 75% of the time.” In addition to reporting numbers of decisions involving ineq-
uitable conduct, which we have already discussed, Mack (2006) and Brown (2008) draw
on the PatStats data to summarize patentee and alleged infringer success on an annual
basis. Mack reports ranges of 15–29 instances of patentees prevailing on inequitable
conduct between 2000–04, and 5–18 instances of alleged infringers prevailing, while
Brown reports ranges of 33–55 instances of patentees prevailing and 12–40 instances of
alleged infringers prevailing for the period 2005–07. Neither calculates the ratios as a win
rate and, due to the cautions on the PatStats page itself and reasons discussed earlier, we
do not do so either.
Finally, Swanson (2014) examines “every final disposition of inequitable conduct on
the merits at the district court level between January 1, 2008, and April 10, 2013.” The
results of Swanson’s study are reported in Table 16.4.
When interpreting the results reported by Swanson, it should be kept in mind that
the Therasense en banc order issued during the period following Exergen and well
10
To obtain this data, the following searches were performed on the Lex Machina database for
documents filed during the relevant year: (1) Patent case type and answer or pleading document
type selected; (2) Search term “inequitable conduct” with the Patent case type and answer or
pleading document type selected. Periods were: January 1, 2008 to August 3, 2009 (pre-Exergen);
August 5, 2009 to May 24, 2011 (post-Exergen, pre-Therasense; May 26, 2011 to April 10, 2013
(post-Therasense). Results were then limited to cases containing a document matching the search
algorithm.

Table 16.4 Swanson data on inequitable conduct in the district court in relation to the
Exergen and Therasense decisions11
Pre-Exergen Post-Exergen, Post-Therasense

pre-Therasense
Number of cases finding inequitable conduct
Inequitable conduct 13 (23%) 8 (12%) 6 (9%)
No inequitable conduct 43 (77%) 57 (88%) 58 (91%)
Reasons for not finding inequitable conduct*
Lack of intent 37 (90%) 49 (86%) 46 (87%)
Lack of materiality 21 (51%) 30 (53%) 30 (57%)
Note: * Percentages exclude cases for which no opinion was available. Some cases failed to find either intent
or materiality.
before Therasense (i.e., during the middle period depicted above). Prior work in other
patent law contexts has shown that decisional behavior can change when signals of large
jurisprudential events are given (Mojibi, 2010; Wagner and Petherbridge, 2012). As the
Therasense order issued into an atmosphere of hostility toward the doctrine and indicated
the court’s willingness to reconsider the doctrine to an unprecedented extent (Rantanen
and Petherbridge, 2012), it is possible that the order itself may have impacted litigant and
district court behavior.
2. The Federal Circuit

11
There are several studies reporting information concerning inequitable conduct outcomes
at the Federal Circuit. In the earliest, Dunner (1986) examines the Federal Circuit’s treat-
ment of inequitable conduct claims during its first three years and observes that the court
found inequitable conduct in four of the 23 cases raising the issue as of October, 1985, and
contends further that the court was even less hospitable to inequitable conduct charges
than the base statistics indicated.
Generally speaking, the pattern of appellate treatment of inequitable conduct observed
by Dunner has persisted. Three significant studies reporting outcome data all indicate
that the Federal Circuit rarely reaches the conclusion that a patent is unenforceable due
to inequitable conduct:
● Dunner (1995): history of Federal Circuit until 1994; 37 unenforceable/134 enforce-

able (1982–94) [per patent] (22 percent).
● Mammen (2009): history of Federal Circuit until 2008; 64 unenforceable/226
enforceable (1983–2008) [per case] (22 percent).
● Petherbridge et. al. (2011): history of Federal Circuit until 2010; 75 unenforceable/
286 enforceable (1983–2010) [per analysis] (21 percent).
11
Based on Table 1 and accompanying text from Swanson (2014). Note that Swanson states
that the percentages of reasons for not finding inequitable conduct “exclude cases for which no
opinion was available” and that “some cases failed to find either intent or materiality.”

Although covering different time ranges, these studies are generally consistent in terms of
methodology. All three studies examine all written decisions, including both precedential
and non-precedential, and do not include Rule 36 summary dispositions. There are minor
methodological differences in the way that data units are assigned. Dunner reports results
at a per-patent level, Mammen reports results at a per-case level, and Petherbridge et al.
report based on each distinct judicial analysis. Despite these methodological differences,
the results reported by these studies are surprisingly consistent.
E. Multi-layer Studies
One challenge in studying litigated cases is that cases may be terminated at different points
along the lifecycle of a case. The parties may settle the litigation, or the defendant may
default, shortly after the complaint is filed. Or the case may proceed to summary judg-
ment, at which point the losing party may cease to contend. A few cases proceed to trial
and a subset of those are appealed. The result is that outcomes may differ at each level.
While population analyses make it possible to draw some comparisons across different
points in litigation, one study has taken the approach of examining outcomes at all levels
for a given period.
In a 2011 report (Lex Machina, 2011), researchers at Lex Machina collected all inequi-
table conduct outcomes in U.S. patent litigation cases terminated between January 1, 2005
and May 31, 2010. This data was then presented in graphical form, largely aggregated
for the entire time period. Relatively little methodology is provided, which may limit the
usefulness of the data. Some results are largely consistent with the data above for the pre-
Therasense time period, while others are less consistent. Over the approximately five-year
period, the researchers observed that about 3000, or about 22 percent of the total cases
terminated during that period, contained inequitable conduct assertions—an amount
substantially lower than observed in the studies discussed above. On the other hand, of the
211 cases with judicial determinations on inequitable conduct at the district or appellate
court level, 41 (about 19 percent) resulted in a finding of inequitable conduct, a result
consistent with the studies discussed above. The 2011 Lex Machina study also contains
data on the stage of the case at which the issue was determined (summary judgment, trial,
post-trial, or at the Federal Circuit), on outcomes at the Federal Circuit, and at different
district courts.
F. Concluding Thoughts: Inequitable Conduct
Empirical research into inequitable conduct has made progress over the last quarter-
century, with some of the most significant advances coming quite recently. Nevertheless,
perhaps the most significant teaching of the studies we reviewed is how little we presently
know about inequitable conduct. While the studies reviewed here undoubtedly fill
important gaps in our empirical understanding of this area of law, they also reveal the
need for additional empirical work and highlight significant limitations in our theoretical
and policy understandings of the doctrine.
Focusing on the empirical, it seems clear that in order to develop a more complete
understanding of the inequitable conduct doctrine, future work will need not only to
examine Federal Circuit opinions, but also to draw more than current work does from

district court decisions, patents and prosecution histories, and, probably, from other
features of cases that do not find expression in written opinions.
Overall, studies of inequitable conduct might benefit from a more reliable and uniform
way of gathering basic data, particularly when one moves away from the Federal Circuit
where opinion collection is pretty reliable and consistent. This will allow researchers to
start from more or less the same basic information, and the consistency might make it
easier to compare future studies and understand their contributions. We also suspect,
however, that idiosyncratic data collections will be of great value to future work, and think
that their interpretation will be enhanced if they can be related back to a more uniform
platform of data.
Turning to the theoretical, more attention should be paid to constructing a comprehen-
sive theoretical framework for inequitable conduct that is capable of being scientifically
applied. At present we have only a narrow example of a rigorous theoretical treatment
(Cotter, 2012), and very general statements of theory—for example, policing inefficient
behaviors (Petherbridge et al., 2013). Further development in this area seems paramount.
Theory and empirics are usually complementary endeavors. Meaningful theoretical work
is, for example, almost impossible without at least some empirical foundation, and empir-
ics are ultimately required to support, and perhaps someday validate, a theory. Given
how undeveloped inequitable conduct theory appears to be, it seems likely that advances
in theory could open new doors to inequitable conduct research as easily as could new
empirical work, and encourage the asking of questions more meaningful than those
concerning case outcomes and frequency of pleadings.
III. PATENT MISUSE
Like inequitable conduct, the doctrine of patent misuse infuses patent law with a public
interest component—one that addresses patent-related behavior deemed to be socially
harmful. While inequitable conduct revolves around an applicant’s conduct in connection
with obtaining a patent, patent misuse centers on what the patent holder does with the
patent once it has issued. And given the utilitarian underpinnings of patent law over
at least the past century, “socially harmful” has in large part meant “anti-competitive”
(Bohannan and Hovenkamp, 2012; Hovenkamp et al., 2013).
The origins of the misuse doctrine lie in tying arrangements in which a seller required
purchasers to purchase an unpatented product or services in order to acquire the patented
product. A classic example is Morton Salt Co. v. G.S. Suppiger Co., 314 U.S. 488 (1942), in
which Suppiger, the holder of a patent on a canning machine, required purchasers of its
machines to also purchase its unpatented salt for use in the machine. Morton, a competi-
tor, sold both canning machines and salt for the machines. Suppiger sued for infringement
of its patent. With heavy reference to the role of public interest in enforcing a patent,
the Supreme Court held that Suppiger had engaged in an unlawful tying arrangement,
the consequence being the dismissal of the suit. Thus, not only was the patent holder
unable to use its patent to restrain competition in the salt market, but its attempt to do
so rendered its patent on the canning machine unenforceable as against its competitor.
The theoretical underpinnings of misuse are woven into its doctrinal analysis. One
strand of misuse theory views it as paralleling—even overlapping with—antitrust law

(Bohannan and Hovenkamp, 2012). Under this approach, the analysis largely resolves
into an antitrust analysis. The legislative pushback against the misuse doctrine contained
in the successive amendments to 35 U.S.C. § 271(d) reflects this perspective: these changes
move misuse toward antitrust law, although not completely (Cotter, 2007).
A second strand of misuse theory views it as applying when the patent holder attempts
to expand its control beyond the statutory grant of rights. Thus, for example, an agree-
ment granting a license to a patent that requires the payment of royalties for a period
extending beyond the patent term is a form of misuse because “a patented invention
‘become[s] public property once [its term] expires,” and “[a]ny attempt to limit a licensee’s
post-expiration use of the invention, ‘whatever the legal device employed, runs counter
to the policy and purpose of the patent laws’”. Kimble v. Marvel Entertainment, LLC, 135
S. Ct. 2401 (2015) (quoting Brulotte v. Thys Co., 1964).
A third theory of misuse focuses on the use of patents (and copyrights) to foreclose
competition, innovation, or access to the public domain (Bohannan and Hovenkamp,
2012). In this view, misuse doctrine should remain tied to intellectual property law’s core
policies, but in a way that moves beyond the vague, yet formalistic, “beyond the scope”
conception.
While patent misuse has been highly theorized by several scholars, there is only one
recent, detailed empirical study of patent misuse litigation. Lim (2014) examines “all
reported U.S. federal opinions that provided substantial analysis of patent misuse from
January 1, 1953 . . . through December 31, 2013.”12 It thus covers a 60-year period during
which numerous legislative and judicial developments occurred in both patent law gener-
ally and misuse law specifically. The study has much to recommend it. Lim examines case
features, such as posture and outcomes, as well as the substantive content of the opinions
themselves. It is also informed by interviews of judges, academics, government officials,
and lawyers.
Broadly speaking, Lim indicates that patent misuse has never been a highly litigated
doctrine, and has in recent years become—in terms of patent challenger success—
increasingly a dead letter. Nevertheless, the study contains a useful description of misuse
litigation, some parts of which we highlight below. A note, however, before we begin. Due
to the relatively small number of cases spread across the 60-year period, Lim necessarily
analyzed some issues in his study across the entire time period. Patent misuse law saw
many changes during that period, which may both complicate the understanding of some
of the data and offer novel opportunities to examine the data.
A. The Number of Misuse Opinions and Outcomes
Between 1953–2013, Lim found 378 cases, at any judicial level, involving any substantive
issue of patent misuse. Lim reports that the number of opinions addressing misuse has
risen over time even as the success rate of misuse assertions has vanished to virtually
nothing. Figure 16.313 vividly illustrates this trend. Between 1953–62, for example, misuse
12
Lim also used a slightly earlier version of this dataset, ending on December 31, 2012 in Lim
(2013). Where data reported only in Lim (2013) is referenced, we have so indicated.
13
From Lim (2014), reprinted with permission.

150
100 Outcome (Misuse)

Not Mentioned/Applicable
Count
No Misuse Found
Misuse Found
Issue of Fact/Procedure
50
0
2
3
96
97
98
99
00
01
–1
–1
–1
–1
–2
–2
53
63
73
83
93
03
19
19
19
19
19
20
10-Year Interval
Figure 16.3 Distribution of cases over time and outcome
was found 40 percent of the time and not found 58 percent of the time; between 2003–13,
misuse was affirmatively not found 67 percent of the time and found only 2 percent of
the time.
B. Forms of Misuse
Lim (2013) finds examples of 11 distinct forms of misuse allegations, including tying,
restrictions on use, royalty-based misuse, vexatious litigation, and others. Of these,
tying was a common misuse allegation in all four 15-year intervals of the study,
although assertions of misuse based on vexatious litigation (including vexatious litiga-
tion, collateral pressure, group boycotts, and bad faith enforcements) have become
more frequent.
One category of misuse that has become more frequent is time-extension misuse,
such as that described in Brulotte v. Thys (Lim, 2013). Time-extension misuse is often
procedurally raised as a defense to a breach of contract claim for royalty payments that
extend beyond the term of a patent (Hovenkamp, 2015). The increase in frequency of
the appearance of this misuse claim/defense may reflect some combination of the per se
status of the time-extension misuse rule, its age, and the nearly overwhelmingly scholarly

ridicule of it over the last 40 years. Nevertheless, it may be reasonable to speculate that at
least this form of patent misuse is alive and (reasonably) well, given the Supreme Court’s
recent ruling in Kimble v. Marvel (2015), which reaffirmed it.
C. What Explains the Observed Decline in Patent Misuse Challenges?
Lim reports that between 1952–82, challengers won misuse decisions at the appellate level
42 percent of the time (Lim, 2013). By contrast, between 2003–12 challengers won only
14 percent of the time (and those “wins” did not produce a single appellate holding that
misuse occurred). These observations thus raise an important question: assuming away
selection concerns, why has misuse law become such a dead letter?
Lim notes two important features of the legal system that correlate with a decline
in patent challenger success with the doctrine. First, as noted earlier here, is important
changes in the statutory law designed to limit the reach of patent misuse doctrine. Second,
is the establishment of the Federal Circuit, which since 1982 has “decided almost all patent
misuse appeals” (Lim, 2013).
Of course, the statutory reforms designed to limit misuse should be much pro-
nounced in the 2003–12 period relative to the 1952–82 period. On the other hand, the
Federal Circuit was not deciding misuse cases in the earlier period. For these reasons, it
is difficult to discern a role for either of these factors in the decline in patent challenger
success.
Nevertheless, Lim suggests that the Federal Circuit’s understanding of misuse has
become very influential, observing that all courts have come to increasingly cite the
Federal Circuit’s Windsurfing Int’l v. AMF, Inc., 782 F.2d 995 (Fed. Cir. 1986) opinion,
as opposed to Morton Salt (Lim, 2014). Specifically, Lim reports that, between 1973 and
1982, Morton Salt was cited in 27 percent of cases, and that by 2003–13, Windsurfing “was
the controlling precedent in 81% of the cases.” To the extent the Windsurfing analysis
was not compelled by statute or precedent, and to the extent it frames misuse in a more
difficult to prove manner, it may be an indicator of a role for the Federal Circuit in the
demise of patent misuse.
D. Concluding Thoughts: Patent Misuse
While the new insights into misuse offered by recent research helpfully improve our
understanding, they also highlight how little is really known and how much more needs
to be done in order for legal scholars to properly understand the nature and role of these
important features of patent law.
REFERENCES
Bohannan, Christina, and Herbert J. Hovenkamp. 2011. Creation Without Restraint: Promoting Liberty and
Rivalry in Innovation. Oxford: Oxford University Press.
Brown, Benjamin. 2008. “Inequitable Conduct: A Standard In Motion”, 19 Fordham Intellectual Property,
Media, & Entertainment Law Journal 593–628.
Chiang, Tun-Jen. 2013. “The Upside-Down Inequitable Conduct Defense”, 107 Northwestern University Law
Review 1243–306.

Cotter, Thomas F. 2007. “Misuse”, 44 Houston Law Review 901–64.

Cotter, Thomas F. 2011. “An Economic Analysis of Patent Law’s Inequitable Conduct Doctrine”, 53 Arizona
Dolak, Lisa A. 2010. “The Inequitable Conduct Gyre Widens”, 50 IDEA 215–45.
Duffy, John F. 2013. “The Inequities of Inequitable Conduct: A Case Study of Judicial Control of Administrative
Process”, 51 Houston Law Review 417–54.
Dunner, Donald R. 1986. “Inequitable Conduct in the Patent Office and Bad Faith Enforcement of Patents”,
55 Antitrust Law Journal 415–19.
Dunner, Donald R. 1988. “Inequitable Conduct: Is the Sky Really Falling?”, 16 AIPLA Quarterly Journal 27.
Dunner, Donald R, J. Michael Jakes, and Jeffrey D. Karceski. 1995. “A Statistical Look at the Federal Circuit’s
Flores, Edwin S., and Sanford E. Warren, Jr. 2000. “Inequitable Conduct, Fraud, and Your License to Practice
Before the United States Patent and Trademark Office”, 8 Texas Intellectual Property Law Journal 299–330.
Goldman, Robert J. 1993. “Evolution of the Inequitable Conduct Defense in Patent Litigation”, 7 Harvard
Journal of Law and Technology 37–99.
Hovenkamp, Herbert J., Mark D. Janis, Mark A. Lemley, and Christopher R. Leslie. 2013. IP and Antitrust:
An Analysis of Antitrust Principles Applied to Intellectual Property Law, 2d ed., Section 3.3. New York, NY:
Wolters Kluwer Law and Business.
Johnson, et al. 2005. “Patstats U.S. Patent Litigation Statistics”, University of Houston Law Center, https://web.
archive.org/web/20050522021128; www.patstats.org/editors_page.html.
Lemley, Mark A. 2012. “The Fractioning of Patent Law”, in Shyamkrishna Balganesh, ed., Intellectual Property
and the Common Law. New York: Cambridge University Press 504–13.
Lex Machina, Inequitable Conduct Defense in Patent Litigation (2005–2010) (Jan. 15, 2011), https://lexmachina.
com/inequitable-conduct-defense-in-patent-litigation-2005-2010/.
Lim, Daryl. 2013. Patent Misuse and Antitrust Law: Empirical, Doctrinal and Policy Perspectives. Northampton,
MA: Edward Elgar Publishing.
Lim, Daryl. 2014. “Patent Misuse and Antitrust: Rebirth or False Dawn?”, 20 Michigan Telecommunications
and Technology Law Review 299–389.
Lynch, John F. 1988. “An Argument for Eliminating the Defense of Patent Unenforceability Based on
Inequitable Conduct”, 16 AIPLA Quarterly Journal 7–20.
Mack, Kevin. 2006. “Reforming Inequitable Conduct to Improve Patent Quality: Cleansing Unclean Hands”,
Mammen, Christian E. 2009. “Controlling the ‘Plague’: Reforming the Doctrine of Inequitable Conduct”, 24
McGowan, David. 2010. “Inequitable Conduct”, 43 Loyola of Los Angeles Law Review 945–80.
Mojibi, Ali. 2010. “An Empirical Study of the Effect of KSR v. Teleflex in the Federal Circuit’s Patent Validity
Jurisprudence”, 20 Albany Law Journal of Science and Technology 559–96.
Nolan-Stevaux, Katherine. 2005. “Inequitable Conduct Claims in the 21st Century: Combating the Plague”, 20
Petherbridge, Lee, Jason Rantanen, and Ali Mobiji. 2011. “The Federal Circuit and Inequitable Conduct: An
Empirical Assessment”, 84 Southern California Law Review 1293–356.
Petherbridge, Lee, Jason Rantanen, and R. Polk Wagner. 2013. “Unenforceability”, 70 Washington and Lee
Rantanen, Jason. 2013. “Recalibrating Our Empirical Understanding of Inequitable Conduct”, 3 Intellectual
Property Theory 98–113.
Rantanen, Jason, and Lee Petherbridge. 2012. “Therasense v. Becton Dickinson: A First Impression”, 14 Yale
Journal of Law and Technology 226–65.
Swanson, Robert D. 2014. “The Exergen and Therasense Effects”, 66 Stanford Law Review 695–724.
Wagner, R. Polk, and Lee Petherbridge. 2012. “ Did Phillips Change Anything? Empirical Analysis of the
Federal Circuit’s Claim Construction Jurisprudence”, in S. Balganesh, ed., Intellectual Property and the
Common Law. New York: Cambridge University Press.
Wasserman, Melissa F. 2008. “Limiting the Inequitable Conduct Defense”, 13 Virginia Journal of Law and
Technology 1–26.
Cases
Burlington Industries v. Dayco Corp. (1988), 849 F.2d 1418.

Exergen Corp. v. Wal-Mart Stores, Inc. (2008), 575 F.3d 1312.
Kimberly-Clark Corp. v. Johnson & Johnson (1984), 745 F.2d 1437.
Kimble v. Marvel Entertainment, LLC (2015), 135 S.Ct. 2401.

Larson Mfg. Co. of South Dakota, Inc. v. Aluminart Prods. Ltd. (2009), 559 F.3d 1317.
Morton Salt Co. v. G.S. Suppiger Co. (1942), 314 U.S. 488.
Precision Instrument Mfg. Co. v. Auto. Maint. Mach. Co. (1945), 324 U.S. 806.
Therasense, Inc. v. Becton Dickinson & Co. (2011), 649 F.3d 1276.
Windsurfing Int’l v. AMF, Inc. (1986), 782 F.2d 995.

17. Remedies
Thomas F. Cotter* and John M. Golden**
14
Contents
I. Introduction
II. Injunctions
A. Legal Overview
B. Economic Debate: Property Rules versus Liability Rules
C. Permanent Injunctions in Law and Practice
1. U.S. law and practice in district courts
2. U.S. law and practice before the International Trade Commission
3. Law and practice in other countries
4. Preliminary injunctions in law and practice
5. Stays
6. Empirical literature
7. Circumstances and frequency of injunctive relief or motions for
same
D. Other Fronts in Empirical Study
1. Injunction content and scope
2. Injunction violation, enforcement, and stays
III. Damages
A. Overview of Legal Doctrine
1. Types of damages
2. Calculation issues
B. Empirical Literature
IV. Other Remedies
V. Conclusion
References
I. INTRODUCTION
Recent debates and legal developments have highlighted the importance and variety of
patent remedies. Empirical work has followed, but is still a work in progress. Until the
last few years, there were relatively few empirical studies of patent remedies, particularly
* Briggs and Morgan Professor of Law, University of Minnesota Law School.
** Loomer Family Professor in Law, University of Texas School of Law. The author thanks
Tsuki Hoshijima and Paulius Jurcys for research assistance. For helpful comments, both authors
thank participants in a conference associated with the preparation of this research handbook.
390

Remedies 391
for countries other than the United States. Further, most published studies have been
descriptive in nature, reporting aspects of the distributions of case characteristics or
results, such as the percentage of cases in which injunctions were awarded or the median
amounts of damages awards. Moreover, certain topics within the field of patent remedies
have never been the subject of empirical analysis at all. Policymakers would benefit not
only from further descriptive studies to fill existing gaps in the literature, but also from
studies (which, fortunately, appear to be on the rise) using more sophisticated statistical
techniques to test explanatory hypotheses.
Section II discusses the literature on injunctions, beginning with an overview of rel-
evant economic arguments and legal doctrine, and concluding with a summary of existing
empirical studies on the frequency and circumstances of injunction grants as well as on
injunctions’ content and scope. Section III discusses the literature on patent damages,
beginning with an overview of the law and economics of damages, before proceeding
to a review of the empirical literature on the prevalence of different types of damages,
damages amounts, and possible explanatory factors. Section IV discusses other remedies
including declaratory judgments, about which there appears to be little or no relevant
empirical literature. Section V concludes by summarizing aspects of the preceding parts
and discusses possible avenues for future research.
In describing the current state of the empirical literature on patent remedies, this chap-
ter largely sidesteps detailed questions about the meaning of results that have so far been
reported. Generally speaking, the state of empirical studies in this area is best described
as nascent, and in many areas rigorous empirical studies are either non-existent or sparse.
Even when substantial raw data has been collected and subjected to appropriate statistical
tests, the interpretation of results is commonly subject to questions about selection effects
that are virtually inescapable in studies of litigation-based data. Because parties can
choose not to litigate questions on which one side is much more likely to prevail, these
selection effects can substantially disconnect win rates in litigation from the degree of
laxity or strictness of the relevant legal test. On top of normal methodological difficulties
in fitting the study of complex legal phenomena to standard statistical techniques, these
complications and limitations mean that many, if not most conclusions about the detailed
state of patent remedies in practice are at best tentative. With these general caveats, this
survey proceeds.
II. INJUNCTIONS
A. Legal Overview
In suits for patent infringement, courts have frequently issued injunctions in order to
enforce a patent holder’s “right to exclude.” These injunctions have commonly come in
one of two major types: (1) permanent injunctions following a final determination that
one or more patent claims are infringed, and (2) preliminary injunctions occurring before
such a final determination on the merits. Other types of injunctions available in some
countries are for the purpose of obtaining or preserving evidence or assets, and are more
analogous to discovery procedures or other types of court orders in the United States
(Cotter, 2013c).

Permanent and preliminary injunctions are commonly backed up by courts’ power

to impose sanctions for contempt, which may include compensatory damages, fines,
further injunctive relief, and even imprisonment (Golden, 2012). Historically, perma-
nent injunctions have tended to issue as a matter of course in patent cases in which a
patent holder has prevailed on the merits. In contrast, preliminary injunctions have
typically been more difficult to obtain, in part because patent claims’ scope or validity
is frequently not as clear as courts believe necessary to justify such relief before a full
trial. Recent years have witnessed greater questioning of the propriety of permanent
injunctions, as well as exploration of how preliminary and permanent injunctions might
be tailored or timed in order best to serve the public interest or a proper balance of
competing private interests.
B. Economic Debate: Property Rules versus Liability Rules
At least when considered from an economic perspective, the question of whether to

address a legal wrong by granting injunctive relief to an injured party has commonly
been characterized as one of whether “property rules” or “liability rules” should apply
to protection of the violated legal entitlement (Cotter, 2013c; Lemley and Weiser, 2007;
Sterk, 2008). As formulated in a classic law-and-economics article by Guido Calabresi and
Douglas Melamed (1972), a “property rule” effectively means “that someone who wishes
to remove the entitlement from its holder must buy it from [that holder] in a voluntary
transaction in which the value of the entitlement is agreed upon by the seller.” Hence, at
least in situations where contempt sanctions are substantial enough to deter violation of
an injunction, a permanent injunction against further patent infringement implements a
property rule: the threat of contempt sanctions effectively acts to require the adjudged
infringer to negotiate a license with the patent holder to the extent the infringer wishes
to act in a way that would otherwise trigger a holding of contempt (Golden, 2012). In
contrast, under a “liability rule,” “someone may destroy the initial entitlement” as long
as they are “willing to pay an objectively determined value for it”—a price to which the
entitlement holder does not have to agree (Calabresi and Melamed, 1972, p. 1092). In the
absence of an injunction or other relief, a monetary award of compensatory damages is
generally viewed as implementing a liability rule, with damages determined by a court
or other government body serving as the objectively determined price for infringement
(Cotter, 2013c; Kaplow and Shavell, 1996; Lemley and Weiser, 2007; Sterk, 2008).
The association of injunctions with a property rule and of monetary relief with a liabil-
ity rule is imperfect. Given legal and practical limits on contempt sanctions, an injunction
might be viewed as simply imposing a higher “price” for continued patent infringement
than would exist in the absence of the injunction. Under certain circumstances, this
price might not suffice to deter continued infringement. Thus, an injunction might be
“liability-rule-like” in the sense that it “merely discourages” but does not absolutely
prevent violation of the associated entitlement (Golden, 2012, p. 1415). Likewise, under
certain circumstances, monetary awards—particularly monetary awards enhanced to a
level that exceeds pure compensation—can have a sufficiently deterrent effect that they are
“property-rule-like” (Golden, 2012) in the sense that they substantially operate to provide
“absolute protection” of the relevant entitlement (Kaplow and Shavell, 1996, p. 723).
Nonetheless, as a matter of empirical fact, violations of injunctions that trigger con-

Remedies 393
tempt sanctions appear to be quite rare.1 Further, as a matter of law and legal practice,
denial of injunctive relief can be associated with the granting of a compulsory license
or “ongoing royalty” (Paice, 2007). Thus, association of patent infringement injunctions
with property rules and compensatory damages for patent infringement with liability
rules seems reasonable on a first cut.
Classic analysis of the relative desirability of property rules and liability rules suggests
that a property rule will be desirable when the obstacles to socially beneficial private trans-
actions are relatively low compared to the information and error costs associated with
government determinations of proper amounts of liability (Cotter, 2013c; Calabresi and
Melamed, 1972; Merges, 1994; Sterk, 2008). Consequently, because tasks of assessing lost
profits or reasonable royalties for patent infringement have commonly been viewed both
as non-trivial and as risking undercompensation of patent holders, there has traditionally
been substantial momentum behind arguments that the discrete parties to a patent case
should generally be left to work out their differences under a property-rule regime, rather
than have a court or other government body impose a compulsory license or “ongoing
royalty” or, alternatively, award retrospective compensatory damages in a later lawsuit
(Cotter, 2013c; Merges, 1994; cf. Golden, 2010). Although the classic property-versus-
liability rule analysis has been subjected to substantial criticism (Sterk, 2008; cf. Ayres
and Talley, 1995), it has continued to provide a basic framework for thinking and debate.
Additional lines of pro-property-rule arguments have appeared. Henry Smith has
highlighted informational benefits of property rules, which can do more than liability
rules to encourage a right holder to develop, exploit, and disseminate information about
potential uses of an object of entitlement (Smith, 2004; Smith, 2007). Even while giving
pole position to liability rules, Louis Kaplow and Steven Shavell have recognized that,
from the standpoint of social desirability, property rules might overtake their rivals by
encouraging private bargains (Kaplow and Shavell, 1996). Relatedly, Robert Merges has
suggested that property rules can encourage investment “in institutions that lower the
costs of [intellectual property rights] exchange” and thereby enable better social outcomes
than a “suboptimal liability rule” that has less generative effect (Merges, 1994, pp. 2655,
2671–72).
At the start of the twenty-first century, however, there has been substantial pushback
against the theoretical underpinnings for a property-rule approach to patent law. This
pushback has been particularly marked in the United States, where commentators have
pointed out that various prominent modern technologies—including many software,
semiconductor, and communications technologies—appear prone to situations in which
obstacles to efficiency-promoting private transactions are high. Such situations can arise,
for example, because of high information costs associated with private parties’ identifica-
tion and comprehension of patent boundaries (Bessen and Meurer, 2008; cf. Sterk, 2008)
or, alternatively, because of a high risk of strategic “holdout” behavior—behavior that
is often facilitated by the complicated, multi-component nature of certain technologies
1
Study of a “data set of 143 patent-infringement injunctions issued by U.S. district courts in
2010” revealed that, as of a few years after their issuance, only three were known to have resulted
in later contempt proceedings, “with only one of these proceedings known to have resulted in an at
least partial finding of contempt” (Golden, 2014, p. 2095).

(Cotter, 2013c; Lemley and Weiser, 2007). At least in the United States, an aggravating,
if not foundational, cause of such modern concerns with patent search and holdout has
been a sharp rise in patent enforcement by patent assertion entities (PAEs), commonly
described as forms of “non-practicing entities” (NPEs) or, more pejoratively, “patent
trolls”—“a class of patent owners who do not provide end products or services them-
selves, but who demand royalties as a price for authorizing the work of others” (Golden,
2007, p. 2112; cf. Golden, 2013). PAEs tend to be less vulnerable to countersuit for patent
infringement than entities, such as major manufacturing firms, that directly exploit the
sorts of technologies in which they own patent rights (Golden, 2007). Thus, PAEs are less
subject to a form of litigation discipline for patent demands that might previously have
more regularly defused problems associated with information costs or holdout opportuni-
ties (Golden, 2007).
In sum, particularly in the United States, there has been substantial recent debate over
when and to what extent injunctive relief is proper in patent law. With this theoretical
backdrop, we now turn to discussion of law and practice with respect to injunctive relief
in the United States and other countries.
C. Permanent Injunctions in Law and Practice
Permanent injunctions against patent infringement are injunctions that courts issue after
determining both that a patent claim is infringed and that the claim survives any relevant
challenges to the claim’s validity or enforceability (Mueller, 2013).2 In many countries,
courts have commonly issued such injunctions as a matter of course. But there have long
been exceptions to the general availability of injunctive relief, and the United States has
experienced a recent shift toward making the requirements for issuance of a permanent
injunction more demanding.
1. U.S. law and practice in district courts

A watershed moment in U.S. law and practice occurred in 2006 with the decision by the
U.S. Supreme Court in eBay Inc. v. MercExchange, L.L.C. (2006). In the wake of eBay,
U.S. law and practice shifted toward markedly greater scrutiny of the appropriateness of
injunctive relief, with denials of motions for permanent injunctions becoming substan-
tially more common than in the past.3
2
Some countries, including Germany and China, have a bifurcated system under which valid-
ity is determined, if at all, not in the infringement proceeding but rather in a separate invalidation
action. In such systems, a court may enter an injunction before the invalidation proceeding has run
its course. If the patent is subsequently invalidated, the injunction is lifted (Cotter, 2013c).
3
Some studies have reported, however, that prior to eBay, prevailing patent owners’ motions
for permanent injunctions led to granted injunctions only about 80 percent of the time. Gupta
and Kesan (2015) reported that 381 out of 485 motions for permanent injunctions were granted
in cases filed between January 1, 2000 and May 10, 2006—a 79 percent rate. Isackson (2007) cited
an unpublished March 27, 2007 presentation by Professor Paul Janicke as reporting an 84 percent
grant rate. Other sources reported a 95 percent grant rate pre-eBay (Downing et al., 2007; Elkind
et al., 2007). The discrepancy might reflect settlements of cases where the settlement came after the
filing of a permanent injunction motion and resulted in no injunction issuing. As in other situa-
tions, lack of methodological clarity can make the different statistics difficult to compare.

Remedies 395
Before the U.S. Supreme Court’s decision in eBay, U.S. courts tended to issue perma-
nent injunctions against continuing infringement as a matter of course after a patent
holder prevailed on questions of patent infringement and validity (Golden, 2007). The
U.S. Court of Appeals for the Federal Circuit—the U.S. court with primary responsibility
for hearing appeals in patent cases (Golden, 2009) —acknowledged the possibility of “a
sound reason for denying” such an injunction (Richardson, 1989), but also indicated that
such a reason would be “exceptional” and that issuance of a permanent injunction was the
“general rule” (MercExchange, 2005; Richardson, 1989). The Federal Circuit noted that
“courts ha[d] in rare instances exercised their discretion to deny injunctive relief in order
to protect the public interest (Rite-Hite Corp., 1995),” and the circuit court specifically
recognized “public health” as a public interest that could trigger such a use of discretion
(MercExchange, 2005).
In eBay, the U.S. Supreme Court rejected the Federal Circuit’s “‘general rule’ . . . ‘that
a permanent injunction will issue once infringement and validity have been adjudged’”
(eBay, 2006, pp. 393–94). At the same time, the Court rejected “the categorical denial of
injunctive relief ” based on “expansive principles,” such as a notion that injunctions should
not be granted to a patent holder who is not practicing its patent and has indicated “will-
ingness to license its patents” (eBay, 2006, p. 393). The Court instructed that applicable
law demanded case-by-case analysis under a four-factor test requiring that a patent holder
seeking an injunction show the following:
(1) that it has suffered an irreparable injury; (2) that remedies available at law, such as monetary
damages, are inadequate to compensate for that injury; (3) that, considering the balance of
hardships between the plaintiff and defendant, a remedy in equity is warranted; and (4) that the
public interest would not be disserved by a permanent injunction (eBay, 2006, p. 391).4
The Federal Circuit subsequently held that the Court’s eBay decision abrogated any
presumption of irreparable injury in favor of the patent holder (Robert Bosch, 2011).
The Federal Circuit separately held that qualifying irreparable injury must have a “causal
nexus” with patent infringement (Apple, 2013). The Federal Circuit has also indicated
that injunctions will be difficult to obtain for enforcement of standard-essential patents
(SEPs) whose owners have committed to licensing them on fair, reasonable, and non-
discriminatory (FRAND) terms (Apple, 2014). The Federal Circuit has recognized that,
in lieu of a permanent injunction, a trial court may award an “ongoing royalty” for
continuing infringement, but the Federal Circuit has suggested that trial courts should
not necessarily be eager to do so (Paice, 2007).
2. U.S. law and practice before the International Trade Commission
A holder of a U.S. patent who believes that infringing matter is being imported into
the United States may seek relief for injury to a U.S. “domestic industry” by filing a
complaint with an independent federal agency, the U.S. International Trade Commission
(ITC) (Cotropia, 2011) (19 U.S.C. § 1337). The two basic forms of relief that the ITC may
provide are (1) an exclusion order prohibiting entry of certain articles into the United
4
Gergen, Golden, and Smith (2012) discussed the eBay test and a possible problem with
respect to the Court’s use of the past tense in the “irreparable injury” prong.

States; and (2) a “cease and desist” order that might, for example, prohibit the sale of
infringing matter that has been imported (19 U.S.C. § 1337(d)–(f); Cotropia, 2011). The
ITC generally lacks power to award damages (Mueller, 2013), but for relief from an exclu-
sion order that has not yet passed Presidential review, for example, the ITC may require
payment of a bond to be forfeit to the complainant if the exclusion order survives such
review (19 U.S.C. § 1337(e), (j)(3); Chien and Lemley, 2012). For violation of ITC orders,
the agency may impose a fine (19 U.S.C. § 1337(f)(2)) or, in cases of repeated efforts at a
violation, forfeiture of imported items (19 U.S.C. § 1337(i)). The ITC may also impose
monetary or non-monetary sanctions, including the shifting of attorneys’ fees, for “abuse
of discovery and abuse of process” (19 U.S.C. § 1337(h); 19 C.F.R. § 210.4(d)). The ITC’s
issuance of exclusion orders is not subject to the eBay test for permanent injunctions
issued by district courts, and this has led to concern that the ITC could allow an end-run
around restrictions on injunctive relief that can help prevent disproportionate hardship
to adjudged infringers (Chien and Lemley, 2013; Cotter, 2013c).5 On the other hand, ITC
exclusion orders are subject to Presidential review, with the President having the authority
to abrogate exclusion and cease-and-desist orders “for policy reasons” (19 U.S.C. § 1337(j)
(2)). In August 2013, the U.S. Trade Representative, acting on behalf of the President,
abrogated an ITC exclusion order and cease-and-desist orders issued against Apple Inc.,
substantially because of policy concerns with enforcing such remedies in relation to
SEPs subject to FRAND obligations (Letter of U.S. Trade Representative, 2013). The
Trade Representative’s letter urged the ITC itself to consider similar concerns in future
determinations of whether exclusion or cease-and-desist orders would be in the public
interest (Letter of U.S. Trade Representative, 2013).
3. Law and practice in other countries

In common law countries other than the United States, such as the United Kingdom,
Australia, and Canada, courts generally have discretion to grant or deny motions for per-
manent injunctions, but the granting of such injunctions after a patent holder prevails
on the merits appears to remain a significantly stronger default than in the United States
(Cotter, 2013c; Sprigman, 2013). In a 2009 decision, the Court of Appeal of England and
Wales indicated that, “although the case for withholding the injunction has to be strong,
it is clear that a permanent injunction can be withheld,” with the test being “whether
enforcement would be ‘grossly disproportionate’” in terms of the hardship inflicted on
the adjudged infringer (Virgin Atlantic, 2009). Courts in the United Kingdom, Canada,
and Australia also appear to feel freer to order the delivery or destruction of infringing
items than courts in the United States (Cotter, 2013c). The U.S. Patent Act does not
explicitly authorize such orders, but U.S. courts occasionally require “destruction,
5
Article 51 of the Agreement on Trade-Related Aspects of Intellectual Property Rights only
explicitly requires members of the World Trade Organization to provide “border measures” against
the importation of “counterfeit trademark or pirated copyright goods,” but, like the U.S., various
countries extend such measures to patent-infringing imports (Cotter, 2013c, pp. 23–24 (internal
quotation marks omitted)). As far as we know, however, South Korea is the only other country
that has a separate forum to deal with infringing imports analogous to the ITC (Cotter, 2013d; cf.
Cotter, 2013c (citing empirical studies of the use of border measures to exclude patent-infringing
imports in Europe)).

Remedies 397
disablement, or delivery of specified material” as part of injunctive relief (Golden, 2012,

pp. 1451–52).
In contrast with common law countries, courts in civil law countries such as Germany,
Japan, the Netherlands, and Switzerland appear typically to award permanent injunctions
to prevailing patent owners as a more straightforward matter of right (Cotter, 2013c). On
the other hand, concerns of disproportionate hardship may provide grounds for denying
an injunction in at least some jurisdictions (Cotter, 2013c). Further, various jurisdictions
allow for compulsory licensing of patent rights in the public interest, a practice explicitly
allowed under specified conditions by the Agreement on Trade-Related Aspects of
Intellectual Property Rights (TRIPs), to which the more than 150 countries that are
members of the World Trade Organization are subject (Cotter, 2013c). Antitrust concerns
or a patentee’s abuse of right can provide further grounds for denying otherwise routine
injunctive relief (Cotter, 2013c).
There are some cross-border issues with respect to injunctions that have been particu-
larly prominent in the European Union. First, the European Union has witnessed intense
debate over the legality of so-called “cross-border injunctions,” which depart from normal
principles of patent rights’ national territoriality by enjoining patent infringement in
other European nations outside the country whose courts issue the injunction (Cotter,
2013c; Trimble, 2009). There has also been controversy about “torpedo actions” by
which a potential infringer can effectively stay proceedings in other jurisdictions by filing
a preemptive declaratory judgment action in a jurisdiction of the potential infringer’s
choice (Cotter, 2013c). Efforts to enable patents with unitary effect for European Union
members and to constitute a Unified Patent Court could help resolve such cross-border
complications (Cotter, 2013c).
4. Preliminary injunctions in law and practice

A preliminary injunction, sometimes called an “interlocutory injunction” (Mullane,
Humphris, and Shand, 2015), is a court order that issues to provide temporary relief
before a final decision on the merits (Mueller, 2013). In the United States, the test for
whether a court should issue a preliminary injunction is a four-factor test similar to that
for whether a court should issue a permanent injunction:
A plaintiff seeking a preliminary injunction must establish that he is likely to succeed on the
merits, that he is likely to suffer irreparable harm in the absence of preliminary relief, that the
balance of equities tips in his favor, and that an injunction is in the public interest (Apple, 2012).
The major difference between the tests for preliminary and permanent injunctions is that,
the merits not yet having been adjudicated, the test for a preliminary injunction requires
a showing of likelihood of success. This showing can be difficult because “[a]n accused
infringer can defeat a showing of likelihood of success on the merits by demonstrating a
substantial question of validity or infringement” (Trebro, 2014, p. 1165). As for perma-
nent injunctions, the party seeking an injunction must show a “causal nexus” between
the alleged irreparable injury and the alleged patent infringement (Apple, 2012). After the
Supreme Court’s decision in eBay, courts no longer apply a rebuttable presumption of
irreparable injury after a showing of likelihood of success on the merits (Robert Bosch,
2011).

With respect to granting preliminary injunctions, other common law jurisdictions

follow approaches similar to that in the United States, although perhaps without as
demanding standards for proof of associated factors (Cotter, 2013c; cf. Sprigman, 2013).
Similar factors are considered in a variety of civil law jurisdictions, including countries
such as Germany, France, Japan, Italy, the Netherlands, and Switzerland (Cotter, 2013c).
5. Stays
As an alternative to denying an injunction in situations where giving the injunction
immediate effect will cause substantial hardship to its object or the public, a court may
issue an injunction but stay its effect.6 U.S. courts may also stay an injunction while its
bases, such as an infringement determination underlying a permanent injunction, are
subject to challenge on appeal. Either a U.S. trial court or the Federal Circuit may grant
such a stay, with relevant considerations—likelihood of success, irreparable injury in the
absence of a stay, balance of hardships, and public interest—largely mirroring those for
a preliminary injunction.7
Other countries may permit stays under similar circumstances (Cotter, 2013c). In
countries that bifurcate infringement and validity determinations, a court that has found
infringement sometimes may stay an injunction pending the conclusion of the validity
proceedings (though the practice is believed to be uncommon in Germany) (Cotter,
2013c).
6. Empirical literature

Until recently, publicly available empirical work on injunctive relief in patent law has been
relatively scarce. This is not surprising given (1) the relative immaturity of empirical studies
of patent law generally (Golden, Merges, and Samuelson, 2014); (2) the relative scarcity,
compared to patent applications, issued patents, and patent case filings, of court-ordered
remedies such as injunctions; and (3) the fact that, before the U.S. Supreme Court’s deci-
sion in eBay, practices with respect to injunctive relief might not have naturally attracted
too much interest because, in many jurisdictions with robust patent systems and no more
than exceptional compulsory licensing, permanent injunctions for patent infringement
tended to issue as a matter of course. This last point might explain the fact that, before
eBay, perhaps the most notable empirical work in the area was that by Jean Lanjouw and
Josh Lerner with respect to preliminary injunctions (Lanjouw and Lerner, 2001). After
eBay, there has been a boomlet of empirical work on newly interesting issues relating to
the availability of permanent injunctions in the United States. A still relatively neglected
concern has been the precise form of injunctive relief that courts provide: injunctions
have commonly been modeled as “off switches” for potential infringing activities, but the
form of injunctions against patent infringement can vary significantly, albeit in ways not
as easy to summarize or study quantitatively as variations in amounts of money damages
(Golden, 2012).
6
In Retractable Techs. Inc. v. Becton, Dickinson & Co. (2010), the district court granted a
motion for a permanent injunction but stayed the injunction “for the longer of: exhaustion of an
appeal . . . or twelve (12) months from the date of entry” because of the possibility of “a public
health emergency” (emphasis omitted).
7
Examples include Standard Havens (1990) and E.I. DuPont (1987).

Remedies 399
7. Circumstances and frequency of injunctive relief or motions for same

Preliminary injunctions have attracted some attention in empirical studies, but perhaps
because of their relative rarity (Golden, 2007), that attention has not been too great,
and the datasets used have tended to be relatively small. Practitioner-oriented studies
have developed some basic descriptive data on preliminary injunctions. A report from a
private litigation analysis firm gave rates for U.S. district courts’ granting of preliminary
injunctions from 1991 to the start of 2013 that vacillate around an average rate of about
30 percent per preliminary injunction motion, with temporary restraining orders having
been granted at a rate of about 38.5 percent (LegalMetric, 2013). For the 1991–2013
period, the firm reported that average numbers of preliminary injunctions and temporary
restraining orders issued annually by U.S. district courts were about 20 and 10, respec-
tively (LegalMetric, 2013).8 A 2015 study by Kirti Gupta and Jay Kesan of injunction
motions in U.S. district courts gave somewhat different numbers: lower preliminary
injunction grant rates of about 23 percent pre-eBay and about 19 percent post-eBay, and
a total number of preliminary injunctions for cases filed from 2000–2012 of 361 (Gupta
and Kesan, 2015). Gupta and Kesan also observed a decline in the percentage of cases
in which preliminary injunction motions were filed: according to their data, 5 percent or
more of cases filed in years before 2006 featured such motions, and less than 3 percent of
the cases filed in the years 2010–2012 featured such motions (Gupta and Kesan, 2015).
There have been some empirical studies of preliminary injunctions outside the United
States. A 2002 study of applications for preliminary injunctions in German district
courts from 1996–2000 and a 2015 study by legal practitioners of preliminary injunction
motions in Australia found even higher grant rates, with a rate of roughly half in the
German courts (Falck, 2002)9 and of about 70 percent in the Australian courts (Mullane,
Humphris, and Shand, 2015). But the sample size in the Australian study was very small:
the researchers had identified only 18 cases in Australia in which a court had ruled on
such a motion since 2009 (Mullane, Humphris, and Shand, 2015). Given the small sample
8
LegalMetric’s average figure of about 30 preliminary injunctions and temporary restraining
orders issued by U.S. district courts per year is greater than a separately developed figure of 19
preliminary injunctions and temporary restraining orders issued by district courts in 2010 (Golden,
2012), but the discrepancy might be explained by the fact that, according to LegalMetric, the rate
of granting preliminary injunctions in 2010 was only about half the 20-year average of 30 percent
(LegalMetric, 2014).
9
Using information provided by district courts in Düsseldorf, Mannheim, Frankfurt,
Hamburg, Braunschweig, Nuremberg, and Munich, Falck reported that German courts considered
over 100 applications for preliminary injunctions annually from 1996–2000. A more recent study
by Köllner and Weber (2014) of publicly accessible data culled from the Darts-IP database for
the district and appellate courts of Düsseldorf from 2002–2013 reported the number of motions
for preliminary injunctions filed there ranging from a low of one (in 2002) to a high of 17 (in
2008), with success rates ranging from 0 (2002–2004) to 100 percent (2006 and 2013). The authors
conceded that there were gaps in the data, however, noting in particular that, according to figures
Kühnen and Claessen attributed to Presiding Judge Ulrike Voß, about 80 such motions were
filed in 2012 (Kühnen and Claessen, 2013). Another study reported that, in France, requests for
preliminary injunctions were made in only 1 percent to 2 percent of cases (approximately five per
year) from 1984–2004, and that on average only one was granted (Véron and Mandel, 2005). For
a discussion of studies reporting the number of motions for preliminary injunctions filed in Japan
from 1998–2002, but not reporting grant rates, see Cotter (2013c).

size, the study understandably devolved largely into a largely non-statistical discussion of
factors considered in individual cases or small subsets thereof (Mullane, Humphris, and
Shand, 2015).
The Lanjouw and Lerner study of 2001 represented an effort at more rigorous statisti-
cal investigation of preliminary injunction motions in the United States. Specifically,
Lanjouw and Lerner (2001) investigated the circumstances of 48 requests for preliminary
injunctions among 252 patent cases filed in six U.S. district courts from January 1990
through June 1991. Using t tests and probit regressions, they found statistically significant
and substantial relationships between plaintiff size, as indicated by a measure such as
employment or sales, and the likelihood that the plaintiff would seek a preliminary injunc-
tion (Lanjouw and Lerner, 2001). They suggested that these empirical results could mean
that the preliminary injunction is a legal mechanism that might improperly tilt markets in
favor of incumbents (Lanjouw and Lerner, 2001).
Some empirical work relating to injunctions has featured in more general studies of
patent litigation. For example, in a pre-eBay study of outcomes in patent cases filed in
1995, 1997, or 2000, Jay Kesan and Gwendolyn Ball (2006) reported that, despite the sense
that permanent injunctions were generally available to prevailing patent holders before
eBay, these injunctions were actually fairly “rare in adjudicated cases: only 19% of cases
ending in trials and only 4% of those terminating in summary judgments included an
injunction.” Further, the permanent injunctions that issued were “most commonly found
in consent judgments and even formal settlements” (Kesan and Ball, 2006). In a later study
comparing (1) patent litigation in 917 cases involving very litigious PAEs with (2) patent
litigation in a matched set of 1311 non-PAE cases, Michael Risch reported that an injunc-
tion had issued in about 22 percent of non-PAE cases as opposed to less than 2 percent of
PAE cases (Risch, 2015). Like Kesan and Ball, Risch noted that most injunctions “were
obtained as part of consent or default judgments” (Risch, 2015).
In the wake of the U.S. Supreme Court’s eBay decision, both legal practitioners and
academics have shown great interest in assessing the likelihood of a grant of a perma-
nent injunction or analogous order by a district court or the ITC. Consequently, there
have been a number of studies specifically directed at assessing overall probabilities of
injunctive relief and relationships between such probabilities and case or party char-
acteristics. Of course, however, the percentages of injunction grants and denials found
directly through empirical studies are potentially subject to substantial selection effects.
Consequently, what studies of such percentages reveal about background probabilities for
parties’ prevailing is often open to question.
Generally speaking, studies of courts’ injunction practices have tended to show that
post-eBay courts have granted motions for permanent injunctions about 75 percent of
the time (Adams and Wicken, 2007; Beckerman-Rodau, 2007; Chien and Lemley, 2012;
Diessel, 2007; Ellis, Jarosz, Chapman, and Oliver, 2008; Grumbles, Hughey, and Perera,
2009; Patstats.org, 2015b; Seaman, 2015; see also Cotter, 2013c). In contrast, in the years
immediately following eBay, the ITC continued to issue exclusion and cease-and-desist
orders as a matter of course after patent infringement was found (Chien, 2008; Chien
and Lemley, 2012). The fact that two other studies yielded a lower average grant rate
for district courts of about three-fifths to two-thirds might be explained by factors
such as (1) limitation to a subset of district courts, (2) further exclusion of decisions
whose rationale could not be determined, and (3) relatively small sample size (Manzo,

Remedies 401
2007; McClelland, 2011). The 2015 study of injunctions by Gupta and Kesan indicated
relatively steady grant rates of about 80 percent for permanent injunction motions filed
both pre-eBay and post-eBay, with the slight inflation of the post-eBay grant rate relative
to other studies possibly reflecting settlements that lead to an agreed-to injunction after
a permanent injunction motion is initially filed (Gupta and Kesan, 2015). As suggested
above, some discrepancies between studies’ results might reflect sometimes subtle dif-
ferences in the classes of injunction motions that are investigated. As in other areas of
empirical study, authors can aid comparison of their results by clearly indicating the
nature of the datasets they have constructed.
Other features of context also appear relevant to a rich understanding of current prac-
tices with respect to injunctive relief, and a number of studies have begun to explore these.
Whether to identify the factors that appear likely to affect courts’ decisions (McClelland,
2011) or more simply to understand the conditions in which injunctions are likely to
appear, various researchers have explored the circumstances in which permanent injunc-
tions have issued or been denied (Golden, 2014). In the first years after eBay, multiple
studies pursued such contextual analysis by tabulating the invocation of various factors in
decisions granting or denying injunctive relief (e.g., Beckerman-Rodau, 2007; Malin and
Rafilson, 2009). Common conclusions were that direct competitive effects from infringe-
ment were substantially associated with courts’ granting of permanent injunctions, and
that a patent owner’s status as a PAE was substantially associated with a tendency to deny
injunctive relief (e.g., Adams and Wicken, 2007; Beckerman-Rodau, 2007; Diessel, 2007;
Malin and Rafilson, 2009).
More recently, and with a larger dataset of decisions for analysis, Christopher Seaman
has used tests for statistical significance of differential grant rates and multivariable
regression models to explore relationships between various contextual factors and the
granting of permanent injunctions (Seaman, 2016). Focusing on the post-eBay period
from mid-May 2006 through the end of 2013, Seaman analyzed a dataset of 218 district
court decisions on contested motions for permanent injunctions that did not involve
enforcement of a design patent (Seaman, 2016). Among other things, Seaman’s results
suggested that injunction grant rates differed markedly for different fields of technologies:
Seaman found that 92 percent of 25 requested injunctions targeting drugs were granted,
whereas the grant rate for injunctions targeting software was only 53 percent (Seaman,
2016). Seaman’s regression results were consistent with at least some of the conventional
wisdom about the relevance of contextual factors. Seaman’s regression results indicated
a statistically significant positive association between the granting of injunctions and a
court finding that the litigants were competitors (Seaman, 2016). Seaman’s regression
results also indicated a statistically significant negative association between the granting
of injunctions and court findings that a patent covered only a relatively small component
of the infringing products or processes (Seaman, 2016).
D. Other Fronts in Empirical Study
The post-eBay boomlet in empirical study of the availability of injunctive relief has left
largely neglected other potential fronts in the study of injunctions in patent cases. But
there have been occasional empirical investigations into other aspects of injunctions, such
as their content and scope or their violation and enforcement.

1. Injunction content and scope

At least two researchers, Marketa Trimble and John Golden (a co-author of this chapter),
have conducted empirical investigations of injunction content and scope. For a 2009
article, Trimble used two approaches to identify patent-infringement injunctions directed
solely at foreign entities: (1) she reviewed all patent cases filed in U.S. district courts in
2004, and (2) she reviewed the 54 injunctions associated with a post-eBay decision dataset
developed by the University of Houston Law Center (Trimble, 2009). Having identified
13 such injunctions, Trimble analyzed their content, noting (1) that some of the injunc-
tions imposed affirmative requirements, such as labeling requirements, as an alternative
or supplement to more typical prohibitory language, and (2) that at least one injunction
included general “do not infringe” language that the Federal Circuit had indicated violates
the specificity requirement of Rule 65(d) of the Federal Rules of Civil Procedure (Trimble,
2009).
These two phenomena observed by Trimble—the occurrence of apparently improper
“do not infringe” language in injunctions issued by U.S. courts and the existence of
different forms for patent-infringement injunctions—became focal points of Golden’s
2012 study of 143 injunctions issued by U.S. district courts in 2010. Golden’s 2012 study
reported on the occurrence of non-standard forms of injunctions, including the mul-
tiple occurrences of each of four subtypes of specially tailored injunctions that did at
least one of the following: (1) prohibited activities such as shipping that were correlated
with patent infringement but did not necessarily constitute infringement; (2) required
action such as destruction, disablement, or delivery of infringing items; (3) formulated
bounds for encompassed subject matter that did not simply reflect patent scope or
the range of items or processes specifically adjudged to infringe; or (4) moderated
injunction scope by including an explicit carveout for exempt activity (Golden, 2012).
Golden’s 2012 study also reported that 82 of the 143 injunctions in the 2010 dataset
featured general “do not infringe” language that Trimble had previously flagged as
apparently improper (Golden, 2012). Occurrence of such language appeared correlated
with at least some contextual factors. General “do not infringe” language occurred at
an 89 percent rate for 19 unconsented-to but unopposed injunctions, at a 57 percent
rate for 83 consented-to injunctions, and at a 44 percent rate for 41 actively opposed
injunctions (Golden, 2012). There was some evidence of subject-matter dependence:
only 12 percent of 25 injunctions directed at biomedical-substance technologies (e.g.,
pharmaceuticals) featured such language, whereas the corresponding percentages were
about 63 percent for 105 injunctions directed at non-biomedical-substance technology
and about 92 percent for 13 injunctions directed solely at patented ornamental designs
(Golden, 2012).
In a follow-up study of 2014, Golden used a slightly updated version of the original
dataset of 2010 injunctions to investigate the types of technologies at which the 2010
injunctions were directed. Golden reported that 61 of the 124 permanent injunctions in
the dataset were directed primarily at relatively mundane mechanical technologies, with
12 additional injunctions being directed purely at ornamental designs for “mechanical
or otherwise tangible macroscopic objects such as forms of furniture” (Golden, 2014,
p. 2097). Among the 19 preliminary injunctions in the dataset, nine each were directed at
mechanical technologies, with a tenth targeting the ornamental design of a box (Golden,
2014). Despite the prominence of software patents among filed patent cases and in

Remedies 403
policy debates, injunctions categorized as being primarily directed at software accounted

for only six of the dataset’s 124 permanent injunctions and none of its 19 preliminary
injunctions (Golden, 2014).
2. Injunction violation, enforcement, and stays

Empirical study of injunction violation and enforcement appears even scarcer than
empirical study of injunction content and scope. This likely reflects the apparently
extreme rarity of contempt proceedings. Golden’s 2014 study reported that contempt
proceedings or an effort to initiate contempt proceedings had been observed for no
more than four of the 143 injunctions in the updated 2010 dataset (Golden, 2014). With
respect to stays, a study by Thomas Kühnen and Rolf Claessen indicated that, in certain
German district courts (2013), requests for stays had success rates of between 9 percent
and 12 percent, with success rates at the appellate level lying between 4 percent and 17
percent.
III. DAMAGES
A. Overview of Legal Doctrine
1. Types of damages

Monetary awards generally can be classified as falling into one of three types: compensa-
tory, restitutionary, or punitive (exemplary). Compensatory damages attempt to restore
the patent owner to the position it would have occupied but for the infringement, whereas
restitutionary awards require the defendant to disgorge some or all of the benefit it
derived from the wrongful act. In patent cases, compensatory damages include awards of
lost profits from lost sales (profits the patent owner would have earned on sales that were
lost as a result of the infringement) (Grain Processing, 1999), as well as, where available,
the recovery of attorneys’ fees and pre-judgment interest. Subject to the laws of specific
jurisdictions, other actual losses may be compensable as well, including price erosion; loss
of profits on complementary products; “springboard” damages intended to compensate
for the loss of future business opportunities; and harm to the plaintiff’s business reputa-
tion. In the U.S., courts routinely award pre-judgment interest (General Motors, 1983),
but under 35 U.S.C. § 285 attorneys’ fees are available only in “exceptional” cases. In a
2014 decision in Octane Fitness, LLC v. Icon Health & Fitness, Inc., the U.S. Supreme
Court overturned a Federal Circuit precedent, Brooks Furniture Manufacturing, Inc. v.
Dutailier International, Inc. (2005), under which fees would be awarded to the prevailing
defendant on the basis of the weakness of the plaintiff’s case only if the plaintiff asserted
claims that were “objectively baseless” and brought “in subjective bad faith,” holding
instead that courts should consider whether a case is exceptional based on the “totality
of the circumstances.” In many other countries, however, courts routinely award at least
some portion of the prevailing party’s attorneys’ fees (Cotter, 2013c), though practice
varies with respect to the recoverability of pre-judgment (and compound) interest (Cotter,
2013c).
In contrast to lost profits, awards of reasonable royalties—commonly described by U.S.
courts as the royalty a willing licensor and willing licensee would have agreed to, from a

date just prior to the date on which the infringement began10—can be viewed either as
compensatory (to the extent they restore the patent owner to the position it would have
occupied had the infringer taken a license) or as restitutionary (to the extent the infringer
must pay back a benefit in the form of the royalty it should have paid, but did not pay,
for the use of the patented invention) (Restatement (Third), 2011, § 42 cmt. f). Another
form of restitutionary relief, disgorgement of the defendant’s profits attributable to
the infringement, is common in many countries (Cotter, 2013c),11 though in the U.S.
it is available only for the infringement of design patents12 (as well as, in certain cases,
copyright and trademark infringement) (35 U.S.C. § 289).
Awards of compensatory damages generally can be justified on the ground that restor-
ing the patent owner to the position it would have occupied but for the infringement is
necessary to preserve the patent incentive (Cotter, 2014a; Sichelman, 2014). In addition,
compensatory damages may serve a deterrent rationale, though one might argue that, to
the extent reasonable royalties merely replicate the bargain the parties themselves would
have struck, they may not adequately deter. Additional deterrence may be achieved,
however, if the defendant would incur costs in complying with an injunction; or if
compensatory damages include the recovery of the prevailing party’s attorneys’ fees and
costs; or if supra-compensatory damages in the form of awards of defendant’s profits or
enhanced (e.g., punitive or treble) damages are permitted.
On the other hand, supra-compensatory damages awards could result in overdeter-
rence of lawful conduct on the part of defendants, a problem that may be addressed
by conditioning such awards on the defendant’s state of mind. In the U.S., for example,
35 U.S.C. § 284 permits courts to award enhanced damages of up to three times the
amount of actual damages sustained. By long-standing judicial interpretation, however,
such awards are available only for “willful” or “egregious” infringement. For a time,
awards of enhanced damages were further constrained by the Federal Circuit’s ruling in
In re Seagate Tech., LLC, which defined “willful” infringement as requiring a showing
of both objective and subjective recklessness (In re Seagate, 2007),13 and by a 2012
10
A less common method U.S. courts sometimes use for calculating reasonable royalties,
known as the “analytical” approach, “focuses on the infringer’s projections of profit for the
infringing product” (Lucent, 2009). In cases in which the patent owner licenses its patent at an
established royalty, a U.S. court may award that royalty, but the prerequisites for establishing an
established royalty are strict (Nickson, 1988). Courts outside the United States award lost profits
and reasonable royalties, as circumstances dictate, subject to standards that are roughly similar,
though occasionally different in certain material respects (Cotter, 2013c).
11
Siebrasse et al. (2008) describe accountings of profits as the “dominant” monetary remedy
in Canadian patent litigation.
12
The statute has been understood to require courts to award the defendant’s entire profits
from the “article of manufacture” bearing the infringing design, not just the profits attributable
to the infringing design itself, but there is room to argue that the relevant “article of manufacture”
is no more than a component or portion of an overall product that is sold as a whole (Samsung,
2016). In other countries in which awards of defendant’s profits are available, courts must grapple
with issues such as how to determine the portion of the defendant’s profits that was attributable to
the patented feature, and whether a portion of allocable overhead or only variable costs should be
deducted from the defendant’s turnover.
13
Seagate overruled earlier case law holding that “[w]here . . . a potential infringer has
actual notice of another’s patent rights, he has an affirmative duty to exercise due care to

Remedies 405
ecision holding that objective recklessness was to be determined by the judge alone (not
d
the jury) (Bard Peripheral, 2012).14 In 2016, however, the U.S. Supreme Court overruled
these precedents in Halo Electronics, Inc. v. Pulse Electronics, Inc., holding instead that,
“although there is “‘no precise rule or formula’ for awarding enhanced damages, a court’s
‘discretion should be exercised in light of the considerations’ underlying the grant of that
discretion,” which the Court characterized as “punitive” or “vindictive.”
Enhanced or punitive damages are also available in a few other countries (including
Taiwan, Australia, Canada, and the United Kingdom), though anecdotal evidence sug-
gests they are rarely awarded in practice (Cotter, 2013a, 2013b, 2013c, 2013e). In addition,
courts in Germany and France sometimes adjust reasonable royalties upwards to reflect
the fact that an infringer avoids some of the burdens faced by a real-world licensee (for
which a real-world licensee might be “compensated” in the form of a lower royalty), such
as monitoring and the payment of licensing fees that are not recoverable even if the patent
is subsequently invalidated (Cotter, 2013c).
Another type of monetary award is an award of “statutory” damages—that is, an
award falling within a range specified by statute and not necessarily tied to any measure
of proven harm to the right holder or proven benefit to the infringer. Statutory damages
are used in several countries, including the United States, in cases involving copyright
infringement, but their use in patent cases is much more limited. China and Russia are two
countries that make available statutory damages for patent infringement. Chinese courts
may consider various factors in determining whether to grant statutory damages at the
low or high end of the specified range (RMB 10 000 to RMB 1 million), and the amount
granted need not be closely tied to proof of actual harm or benefit.15 Statutory damages
can greatly simplify the calculation of damages but, from the standpoint of compensation
or deterrence, risk either over- or undershooting the mark.
2. Calculation issues

In the U.S., damages evidence usually is presented through submission of expert testimony.
Under Federal Rule of Evidence 702, such testimony is admissible only if it “is based
upon sufficient facts or data, . . . is the product of reliable principles and methods, and
. . . the witness has applied the principles and methods reliably to the facts of the case.”
The amount of damages awarded is subject to appellate review under the “clearly errone-
ous” (in the case of a bench trial) or “substantial evidence” (in the case of a jury trial)
determine whether or not he is infringing,” which “affirmative duty includes, inter alia, the duty
to seek and obtain competent legal advice from counsel before the initiation of any possible
infringing activity” (Underwater Devices, 1983). The Federal Circuit had partially overruled this
standard in 2004 in Knorr-Bremse Systeme Fuer Nutzfahrzeuge GmbH v. Dana Corp. (2004) (en
banc), which (among other things) eliminated the so-called adverse inference rule under which
the trier of fact could draw an inference of willfulness from the defendant’s failure to offer in
evidence an opinion of counsel, but the “duty of due care” standard remained in place until
Seagate.
14
The amount of the enhancement (which, as noted above, can be up to three times the actual
damages sustained) following a finding of willfulness has traditionally been decided by the judge
as well.
15
Cotter (2013c) discusses statutory damages under Chinese law. Cotter and Wood (2014)
discuss statutory damages under Russian law.

standard.16 In other countries, parties sometimes present expert testimony or statements

at trial, and courts are sometimes assisted by their own appointed experts (Cotter, 2013c).
The calculation of damages is often a fact-intensive inquiry, sometimes involving
considerable economic sophistication. In a case in which the patent owner claims that
the infringement caused it to lose profits, the patent owner must offer proof of what its
sales would have been but for the infringement (including proof of its capacity to make
those sales), as well as the applicable profit margin that would have been earned on those
sales. In the U.S. and some other countries, proof of lost profits also normally requires
consideration of whether the defendant could have made some or all of the sales it actually
made using a non-infringing alternative.17
As stated above, in the U.S., the most common approach for estimating a reasonable
royalty involves the reconstruction of the hypothetical bargain the parties would have
negotiated as of a date just prior to the date on which the infringement began. The
hypothetical bargain assumes (counterintuitively) that the parties would have bargained
knowing that the patent was valid and infringed, and often involves consideration of the
so-called Georgia-Pacific factors as a supposed aid to this reconstruction (Georgia-Pacific,
1970). Among the more important of these factors are the “rates paid by the licensee for
the use of other patents comparable to the patent in suit,” the “utility and advantages
of the patent property over the old modes or devices, if any, that had been used for
working out similar results,” and the “extent to which the infringer has made use of the
invention; and any evidence probative of the value of that use.” Additional tools that
have sometimes been employed to assist in calculating royalties are the 25 Percent Rule
of Thumb, which assumes that the inventor would receive 25 percent “of the profits from
any infringing sales” (i4i, 2010); the Nash Bargaining Solution, which posits that parties
with equal bargaining power will agree to split the profits from the use of the invention
50/50; and conjoint analysis, which employs surveys to determine the relative importance
of the patented features of a complex device. In recent years, however, courts in the U.S.
have rejected the Rule of Thumb as lacking a sufficiently reliable basis (Uniloc, 2011) and
also use of the Nash Bargaining Solution absent evidence specifically tying the theory to
the facts of the case (Virnetx, 2014). Courts have permitted testimony based on conjoint
analysis in some cases (Apple, 2013),18 and statistical techniques such as regression
analysis also might be useful in appropriate circumstances.19
Other variations on reasonable royalties include awards of FRAND royalties for the
16
In the U.S., the majority (80 percent) of patent trials—excluding those involving abbreviated
new drug applications—are conducted before juries (Barry et al., 2017). Other countries do not
employ juries in patent cases.
17
For example, Grain Processing Corp. (1999), and Panduit Corp. (1978). Cotter (2013c)
discusses varying practices elsewhere.
18
The opinion approved the use of conjoint analysis to determine causal nexus for purposes
of injunctive relief.
19
Weinstein (2010) discusses possible uses of other techniques, including regression analysis, to
estimate the relationship between the patented feature of a product and sales. In addition, courts
in Germany sometimes consult awards made in arbitration decisions involving employee inventor
compensation for guidance in determining royalty rates, and courts in Japan sometimes consult
standard rates for various technologies as reported by surveys conducted by the Japanese Institute
of Invention and Innovation (Hatsumei Kyokai) (Cotter, 2013).

Remedies 407
infringement of SEPs; ongoing royalties awarded in lieu of injunctive relief; and royalties
awarded for purposes of compulsory licensing. For purposes of calculating FRAND
royalties, courts in the U.S. thus far have employed variations on the conventional
Georgia-Pacific factors, with the caveats that the relevant date for the hypothetical negotia-
tion should be just prior to the date on which the standard was adopted, and that the
royalty should not reflect the value of standardization.20 As for ongoing royalties, the U.S.
case law generally permits courts to award a royalty at a higher rate than the rate awarded
for the pre-judgment infringement, on the theory that the patentee is now in a much
stronger bargaining position (ActiveVideo, 2012; Paice, 2007)—a position that, from an
economic standpoint, is dubious (Lemley, 2011). Finally, although many countries’ laws
permit compulsory licensing (i.e., the government compels the patent owner to license the
patented technology, for example to ensure access to essential medicines at an affordable
price), such licensing must satisfy several stringent conditions under Article 31 of the
TRIPs Agreement, and at present is not widely used.21 One major exception is when
the government itself takes a license to use patented technology. In situations in which
the U.S. government uses a patented invention without a patentee-authorized license, the
Court of Federal Claims determines reasonable royalties due using the same standards as
for royalties in other patent cases (Honeywell, 2012).22
B. Empirical Literature
Much of the empirical literature on patent damages consists of reports on the median or
average damages awards in various countries, the prevalence of various forms of damages,
and similar tabulations. Most of this information relates to U.S. patent litigation only,
though there are a few studies focusing on other countries.
An important source of information on patent damages in the U.S. is
PricewaterhouseCoopers (PwC), which publishes annual Patent Litigation Studies based on
information culled from LexisNexis (Barry et al., 2017).23 The 2017 study reported, among
other things, that:
The annual median damages award between 1997 and 2016 ranged from $2.0 million to over
$17.0 million, with an overall median award of $5.8 million over the last 20 years . . . . Excluding
damages awarded before trial (i.e., summary judgment and default judgment), the overall median
award over the last 20 years jumps to $8.0 million (p. 9).
In 2016, the median damages award was $6.1 million, down from $10.2 million in 2015.
In addition, (1) the median jury award was about 15 times the median bench award from
20
Siebrasse and Cotter (2015) discuss the cases to date and the principles they have articulated.
21
Hilty and Liu (2015) recently collected papers on compulsory licensing, including informa-
tion on practice in various countries.
22
Awards of ongoing royalties in lieu of injunctive relief might be analogized to compulsory
licenses, although the license would not necessarily be available on the same terms to non-parties
(Seaman, 2015).
23
“PwC identified final decisions at summary judgment and trial recorded in two Lexis
Advance databases, US District Court Cases and Jury Verdicts and Settlements, as well as in cor-
responding docket entries from LexisNexis CourtLink” (Barry et al., 2017).

2012–2016—that is, $9.5 million versus $0.6 million (p. 10);24 (2) median “damages
awards [we]re almost 4x greater for NPEs” than for practicing entities from 2012–2016
($15.7 million versus $4.1 million) (p. 16);25 (3) from 2007–2016, courts awarded reason-
able royalties in 80 percent of the cases in which courts awarded damages to practicing
entities, whereas courts awarded lost profits in 40 percent of such cases (p. 11);26 and
(4) median damages awards were highest in the medical devices sector and the biotech/
pharma sectors (p. 19).27
Lex Machina’s 2014 Patent Litigation Damages Report covered some of the same
ground for the period 2000–2013, though using a somewhat different methodology from
the PwC study and reporting somewhat different numbers.28 According to this report:
(1) “[o]ut of 36,629 patent cases filed and terminated from 2000 through 2013, only 708
cases (1.9%) involved compensatory damages awards”; (2) “[o]ver $15 billion in damages,
fees, costs and interest have been awarded” during this period, including “over $13 billion
in compensatory damages” consisting of (a) slightly more than $8 billion in reasonable
royalties; (b) slightly under $3 billion in lost profits; and (c) slightly more than $2 billion
in “compensatory lump”—that is, where “the specific sub-type (reasonable royalties or
lost profits) is not specified or the apportionment of the award between sub-types is not
specified”;29 and (3) the median compensatory award for 2013 was $688 000, with the
median reasonable royalty award (based on 25 cases) amounting to $403 000; the median
24
One might speculate that this statistic is driven in part by selection bias—in a low-stakes
case, the parties may prefer to forgo a jury.
25
Note, however, that “[o]ver the last 20 years, practicing entities fared better than NPEs,
enjoying an 11% premium in their overall success rate” (36 percent versus 25 percent), though
success rates in cases that proceed to trial by jury are closer (77 percent versus 70 percent) (Barry
et al., 2017).
26
Some cases involved both types of awards.
27
The study also ranked district courts by median damages awards from 1997–2016, but the
study did not report the number of damages cases on which these medians are based.
28
PwC’s study was based on “final decisions at summary judgment and trial recorded in two
Lexis Advance databases, US District Court Cases and Jury Verdicts and Settlements, as well as in
corresponding docket entries from LexisNexis CourtLink,” involving 2446 “district court patent
decisions issued since 1997”, with monetary amounts adjusted for 2016 dollars (Barry et al., 2017).
The Lex Machina study was based on 36629 cases filed and terminated from 2000–2013, including
708 in which compensatory damages were awarded. These 708 appear to include cases ending in
default and consent judgments. Lex Machina separately reported enhanced damages, attorneys’
fees, costs, and prejudgment interest. (Byrd et al., 2014). Lex Machina excluded “cross-category”
decisions, which it defines as follows:
A cross-category damages award is one awarded on the basis of different claim types, without
apportionment of the amount among those claim types, (or where documents specifying the
types/apportionment are not available). For example, the much-publicized $290 million award
in the November 2013 Apple v. Samsung case (N.D. Cal., 5:11-cv-01846-LHK) was not appor-
tioned between patent infringement and trade-dress infringement; thus Lex Machina classifies
this award as cross-category. (Byrd et al., 2014 )
It appears that PwC would include such cases (Barry et al., 2017).
29
Note, however, that:
the report is limited to the actions of the district courts to best help readers understand the
actual (rather than the legally proper) behavior of those courts. Therefore, this report does not

Remedies 409
lost profit award (based on only four cases) $5 million; and the median compensatory
lump, based on 26 cases, $669 584. The report also tabulated median awards by year,
judge, and law firm (Byrd et al., 2014).30 Lex Machina’s more recent Patent Litigation
Year in Review 2016 updated some of these numbers, reporting median reasonable royalty
damages in 2016 of $3 552 600, based on 38 cases; median lost profits damages of $1
631 231, based on eight cases; and a median “Other/Mixed Damages” of $67 785, based
on 18 cases (Howard and Maples, 2017, p. 32).31 Because of the relatively small number of
damages awards of each type each year, it might come as no surprise that the size of mean
or even median damages awards can vary substantially from year to year.
As for enhanced damages, a study by Moore (2000) of all 1209 patent cases tried in
the United States between 1983–99 disclosed that (1) the trier of fact passed judgment on
the issue of willfulness in 547 cases (45 percent), and found willfulness in 64 percent of
these; and (2) judges considered a damages enhancement in 219 of these cases, awarding
treble damages in 35 percent of them and no enhancement in 30 percent, with a mean
enhancement of 1.69. Moore (2004) reviewed 1721 of the 4254 complaints filed in patent
cases terminating in 1999–2000 and reported that plaintiffs pleaded willfulness in 92.3
percent of these cases. Willfulness was decided in 143 of these cases, or about one-third of
those that went to trial, and was found in 55.7 percent of the cases in which the issue was
decided; and courts increased damages in 55.7 percent of these cases (comprising about
32 percent of all cases reaching the issue, and about 8 percent of all cases going to trial).
In addition, in the 1585 patent cases tried from 1983–2000, the issue of willfulness was
decided in 664, with juries finding willfulness in 67 percent of these and judges in 52.6
percent. According to Seaman (2012), findings of willful infringement from August 2007
(when the Federal Circuit’s Seagate decision tightened the standard for finding willful
infringement) to July 2010 were down only 11 percent (from 48.2 percent to 37.2 percent
in cases in which the issue was decided) from the pre-Seagate (September 2004–August
2007) level.32 Post-Seagate, however, the disparity between juries (61.9 percent) and
contain data on whether or not damages awarded by district courts were ultimately modified
(or overturned) on appeal, or on outcomes in the district courts on remand. (Byrd et al., 2014)
30
The report also showed that damages awards are highly skewed, with 75 percent of com-
pensatory awards from 2000–2013 under $5 million and 75 percent of enhanced damages under
$3.2 million (Byrd et al., 2014). For earlier studies, see Lemley and Shapiro (2007), which reported,
based on a Westlaw search of cases from 1982–2005 in which a court rendered “an opinion
disclosing the royalty awarded,” 58 such cases, with an average royalty rate of 13.13 percent; Moore
(2000), which reported mean and median U.S. damages awards rendered by judges and by juries
in 1209 cases tried from 1983–1999, including 501 in which the factfinder awarded damages; and
Opderbeck (2009), which reported mean and median damages awards based on a random sample
of cases decided from 2002–2007. Finally, Patstats.org published a list of patent jury verdicts from
February 1, 2005, to December 31, 2013, from highest award to lowest (Patstats.org 2015a).
31
The report also observed: “Most individual awards are small, with a few outliers driving
the high totals. Among all damages awarded in cases filed since the year 2000, 90% of the total
compensatory awards have been less than $9.6M, 75% less than $1.7M, and half less than approxi-
mately $170,000” (Howard and Maples, 2017).
32
Seaman’s study was based on the author’s compilation of all “cases that reached a final
decision on the merits regarding willfulness” during the relevant time period, including decisions
on dispositive pre-trial motions, but not default judgments (Seaman, 2012). Moore’s 2004 study
included all decisions from 1999–2000 in which willfulness was decided either “at trial or on

judges (18.5 percent) was more pronounced, and the frequency of damages enhancements
in cases in which willfulness was found decreased from 81.4 percent to 54.9 percent (with
a mean enhancement of 1.98 percent). Seaman’s regression analysis revealed a statistically
significant relationship between a finding of willfulness and two factors—a substantial
defense to infringement and evidence of copying—but not with respect to other factors
such as an opinion of counsel (Seaman, 2012).33
Other developments in the law have brought about corresponding changes. For exam-
ple, Seaman (2015) reported on awards of ongoing royalties from the date of the eBay
decision through January 2015 (57 decisions in all involving 54 separate awards, including
ten in 2014). Seaman reported, among other things, that 40 percent (23) of these cases
were litigated in the Eastern District of Texas; that the leading technologies in these cases
(accounting for 81 percent of them) were software (21), electronics (14), and medical
devices (11); and that the mean and median post-judgment royalty rates were 1.84 and
1.34 times the mean and median pre-judgment royalty rates, respectively.
Several studies centered on the 2014 revision to the standard for granting fee awards
in the U.S. have now been published. Jiam (2015) used “LexisNexis and Westlaw to
consolidate all cases from April 29, 2014, to March 1, 2015, involving a 35 U.S.C. § 285
motion. The cases retrieved were checked against a list on the blog Patently-O.com that
consolidated several fee-shifting motions post-Octane.” Citing research by Chien (2012),
Jiam stated that “[i]n 2011, approximately twenty awards were granted out of eighty-six
cases, while in 2002, approximately ten awards were granted out of fifty cases.” By con-
trast, “[a]fter Octane Fitness, the proportion of fee awards granted under § 285 has more
than doubled” to 26 out of 59, although the actual fees awarded in the 12 cases that had
reached that issue by March 1, 2015, tended to be modest, mostly ranging from $200 000
to $300 000.
In another recent article, Flanz (2016) reviewed all cases reported on WestlawNext
as (1) having been filed prior to October 1, 2013, the date on which the Supreme Court
granted certiorari in Octane Fitness), and (2) having decided a § 285 fee motion between
January 1, 2013, and September 30, 2013 (the nine-month period prior to the grant of
certioriari in Octane Fitness), or from April 29, 2014, through February 1, 2015 (the nine-
month period after). Flanz reported that the court granted the fee motion in whole or in
part in 31 percent of the 36 decisions falling into the former category, and in 44 percent
of the 54 decisions falling into the latter. After analyzing results from Welch’s two-sample
t test, Flanz opined that his results most likely supported a conclusion that Octane Fitness
in fact did change the governing standard for awarding fees, with the effect being more
ispositive motion,” but reported that during that time period “[w]illfulness was only decided if
d
and when the case went to trial” (Moore, 2004). Seaman reported that willfulness was decided on
pretrial motion in 16.8 percent of cases from 2004–2007 and 26.9 percent from 2007–2010 (Seaman,
2012). As noted above, the Federal Circuit’s 2004 decision in Knorr eliminated the adverse inference
rule, among other changes, which may explain why willfulness was more frequently decided pretrial
from 2004–2007 than in earlier years, but not as frequently as in later years. Seaman reported
that findings of willfulness declined more significantly after Knorr (from 63.8 percent in cases in
which the issue was decided from 1983–1999 to 48.2 percent) than after Seagate (to 37.2 percent)
(Seaman, 2012).
33
Seaman (2012) explained the methodology used to identify and compare cases.

Remedies 411
pronounced with regard to motions filed by defendants and in cases involving software
and electronics patents. But as noted above, such conclusions about the meaning of
litigation data is almost necessarily more open to question than the descriptive statistics
themselves.
In an additional study, Barry et al. (2016) “analyzed patent decisions between January
1, 2013 and December 31, 2015 that reference 35 U.S.C. 285 (as reported by LexisNexis)”
and “where a decision on the merits of awarding attorneys’ fees was made.” They reported:
[T]he percentage of attorneys’ fees awards granted in the 16 months prior to Octane and
Highmark was 26%, which jumped significantly in the subsequent 20 months to 41%. The
average number of fees award decisions also increased from about 4 per month to 7 per month.
Barry et al. also reported that:
the median attorneys’ fees awarded was approximately $0.3 million post-Octane and Highmark
while the maximum amount awarded in fees during this time frame was $12.5 million.
Furthermore, post-Octane and Highmark, the median attorneys’ fees awarded represented 82%
of the median amount requested.34
With few exceptions, published data on patent damages for countries other than the
United States are scant. Féliers (2016), for example, used the Darts-IP Database to
34
Similarly, Jones (2015) examined 55 cases interpreting § 285 from April 29, 2014, through
December 31, 2014, of which 26 resulted in a fee award. Defendants brought 42 of these motions
(20 successfully) and plaintiffs 13 (six successfully), for an overall success rate of 47 percent. Using
Docket Navigator, Renaud et al. (2015) reported a large increase in fee motions filed in 2014–2016,
up from 144 in 2013—but that the success rate increased from 35 percent to 37 percent. Lipsitz et
al. (2014) sampled (1) 100 decisions from 2004 or earlier in which district courts considered whether
to grant attorneys’ fee awards under a discretionary totality-of-the-circumstances test; (2) 100 deci-
sions from 2011–2013 that applied the Brooks Furniture rule that the Supreme Court overruled in
Octane Fitness; and (3) the 40 post-Octane fee award cases that were published through September
18, 2014. They reported that fee awards were granted in 42 percent of such cases prior to 2004, in
only 32 percent from 2011–2013, and in 45 percent since Octane Fitness. In addition, Vishnubhakat
(2014) reported that, based on data gathered from DocketNavigator prior to Octane Fitness,
plaintiffs were the recipients of 153 (70.8 percent) of the 216 § 285 district court fee awards from
2003–2013, but that the median award for defendants was substantially higher ($109 466, versus
$46 354 for plaintiffs). Using Lex Machina, Liang and Berliner (2013) reported that fees were
shifted in 208 cases from 2003 through May 15, 2013; that courts granted 57 percent of plaintiffs’
motions (142 out of 248); and that they granted 26 percent of defendants’ motions (66 out of 252).
Using Westlaw to search “for all patent cases from 1985 to 2004 that discussed fee-shifting,” Bessen
and Meurer (2012) identified 352 such cases, with 137 motions for attorneys’ fees having been
granted. Bessen and Meurer also reported mean and median amounts for attorneys’ fees awarded
in the 87 cases for which they were able to determine the amount.
Outside the U.S., McDonagh and Helmers (2013) examined fee decisions from a sample of 18
patent cases filed in the Patents Court of England and Wales from 2000–2008, and Féliers (2016)
reported average awards of costs under Article 700 of the French I.P. Code from 2000–2015.
Anecdotal evidence suggests that fees often are not fully compensatory even in countries in which
they are routinely awarded (Cotter, 2013c). See also European Observatory on Counterfeiting
and Piracy (2012, pp. 55–59), which reported the “percentage of actual legal costs expended by
the” intellectual property rights owner that “is typically recovered in successful civil litigation” in
countries of the European Union, based on legal practitioners’ responses to questionnaires.

tabulate average patent damages awarded from 2000–2015 by the Tribunal de grande
instance de Paris (TGI Paris), the principal trial venue for patent cases in France, as well
as by the Cour d’appel de Paris and the Cour de cassation. Féliers detected no general
trend from the data, reporting that the average peaked at about €900 000 in 2009 (due to
a single decision in which the court awarded over €9 million) and was well under €100 000
in 2015. The number of cases awarding damages was also small, with only 12 of 139
cases filed in 2015 resulting in a damages award under Article 615-7 of the French I.P.
Code.35 Similarly, Dumont (2015) reported mean and median damages of €323 270 and
€60 000, respectively, based on analysis of “483 patent infringement suits encompassing
673 patents” filed in the TGI Paris from 2008 to 2013. Véron (2010) reported average and
median damages awards in the TGI Paris of €220 000 and €40 000, respectively, from 2000
to 2009 (involving 214 cases).36 More recently, the French government published a study
comparing awards in France, Germany, and the United Kingdom from January 1, 2010,
to August 1, 2013, based on decisions available on publicly accessible databases (believed
to cover approximately 25 percent of all decisions rendered during the applicable time
periods). The study reported that cases involving higher damages awards (more than
€100 000) were much more prevalent in the United Kingdom, and that very few (3 percent)
of French judicial opinions on patent damages took into consideration the infringer’s
profits (compared with 36 percent in Germany and 19 percent in the United Kingdom)
(République Française, 2014).
In Japan, Matsunaka (2004) reviewed all cases:
published in the list of IPR related judgments on the Supreme Court website, in which the right
holder claimed damages relating to IP . . . and for which judgment affirming all or part of the
claim was rendered during the period from January 1, 1998, to December 31, 2003.
Matsunaka reported the number of such cases involving patents, utility models, and
design rights for each year as well as the basis on which damages were assessed. According
to Matsunaka, reasonable royalty awards made up the plurality in both patent (40)
and utility model (22) cases from 1999–2003, ahead of defendant’s profits (33 and 12,
respectively) and lost profits (19 and 13, respectively).37 A more recent study examined
all 68 patent cases from January 1, 1999, to March 5, 2013, in which courts awarded
reasonable royalties, and reported that courts awarded a 5 percent royalty in 28 percent
of cases, 3 percent in 22 percent, and 10 percent in 16 percent (all based on the value of
35
Cotter (2016) discusses Féliers in more depth.
36
Leroux and Bourguet (2006) reported average damages awards in the TGI Paris for the years
2003–2005 of €146 747, €260 765, and €111 370. Véron (2001) reported an average damages award
of €152 000 in the TGI Paris for the years 1990–99.
37
Lost profits usually are awarded under Article 102(1) of the Japanese Patent Act, and less
commonly under Article 709 of the Civil Code. In addition, awards of defendant’s profits under
Article 102(2) are viewed as a proxy for the plaintiff’s own lost profits. For critique, see Cotter,
2013c. An earlier study by Toyama analyzed 103 patent and utility model cases involving “165
supported claims for damages” from April 1, 1960, through December 31, 1995, and reported that
“the average supported amount [was] approximately ¥15.08 million,” although for 1985–95 the
figure was ¥23.88 million,” or about $233 000 using then-current exchange rates (Toyama, 1996).

Remedies 413
the defendant’s turnover).38 Over this time, there were only five cases in which the award
exceeded ¥200 000 000 (approximately U.S. $1.7 million, using late 2016 exchange rates)
(Second Subcommittee, 2014; Cotter, 2015).39 In addition, Nakamura (2014) provided a
comprehensive list of all Japanese patent damages judgments from January 1, 2003, to
January 30, 2014 and provided a qualitative discussion of the case law concerning lost
profits, reasonable royalties, and accountings of profits.
In China, Miao et al. used data from the Fabao Legal Database at Peking University
to review 416 judgments entered in five major provinces (Beijing, Shanghai, Guangdong,
Zhejiang, and Jiangsu) from 2007 to 2008. The authors reported average damages of
RMB 106 000, median damages of RMB 70 000, and a range from RMB 3000 to RMB
1 050 000. Notably, 411 of the 416 were for statutory damages, which as noted above are
capped at RMB 1 million (about U.S. $160 000) (Miao et al., 2009). More recent studies
have continued to report both the predominance of statutory damages and average dam-
ages awards well below RMB 200 000 (Hu, 2016; Yuan, 2016).40 Similarly, Sepetys, and
Cox (2009) reported median patent damages awards of $34 722, but based on a sample
of only 14 patent cases litigated between 2002–2008. Based on analysis of 179 IP cases
(including 20 patent cases) altogether, the authors also found no meaningful difference in
median damages awards based on whether the plaintiff was Chinese and the defendant
foreign, or vice versa. A more recent paper by Love et al. similarly found no evidence of
bias against foreign companies in the forums studied (Love et al., 2016).41
In recent years, a small number of scholars have begun analyzing the relationship
between damages amounts and various potential predictors. In a 2009 article, Opderbeck
(2009) reported “no overriding patterns” to U.S. damages awards based on field of art
38
Compare with Matsunaka (2004), who reported that royalty rates in patent cases from
1998–2003 ranged from 1 to 10 percent, with an average of 4.3 percent). Toyama (1996) reported an
average awarded royalty rate of 4.2 percent in 75 patent and utility model cases decided from April
1, 1960, through December 31, 1995.
39
The authors also presented a formula which they claim can predict the rate a court will award
in a given case, starting with the rate used in comparable licenses or the industry standard rate, then
adjusting that rate in light of various weighted positive and negative factors and multiplying the
adjusted rate by a sales coefficient. Overall, in 70 percent of the 54 cases to which this analysis was
applied, the authors were able to predict the rate awarded to within one percentage point. In addi-
tion, Yamaguchi (2016) reported 14 first instance patent damages judgments in 2014, the largest in
the amount of ¥1 568 040 000.
40
Both Hu and Yuan relied on an unpublished study conducted by the IP Research Center of
Zhongnan University of Economics and Law, covering 839 cases in which damages were awarded
from June 1, 2008, to December 31, 2011. According to Hu, this study revealed that statutory
damages were awarded in over 90 percent of cases, and the average compensation for invention
patents came to only RMB 155 412. Hu also cited data provided by the searchable database
China IP Litigation Analysis (CIELA) indicating that, at least in published judgments of China’s
principal IP courts, statutory damages are awarded in over 90 percent of cases in which damages
are awarded. A study by Love, Helmers, and Eberhardt also used data from CIELA and reported,
inter alia, that, based on a study of 471 IP suits “that included at least one claim for [invention]
patent infringement” from 2006–2011, “[s]even of the top eight jurisdictions have an injunction
grant rate of roughly 90 to 100 percent and a median damages award between 80,000 and 150,000
RMB” (Love et al., 2016). Because of CIELA’s restriction to data from published judgments from
a subset of China’s courts, it is unclear how representative results based solely on CIELA data are.
41
Cohen (2016) partially critiqued this, however.

or type of remedy. More recently, however, Mazzeo et al. (2013a) reviewed, among other
things, “340 patent infringement damage awards granted by a judge or jury in United
States district courts from 1995 to 2008 . . . derived from a proprietary dataset owned
by” PwC. The authors presented median annual damages awards for this period of time,
as well as their distribution in absolute terms and by quartile. In addition, Mazzeo et al.
performed regression analyses that they suggested could explain between 64 percent and
77 percent of the variation in the observed patent damage awards. Variables observed to
have the greatest association with higher damages awards include the number of patents
in suit, the average age of the patent, the average number of claims, the average number
of forward citations, whether the defendant was a public company, whether the trial was
to a jury, and the time to trial.
In a subsequent paper, Mazzeo et al. ran regressions involving 261 observations from
1995–2008 to test whether NPE status is independently associated with higher damages
awards. The authors’ regression analyses indicate that once the value-associated factors
identified in the earlier study were taken into account, there was no statistically significant
difference between awards to NPEs and awards to practicing entities over the time periods
studied. On the strength of such results, Mazzeo et al. wrote:
Our analysis suggests that decided cases involving NPEs do not resolve differently than cases
that involve practicing entities as judged along various dimensions. Patent holder success rates
are somewhat lower for NPE cases than for non-NPE cases and, controlling for other factors,
the damages awarded [to NPEs] in cases with valid and infringed patents are somewhat smaller
(though not statistically significantly so)” (Mazzeo et al. 2013b, pp. 901–02).
In a yet more recent study, Ashtor (2015) analyzed “data relating to each patent that has
been held valid and infringed and for which damages have been awarded in U.S. District
Court cases from 2006 to 2011.” The set of patents studied included “nearly 400 patents
from over 200 cases awarding infringement damages during this six-year timeframe,”
and for each patent the author coded “over 70 unique data points” with the informa-
tion coded falling into one of three categories (intrinsic attributes, acquired attributes,
or case and party characteristics). Based on his regression analyses, Ashtor concluded
that there was “strong evidence that upstream technology patents tend to have higher
enforcement values,” and also that “one of the most significant indicators of patent
enforcement value is the extent to which a patent is practiced by its owner.” According to
Ashtor’s study, forward citations are also positively associated with enforcement value,
but they “may be somewhat imprecise identifiers of particularly valuable patents.” By
contrast, Ashtor found a negative association between patent assignment and the size
of court awards. Similarly, in a recent study of French patent decisions, Dumont (2015)
used a quality index based on backward and forward citations and other characteristics
and found a substantial association between the damages awarded in the cases that
plaintiffs won and “objective grounds, related to the quality of the patents involved in
the case.”
Along a different line of inquiry, some studies have attempted to determine whether
compulsory licensing negatively impacts firms’ decisions to obtain patents or to innovate,
but as indicated by Haracoglou (2008) and by Chien (2003), the empirical evidence is
ambiguous. For example, Scherer (1977) reported a decline in patenting activity among
38 firms subject to compulsory licensing antitrust decrees from 1954–1956, but he found

Remedies 415
no significant decline in R&D based on a similar survey of 44 firms conducted in 1975.

As noted above, at present, compulsory licensing is not widely used.
IV. OTHER REMEDIES
Most countries permit third parties to challenge patent validity either in post-grant
opposition proceedings or by means of a civil action of nullification (topics discussed in
other chapters of this book). Third parties in many countries may also have the option
of commencing a civil action for a declaration of invalidity or non-infringement. The
conditions under which this type of remedy is available vary widely from one country to
another. Relevant issues may include whether the petitioning party must be threatened
with an infringement action; which party has the burden of proof on non-infringement;
whether a licensee in good standing is stopped from asserting such a claim; and whether a
declaratory judgment of invalidity results in the cancellation of the patent or is applicable
only inter partes (Cotter, 2013c). From a policy perspective, such claims may serve the
public interest by providing third parties with the freedom to compete unhindered by a
threat of liability. On the other hand, reducing the barriers to asserting such claims could
increase adjudication, transaction, and licensing costs (the last because licensors may be
unable to shift the risk of invalidation or noninfringement to even a willing licensee). We
are not aware of any empirical studies addressing these issues, however.
As for other remedies, in theory, patent infringement can result in criminal penalties
in some countries, but such penalties are believed to be rarely if ever imposed (Cotter,
2013c).42 Likewise, penalties or fines sometimes may be imposed against patent owners
for false patent marking, though again, outside of a brief boomlet of associated qui tam
suits in the U.S., such penalties or fines appear to be rare in practice (Cotter, 2013c).43
Finally, in some countries, infringers may be compelled to issue a public apology.44 But
empirical research on such alternative remedies appears quite scarce.
V. CONCLUSION
Empirical study of patent-infringement remedies has grown substantially in recent years,

but still appears to be in a state of relative immaturity. The more systematic empirical
work that has been done seems largely to be concentrated in the U.S., likely a reflection
42
Buccafusco and Masur (2014) discuss the costs and benefits of criminal penalties for IP
infringement.
43
Liddicoat and Nicol (2013) note an “absence of empirical data on false patent marking in
Australia.” The one exception occurred in the U.S. from 2009–2011, when an expansive judicial
interpretation of the U.S. false marking statute led to a surge in false marking claims. The average
settlement in such cases was reported to be $48 500 through August 3, 2011 (Gray, 2011). The trend
ended with passage of the America Invents Act in 2011 (Golden 2013).
44
Although a “reparative injunction” requiring an apology appears contrary to U.S. case law
holding that patent-infringement injunctions in the U.S. must be directed to preventing further
infringement, a study of 143 injunctions issued by U.S. district courts in 2010 revealed that one did
require an apology (Golden, 2012).

of the fact that patent-infringement remedies in the U.S. have recently become a major
battleground for policy debates and legal argument (Golden, 2010). Even with this surge
in interest, however, there are gaps in current understanding of the nature of existing
court awards, their relation to the licensing practices of private parties, and the causal
factors that determine what forms of relief for patent infringement a court or other gov-
ernment actor provides. In addition, to our knowledge there are few systematic empirical
studies on the relation between the amount of damages requested by the plaintiff and the
amount actually awarded,45 and little on the relation between damages awards and market
size.46 Some of these holes can be addressed through more systematic study of already
publicly available court records. Others, such as relative lack of knowledge of licensing
practices, might only be fully addressed if governments mandate greater disclosure of the
existence and terms of privately agreed patent licenses, a reform that would likely be very
controversial.
Finally, some recent experimental studies examine whether the presence of certain
factors (e.g., conferring a property or liability rule entitlement upon the IP owner, or inde-
pendent creation by the defendant) affect matters such as willingness to license one’s IP
or the amount of damages participants likely would award. Such experimental literature,
which is the subject of another chapter of this book, might give further insights into the
optimal design of patent and other IP remedies.
REFERENCES
Cases
ActiveVideo Networks, Inc. v. Verizon Comm’ns, Inc., 694 F.3d 1312 (Fed. Cir. 2012).
Amado v. Microsoft Corp., 517 F. 3d 1353 (Fed. Cir. 2008).
Apple Inc. v. Motorola, Inc., 757 F.3d 1286 (Fed. Cir. 2014), overruled in irrelevant part, Williamson v. Citrix
Online, LLC, 792 F.3d 1339 (Fed. Cir. 2015) (en banc).
Apple, Inc. v. Samsung Elecs. Co., 678 F.3d 1314 (Fed. Cir. 2012).
Bard Peripheral Vascular, Inc. v. W.L. Gore & Assocs., 682 F.3d 1003 (Fed. Cir. 2012).
Brooks Furniture Manufacturing, Inc. v. Dutailier International, Inc., 393 F.3d 1378 (Fed. Cir. 2005), abrogated
in part by Octane Fitness, LLC v. Icon Health & Fitness, Inc., 134 S. Ct. 1749 (2014).
eBay Inc. v. MercExchange, LLC, 547 U.S. 388 (2006).
E.I. DuPont de Nemours & Co. v. Phillips Petroleum Co., 835 F.2d 277 (Fed. Cir. 1987).
General Motors Corp. v. Devex Corp., 461 U.S. 648 (1983).
Georgia-Pacific Co. v. United States Plywood Co., 318 F. Supp. 1116 (S.D.N.Y. 1970), modified, 446 F. 2d 295
(2d Cir. 1971).
Grain Processing Corp. v. Am. Maize-Prods. Co., 185 F.3d 1341 (Fed. Cir. 1999).
45
Hu, however, compared damages claimed with damages “supported” in the Chinese cases
studied, and reported an overall “supporting ratio” of 35.88 percent (Hu, 2016). Yamaguchi com-
pared damages claimed with damages awarded in Japanese district court judgments for 2014, and
claimed amounts with settlements from 2011–2013 (Yamaguchi, 2016). Love et al. (2015) reported
the “case value” as estimated by plaintiff patent owners in the United Kingdom and Germany from
2000–2013, but did not compare them to actual damages awards. The “overwhelming majority” of
the estimates, though, were less than U.S. $1 million.
46
Cotter (2013c) discusses the latter issue, exploring possible explanations for the differences
between damages awards in the U.S. and France.

Remedies 417
Halo Elecs., Inc. v. Pulse Elecs., Inc., 136 S. Ct. 1923 (2016).
Honeywell Int’l Inc. v. United States, 107 Fed. Cl. 659 (2012).
i4i Ltd. Partnership v. Microsoft Corp., 598 F.3d 831 (Fed. Cir. 2010), aff’d on other grounds, 131 S. Ct. 2238
(2011).
Knorr-Bremse Systeme Fuer Nutzfahrzeuge GmbH v. Dana Corp., 383 F.3d 1337 (Fed. Cir. 2004) (en banc).
Lucent Technologies, Inc. v. Gateway, Inc., 580 F.3d 1301 (Fed. Cir. 2009).
MercExchange, LLC v. eBay, Inc., 401 F.3d 1323 (Fed. Cir. 2005), vacated, 547 U.S. 388 (2006).
Nickson Indus. v. Rol Mfg. Co., 847 F.2d 795 (Fed. Cir. 1988).
Octane Fitness, LLC v. Icon Health & Fitness, Inc., 134 S. Ct. 1749 (2014).
Paice LLC v. Toyota Motor Corp., 504 F.3d 1293 (Fed. Cir. 2007).
Panduit Corp. v. Stahlin Bros. Fibre Works Inc., 575 F.2d 1152 (6th Cir. 1978).
Retractable Techs. Inc. v. Becton, Dickinson & Co., No. 2:07-CV-250, 2010 WL 9034911 (E.D. Tex. May 19,
2010).
Richardson v. Suzuki Motor Co., 868 F.2d 1226 (Fed. Cir. 1989), abrogated in part by eBay Inc. v. MercExchange,
LLC, 547 U.S. 388 (2006).
Rite-Hite Corp. v. Kelley Co., 56 F.3d 1538 (Fed. Cir. 1995) (en banc).
Robert Bosch LLC v. Pylon Mfg. Corp., 659 F.3d 1142 (2011).
Samsung Elecs. Co. v. Apple Inc., 137 S. Ct. 429 (2016).
In re Seagate Tech., LLC, 497 F.3d 1360 (Fed. Cir. 2007) (en banc).
Standard Havens Prods., Inc. v. Gencor Indus., Inc., 897 F.2d 511 (Fed. Cir. 1990).
Trebro Mfg., Inc. v. Firefly Equipment, LLC, 748 F.3d 1159 (Fed. Cir. 2014).
Underwater Devices Inc. v. Morrison-Knudsen Co., 717 F.2d 1380 (Fed. Cir. 1983).
Uniloc USA, Inc. v. Microsoft Corp., 632 F.3d 1292 (Fed. Cir. 2011).
Virgin Atlantic v. Premium Aircraft [2009] EWCA (Civ) 1513, (Eng.).
Virnetx, Inc. v. Cisco Sys., Inc., 767 F.3d 1308 (Fed. Cir. 2014).
Statutes, Regulations, and Other Government Documents
19 U.S.C. § 1337.
19 C.F.R. § 210.4.
Letter of Michael Froman, U.S. Trade Representative, to Irving A. Williamson, Chairman, U.S. International
Trade Commission (Aug. 3, 2013), available at https://ustr.gov/sites/default/files/08032013%20Letter_1.PDF.
Books, Articles, and Other Publications
Adams, Darryl J. and Victoria Wicken. 2007. “Permanent Injunctions After eBay v. MercExchange: The Year
in Review,” 15 Texas Intellectual Property Law Journal 417–56.
American Law Institute. 2011. Restatement (Third) of Restitution and Unfair Competition. St. Paul, MN:
American Law Institute Publishers.
Ashtor, Jonathan H. 2015. “Redefining ‘Valuable Patents’: Analysis of the Enforcement Value of U.S. Patents,”
18 Stanford Technology Law Review 497–545.
Ayres, Ian and Eric Talley. 1995. “Distinguishing Between Consensual and Nonconsensual Advantages of
Liability Rules,” 105 Yale Law Journal 235–53.
Barry, Chris, Ronen Arad, Landan Ansell, Meredith Cartier, and Hye Yun Lee. 2016. PwC 2016 Patent
Litigation Study: Are We at an Inflection Point? New York, N.Y.: PricewaterhouseCoopers.
Barry, Chris, Ronen Arad, Landan Ansell, Meredith Cartier, and Hye Yun Lee. 2017. PwC 2017 Patent
Litigation Study: Change on the Horizon? New York, N.Y.: PricewaterhouseCoopers.
Beckerman-Rodau, Andrew. 2007. “The Aftermath of eBay v. MercExchange, 126 S. Ct. 1837 (2006): A Review
of Subsequent Judicial Decisions,” 89 Journal of the Patent & Trademark Office Society 631–60.
Bessen, James and Michael J. Meurer. 2008. Patent Failure: How Judges, Bureaucrats, and Lawyers Put
Innovators at Risk. Princeton, N.J.: Princeton University Press.
Bessen, James and Michael J. Meurer. 2012. “The Private Costs of Patent Litigation,” 9 Journal of Law,
Economics & Policy 59–95.
Buccafusco, Christopher and Jonathan S. Masur. 2014. “Innovation and Incarceration: An Economic Analysis
of Criminal Intellectual Property Law,” 87 Southern California Law Review 275–334.
Byrd, Owen, Brian C. Howard, and Jason Maples. 2014. Lex Machina Patent Litigation Damages Report. Menlo
Park, California: Lex Machina.
Calabresi, Guido and A. Douglas Melamed. 1972. “Property Rules, Liability Rules, and Inalienability: One View
of the Cathedral,” 85 Harvard Law Review 1089–128.

Chien, Colleen V. 2003. “Cheap Drugs at What Price to Innovation: Does the Compulsory Licensing of
Pharmaceuticals Hurt Innovation?,” 18 Berkeley Technology Law Journal 853–907.
Chien, Colleen V. 2008. “Patently Protectionist? An Empirical Analysis of Patent Cases at the International
Trade Commission,” 50 William & Mary Law Review 63–114.
Chien, Colleen V. 2012. “Reforming Software Patents,” 50 Houston Law Review 325–90.
Chien, Colleen V. and Mark A. Lemley. 2012. “Patent Holdup, the ITC, and the Public Interest,” 98 Cornell
Law Review 1–45.
Cohen, Mark. 2016. Patent Litigation, Local Protectionism and Empiricism: Data Sources and Data Critiques,
China IPR Blog, Mar. 10, available at https://chinaipr.com/2016/03/10/patent-litigation-local-protectionism-
and-empiricism-data-sources-and-data-critiques/.
Cotropia, Christopher A. 2011. “Strength of the International Trade Commission as a Patent Venue,” 20 Texas
Cotter, Thomas F. 2013a. “After a Six-Month Hiatus, Enhanced Damages for Patent Infringement in Taiwan
Are Back,” ComparativePatentRemedies Blog, Aug. 6, available at http://comparativepatentremedies.blogspot.
com/2013/08/after-six-month-hiatus-enhanced-damages.html.
Cotter, Thomas F. 2013b. “Canadian Federal Court of Appeals Affirms Patentee’s Entitlement to Punitive
Damages,” ComparativePatentRemedies Blog, Oct. 8, available at http://comparativepatentremedies.blogspot.
com/2013/10/canadian-federal-court-of-appeals.html.
Cotter, Thomas F. 2013c. Comparative Patent Remedies: A Legal and Economic Analysis. Oxford, UK: Oxford
University Press.
Cotter, Thomas F. 2013d. “The International Trade Commission: Reform or Abolition? A Comment on Colleen
V. Chien and Mark A. Lemley, Patent Holdup, the ITC, and the Public Interest,” 98 Cornell Law Review
Online 43–54, available at http://cornelllawreview.org/files/2013/05/Cotterfinal.pdf.
Cotter, Thomas F. 2013e. “Punitive Damages for Patent Infringement in the UK?,” ComparativePatentRemedies
Blog, Nov. 27, available at http://comparativepatentremedies.blogspot.com/2013/11/punitive-damages-for-
patent.html.
Cotter, Thomas F. 2014a. “Make No Little Plans: Response to Ted Sichelman, Purging Patent Law of ‘Private
Law’ Remedies,” 92 Texas Law Review See Also 25–33.
Cotter, Thomas F. 2014b. “U.S. Supreme Court Reverses Federal Circuit on Declaratory Judgment Burden of
Proof; Comparison with Germany and the U.K.,” ComparativePatentRemedies Blog, Jan. 22, available at
http://comparativepatentremedies.blogspot.com/2014/01/us-supreme-court-reverses-federal.html.
Cotter, Thomas F. 2015. “A Study of Reasonable Royalty Awards in Japan,” ComparativePatentRemedies Blog,
Mar. 23, available at http://comparativepatentremedies.blogspot.com/2015/03/a-study-of-reasonable-royalty-
awards-in.html.
Cotter, Thomas F. 2016. “Patent Damages in France,” ComparativePatentRemedies Blog, Oct. 5, available at
http://comparativepatentremedies.blogspot.com/2016/10/patent-damages-in-france.html.
Cotter, Thomas F. and Nadia Wood. 2014. “Statutory Damages for Patent Infringement in China and Russia, Part
1,” ComparativePatentRemedies Blog, Apr. 28, available at http://comparativepatentremedies.blogspot.com/
search/label/Russia?updated-max=2014-05-01T04:15:00-07:00&max-results=20&start=4&by-date=false.
Diessel, Benjamin H. 2007. “Trolling for Trolls: The Pitfalls of the Emerging Market Competition Requirement
for Permanent Injunctions in Patent Cases Post-eBay,” 106 Michigan Law Review 305–45.
Docket Navigator. 2009. Success Rates for Permanent Injunction after eBay v. MercExchange, Docket
Report Blog, Nov. 20, available at http://docketreport.blogspot.com/2009/11/success-rates-for-permanent-
injunction.html.
Downing, Kristy, et al. 2007. Injunctive Relief after eBay v. MercExchange, Presentation to the AIPLA 2007
Annual Patent Law Committee, Oct. 19, available at http://nebula.wsimg.com/2a9d2d63d6fef6edb24b5082c9
0922da?AccessKeyId=9353E3DADDC2EC8B09B5&disposition=0&alloworigin=1.
Dumont, Béatrice, 2015. “Does Patent Quality Drive Damages in Patent Lawsuits? Lessons from the French
Judicial System,” 11 Review of Law & Economics 355–83.
Elkind, Thomas I., et al. 2007. Strategic Venue Selection and Use of IP Injunctions, Dec. 4, available at http://www.
foley.com/files/Event/528a386c-167b-4179-9552-f5a74d8cf8a8/Presentation/EventAttachment/5882e1f2-85d
9-489b-9f69-fe934b10519b/Session6_AnatHakim.pdf.
Ellis, Douglas, John Jarosz, Michael Chapman, and L. Scott Oliver. 2008. “The Economic Implications (and
Uncertainties) of Obtaining Permanent Injunctive Relief After eBay v. MercExchange,” 17 Federal Circuit
Bar Journal 437.
European Observatory on Counterfeiting and Piracy, Observatory Update on Costs and Damages. 2012.
Available at https://oami.europa.eu/ohimportal/en/web/observatory/observatory-publications.
Falck, Andreas von. 2002. “Einstweilige Verfügungen in Patent- and Gebrauchsmustersachen,” Mitteilungen
der deutschen Patentanwälte 429–38.
Federal Trade Commission. 2011. The Evolving IP Marketplace: Aligning Patent Remedies with Notice and
Competition.

Remedies 419
Féliers, Antoine. 2016. “Montants alloués par les juridictions en contentieux de brevets,” Cabinet Plesseraud,
available at www.plass.com/fr/actualites/montants-alloues-par-les-juridictions-en-contentieux-de-brevets.
Flanz, Scott. 2016. “Octane Fitness: The Shifting of Patent Attorneys’ Fees Moves into High Gear,” 19 Stanford
Gergen, Mark P., John M. Golden, and Henry E. Smith. 2012. “The Supreme Court’s Accidental Revolution:
The Test for Permanent Injunctions,” 112 Columbia Law Review 203–49.
Golden, John M. 2007. “‘Patent Trolls’ and Patent Remedies,” 85 Texas Law Review 2111–61.
Golden, John M. 2009. “The Supreme Court as ‘Prime Percolator’: A Prescription for Appellate Review of
Questions in Patent Law,” 56 UCLA Law Review 657–724.
Golden, John M. 2010. “Principles for Patent Remedies,” 88 Texas Law Review 505–92.
Golden, John M. 2012. “Injunctions as More (or Less) Than ‘Off Switches’: Patent-Infringement Injunctions’
Scope,” 90 Texas Law Review 1399–472.
Golden, John M. 2013. “Patent Privateers: Private Enforcement’s Historical Survivors,” 26 Harvard Journal of
Golden, John M. 2014. “Litigation in the Middle: The Context of Patent-Infringement Injunctions,” 92 Texas
Golden, John M., Robert P. Merges, and Pamela Samuelson. 2014. “The Path of IP Studies: Growth,
Diversification, and Hope,” 92 Texas Law Review 1757–768.
Gray, Justin E. 2011. “False Marking—Settlement Update,” Gray on Claims, Aug. 9, available at www.grayon
claims.com/home/2011/8/9/false-marking-settlement-update.html.
Grumbles III, Ernest, Rachel C. Hughey, and Susan Perera. 2009. “The Three Year Anniversary of eBay v.
MercExchange: A Statistical Analysis of Permanent Injunctions,” Intellectual Property Today 25–29, available
at www.merchantgould.com/portalresource/Three-Year-Anniversary-of-eBay-v-MercExchange.pdf.
Gupta, Kirti and Jay P. Kesan. 2015. “Studying the Impact of eBay on Injunctive Relief in Patent Cases,”
available at http://ssrn.com/abstract=2629399.
Haracoglou, Irina. 2008. Competition Law and Patents: A Follow-On Innovation Perspective in the Biopharmaceutical
Industry. Cheltenham U.K.: Edward Elgar.
Hilty, Reto M. and Kung-Chung Liu, eds. 2015. Compulsory Licensing: Practical Experiences and Ways
Forward. Berlin, Germany: Springer Verlag.
Howard, Brian C. and Jason Maples. 2017. Lex Machina Patent Litigation Year in Review 2016. Menlo Park,
California: Lex Machina.
Hu, Jingjing. 2016. “Determining Damages for Patent Infringement in China,” 47 IIC 5–27.
Isackson, Robert M. 2007. “After ‘eBay,’ Injunctions Decrease,” National Law Journal, Dec. 3, at S1.
Jiam, Hannah. 2015. “Fee-Shifting and Octane: An Empirical Approach Toward Understanding ‘Exceptional’,”
Jones, Darin. 2015. “A Shifting Landscape for Shifting Fees: Attorney-Fee Awards in Patent Suits After Octane
and Highmark,” 90 Washington Law Review 505–43.
Kaplow, Louis and Steven Shavell. 1996. “Property Rules Versus Liability Rules: An Economic Analysis,” 109
Kesan, Jay P. and Gwendolyn G. Ball. 2006. “How Are Patent Cases Resolved? An Empirical Examination of
the Adjudication and Settlement of Patent Disputes,” 84 Washington University Law Review 237–312.
Köllner, Malte and Paul Weber. 2014. “Rumor Has It: Statistische Analyse der Rechtsprechungspraxis bei
Patentverletzungsklagen,” Mitteilungen der deutschen Patentanwälte 429–38.
Kühnen, Thomas and Rolf Claessen. 2013. “Die Durchsetzung von Patenten in der EU—Standortbestimmung
vor Einführung des Europäischen Patentgerichts,” GRUR 592–97.
Lanjouw, Jean O. and Josh Lerner. 2001. “Tilting the Table? The Use of Preliminary Injunctions,” 44 Journal
of Law & Economics 573–601.
LegalMetric. 2013. LegalMetric Nationwide Summary Report: Metrics in Patent Cases, January 1991–January
2013, available at www.legalmetric.com/presentations/Nationwide%20Patent%20Metrics%20Summary%20
2013-01.pdf.
LegalMetric. 2014. Nationwide Preliminary Injunctions in Patent Cases Report, available at www.legalmetric.
com/nationwide/patent/preliminary_injunction.html.
Lemley, Mark A. 2011. “Ongoing Confusion over Ongoing Royalties,” 76 Missouri Law Review 695–707.
Lemley, Mark A. and Carl Shapiro. 2007. “Patent Holdup and Royalty Stacking,” 85 Texas Law Review
1991–2049.
Lemley, Mark A. and Philip J. Weiser. 2007. “Should Property or Liability Rules Govern Information?,” 85
Texas Law Review 783–841.
Leroux, Isabelle and Frédéric Bourguet. 2006. Litiges de Contrefaçon de Brevets: Une Étude Comparative des
Systèmes Juridictionnels. Paris, France: Institut National de la Propriété Industrielle.
Liang, Mark and Brian Berliner. 2013. “Fee Shifting in Patent Litigation,” 18 Virginia Journal of Law &

Liddicoat, Johnathon E. and Dianne Nicol. 2013. “Re-evaluating False Patent Marking in Australia,” 22 Journal
of Law, Information & Science 128–58.
Lipsitz, Randy, Aaron Frankel and Hanna Seifert. 2014. “Recent Supreme Court Decision Takes Us Back to the
Future: Attorney Fee Award Rate Increases in Patent Cases,” Bloomberg BNA Patent, Trademark & Copyright
Law Daily, Oct. 8, 1489.
Love, Brian J., Christian Helmers, and Markus Eberhardt. 2016. “Patent Litigation in China: Protecting Rights
or the Local Economy?,” 18 Vanderbilt Journal of Entertainment and Technology Law 713–41.
Love, Brian J., Christian Helmers, Fabian Gaessler, and Max Ernicke. 2015. “Patent Assertion Entities in
Europe,” available at http://papers.ssrn.com/abstract_id=2689350.
Malin, Steve and Ari Rafilson. 2009. Presentation Slides, Empirical Analysis of Permanent Injunction
Decisions Following eBay, available at www.ftc.gov/sites/default/files/documents/public_events/evolving-ip-
marketplace/smalin.pdf.
Manzo, Edward D. 2007. “Injunctions in Patent Cases After eBay,” 7 John Marshall Review of Intellectual
Property Law 44–96.
Matsunaka, Masahiko. 2004. “FY 2003 Study Report on the Japanese Economic Structure from a Competition
Policy Perspective—Court Judgments Concerning Calculation of the Amount of Damages in Intellectual
Property Infringement Litigation,” IIP Bulletin 168–77.
Mazzeo, Michael J., Jonathan Hillel, and Samantha Zyontz, 2013a. “Explaining the ‘Unpredictable’: An
Empirical Analysis of U.S. Patent Infringement Awards,” 35 International Review of Law & Economics 58–72.
Mazzeo, Michael J., Jonathan H. Ashtor, and Samantha Zyontz. 2013b. “Do NPEs Matter? Non-Practicing
Entities and Patent Litigation Outcomes,” 9 Journal of Competition Law & Economics 879–904.
McClelland, Scott. 2011. “A Model for Predicting Permanent Injunctions After eBay v. MercExchange,” Bloomberg
Law Reports, available at www.hollandhart.com/files/model_for_predicting_permanent_injunctions_after_eba.
pdf.
McDonagh, Luke and Christian Helmers. 2013. “Patent Litigation in England and Wales and the Issue-Based
Approach to Costs,” 32 Civil Justice Quarterly 369–84.
Merges, Robert P. 1994. “Of Property Rules, Coase, and Intellectual Property,” 94 Columbia Law Review
2655–73.
Miao, Cheng, Wu Yuhe, Li Jiang, Meng-Pu, Li Rongxin, and Wang Zhuojun. 2009. “Theory and Practice
Related to Patent Infringement Damages,” 4 China Patents & Trademarks 12–22.
Moore, Kimberly A. 2000. “Judges, Juries, and Patent Cases—An Empirical Peek Inside the Black Box,” 99
Moore, Kimberly A. 2004. “Empirical Statistics on Willful Patent Infringement,” 14 Federal Circuit Bar Journal
227–40.
Mueller, Janice M. 2013. Patent Law, 4th edition. New York, N.Y.: Aspen Publishers.
Mullane, Andrew, Craig Humphris, and Todd Shand. 2015. “Interlocutory Injunction Applications in
Australia,” 10 Journal of Intellectual Property Law & Practice 170–79.
Nakamura, Nodoka. 2014. “Recent Trends in Court Judgments Concerning Damages in Japanese Patent
Infringement Litigations,” 39 A.I.P.P.I.—Japan 389–410.
Opderbeck, David W. 2009. “Patent Damages Reform and the Shape of Patent Law,” 89 Boston University Law
Review 127–87.
Patstats.org. 2015a. Jury Patent Damages Verdicts (2-1-05 to 12-31-13), available at www.patstats.org/Patstats2.
html.
Patstats.org. 2015b. Post-eBay Permanent Injunction Rulings in Patent Cases to 12-31-13, available at www.
patstats.org/Patstats2.html.
Renaud, Michael, Peter F. Snell, Jack C. Schecter, and Robert J. Moore. 2015. “A ‘Solution’ in Search of
a Problem? The Innovation Act of 2015 and Trends in Fee-shifting in Patent Litigation,” Mintz Levin
Intellectual Property Alert, available at www.mintz.com/newsletter/2015/Advisories/4766-0315-NAT-IP/.
République Française, Ministère du Redressement Productif. 2014. Étude Comparée sur les Dommages et
Intérêts Alloués dans le Cadre des Actions en Contrefaçon en France, au Royaume-Uni et en Allemagne,
available at www.entreprises.gouv.fr/files/files/directions_services/etudes-et-statistiques/etudes/dommages-
interets-dans-actions-en-contrefacon.pdf.
Risch, Michael. 2015. “A Generation of Patent Litigation,” 52 San Diego Law Review 67–132.
Scherer, F. M. 1977. The Economic Effects of Compulsory Patent Licensing. New York, N.Y.: New York University
Monograph Series in Finance and Economics.
Seaman, Christopher B. 2012. “Willful Patent Infringement and Enhanced Damages After In Re Seagate: An
Empirical Study,” 97 Iowa Law Review 417–71.
Seaman, Christopher B. 2015. “Ongoing Royalties in Patent Cases After eBay: An Empirical Assessment and
Proposed Framework,” 23 Texas Intellectual Property Law Journal 203–49.
Seaman, Christopher B. 2016. “Permanent Injunctions in Patent Litigation After eBay: An Empirical Study,”

Remedies 421
Second Subcommittee of the Second Patent Committee. 2014. “Predictability of Monetary Damages under
Article 102(3) of the Japanese Patent Law,” 64 Intellectual Property Management 219–35.
Sepetys, Kristina and Alan Cox. 2009. “Intellectual Property Rights Protection in China: Trends in Litigation
and Economic Damages,” available at www.nera.com/publications/archive/2009/intellectual-property-rights-
protection-in-china-trends-in-liti.html.
Sichelman, Ted. 2014. “Purging Patent Law of ‘Private Law’ Remedies,” 92 Texas Law Review 517–71.
Siebrasse, Norman V. and Thomas F. Cotter. 2015. “The Value of the Standard,” 101 Minnesota Law Review
1159–246.
Siebrasse, Norman V., Alexander J. Stack, A. Scott Davidson, William C. Dovey, Andrew C. Harington, and
Stephen R. Cole. 2008. “Accounting of Profits in Intellectual Property Cases in Canada,” 24 Canadian
Intellectual Property Review 83–136.
Smith, Henry E. 2004. “Property and Property Rules,” 79 NYU Law Review 1719–98.
Smith, Henry E. 2007. “Intellectual Property as Property: Delineating Entitlements in Information,” 116 Yale
Sprigman, Christopher Jon. 2013. “Berne’s Vanishing Ban on Formalities,” 28 Berkeley Technology Law Journal
1565–82.
Sterk, Stewart E. 2008. “Property Rules, Liability Rules, and Uncertainty,” 106 Michigan Law Review 1285–335.
Toyama, Noburo. 1996. “Study with Respect to Proper Civil Remedies for Infringement of Intellectual
Property,” IIP Bulletin 62–76.
Trimble, Marketa. 2009. “Cross-Border Injunctions in U.S. Patent Cases and Their Enforcement Abroad,” 13
Marquette Intellectual Property Law Review 331–69.
Véron, Pierre. 2001. “Le contentieux des brevets d’invention en France: Étude statistique 1990–1999,” available
at www.veron.com/publications/Colloques/Le_contentieux_des_brevets_invention_1990_1999_en_couleurs.
pdf ?card=530.
Véron, Pierre. 2010. “Le contentieux des brevets d’invention en France: Étude statistique 2000–2009,” available
at http://www.veron.com/publications/Colloques/Stats_contentieux_brevets_France_2000-2009_Veron_et_
Associes.pdf.
Véron, Pierre and Olivier Mandel. 2005. “20 ans d’interdiction provisoire de contrefaçon de brevet à travers
l’analyse de 200 decisions,” Propriété Industrielle 14–20.
Vishnubhakat, Saurabh. 2014. “What Patent Attorney Fee Awards Really Look Like,” 63 Duke Law Journal
Online 15–30.
Weinstein, Roy, Janet Thornton, and Paul White. 2010. “New Tools for the Calculation of Infringement
Damages,” 48th Annual Conference on Intellectual Property Law, Center for American and International Law,
Plano, Texas, available at www.micronomics.com/articles/newtoolsforthecalculationofinfringementdamages.
pdf.
Yamaguchi, Kazuhiro. 2016. “Japanese Patent Litigation and Its Related Statistics—Current Environment and
Future Agenda,” A.I.P.P.I.—Japan 128–42.
Yuan, Zhenfu. 2016. “How to Increase Infringement Compensation?,” China IP Magazine, Jan.–Feb., 50–53.

PART IV
TECHNOLOGY-SPECIFIC
STUDIES

18. Patent rights and innovation: evidence from the
semiconductor industry
Rosemarie H. Ziedonis* and Alberto Galasso**
47
Contents
I. Introduction
II. The “Pro-Patent” Shift in the 1980s: Implications for Patent Strategies and the
Organization of R&D
III. Survey Evidence on the Value and Use of Patents
IV. Patent Hold-up and its Resolution
A. The Hold-up Problem
B. Hold-up in the Semiconductor Industry
C. Empirical Evidence of Hold-up
D. Patent Litigation
E. Hold-up in Semiconductor Standard Setting
F. Hold-up Resolution: Licensing and Cross-licensing
V. Conclusion
References
I. INTRODUCTION
Gordon Moore, co-founder of Intel Corporation, famously predicted in 1965 that the
number of computer chips etched onto a silicon wafer would double roughly every two
years. More than 50 years later, the rate of technological progress in the semiconductor
industry remains astonishing. According to Intel estimates, modern transistors are 60 000
times cheaper on a per-unit basis, 3000 times more powerful, and 90 000 times more
energy efficient than they were in 1965 (Weise, 2015). Although “Moore’s Law” appears to
be slowing down and hitting technical limits (Markoff, 2016), few deny that these historic
advances in the design and manufacture of semiconductor devices underpin the modern
information economy.
In the United States alone, the semiconductor industry employs approximately a
quarter of a million people (SIA, 2016). In 2015, U.S. semiconductor companies invested
$34 billion in research and development, representing the highest share of revenue of any
U.S. industry (SIA, 2016). This industry is a fascinating context in which to explore how
** Associate Professor, Strategy and Innovation, Questrom School of Business, Boston
University, and Research Associate, National Bureau for Economic Research.
** Associate Professor, University of Toronto and Research Associate, National Bureau for
Economic Research.
423

patent rights affect the innovative activities of firms. State of the art changes quickly,
with firms racing to move the frontier forward. Semiconductor devices—including the
integrated circuits that power modern smartphones and servers—typically embed thou-
sands of potentially patentable inventions. And the manufacturing process is complex
and costly. These conditions naturally lead to concerns about inadvertent infringement
and patent “hold-up.” Indeed, some credit compulsory licensing restrictions in the 1950s
with the early diffusion of semiconductor know-how and the subsequent development
of a standalone industry of chip design companies and manufacturers (e.g., Levin, 1982;
Grindley and Teece, 1997).
Innovation in the semiconductor industry was also highly cumulative prior to a “pro-
patent” shift in the U.S. legal environment in the 1980s, thus setting a natural stage for
investigating how stronger patent rights affected the incentives of firms in the industry.
According to the USPTO (1995), for example, over 20 000 U.S. patents were issued from
1969–80 on inventions pertaining to semiconductor devices and manufacturing processes.
In contrast, few software or biotechnology-related patents were issued prior to 1980
due to differences in patentable subject matter at the time (Graham and Mowery, 2003;
Merges, 1997), making it difficult to ascertain before-and-after effects in these settings.
This chapter describes empirical findings from studies that examine the relationship
between patent rights and innovative activity in the U.S. semiconductor industry. How
important are patent rights as a stimulus to innovation investment in this sector? What
strategies do firms adopt to navigate the patent landscape and access technologies? To
what extent, if at all, do patent rights deter follow-on innovations?
Even in a single industry setting, answering these questions poses non-trivial measure-
ment and methodological challenges. In early phases of the industry’s development,
pioneering semiconductor companies such as IBM and AT&T engaged in semiconductor
R&D and production primarily for in-house use in their downstream product markets.
Since diversified companies rarely report R&D expenditures at the technology level,
much of the evidence on the relationship between patenting and innovation is based on
specialized firms that primarily compete in semiconductor-related markets. Similarly,
detailed information about the timing of patent license negotiations and the terms of
license agreements is not reported on a systematic basis, leading to a reliance on indirect
litigation-based measures in studies of conflict and settlement among patent owners.
A second, more challenging issue relates to the simultaneous determination of many
economic and innovation outcomes. Suppose, for example, that we are interested in
whether entry of new firms in the semiconductor industry affects the propensity of
incumbent firms to litigate their patents. One may consider addressing this research ques-
tion by comparing patent litigation rates before and after a new competitor enters. But
this is challenging because unobservable factors may drive both market entry and patent
litigation. An increase in market demand may spur entry of new firms into the industry,
for example, and simultaneously increase the incentives of incumbent firms to enforce
their patent rights. In addition to reviewing descriptive findings in the field, we highlight
recent methodological advances that help address these types of challenges.
The chapter is organized as follows. Section II discusses the “pro-patent” shift in the
United States in the 1980s, and its effects on the patenting strategies and organization of
innovative activity within the industry. Section III describes the added insights the role
that patents play in appropriating returns to R&D based on managerial surveys. In Section

Patent rights and innovation 425
IV, we elaborate on the fundamental economics of patent “hold-up” problems, discuss the
challenges of measuring such problems directly, and highlight indirect evidence based on
patterns of legal conflict and license agreements. Although we discuss selected evidence
from the literature on standards-setting organizations (SSOs), we refer readers interested
in multi-party institutional arrangements to the more extensive discussion on that topic
in Chapter 9, “Technical standards, standards-setting organizations, and intellectual
property: a survey of the literature (with an emphasis on empirical approaches)”.
II. THE “PRO-PATENT” SHIFT IN THE 1980s: IMPLICATIONS

FOR PATENT STRATEGIES AND THE ORGANIZATION OF
R&D1
The modern semiconductor industry stems from the invention of the point-contact tran-
sistor at Bell Labs in late 1947 and the subsequent licensing of this invention by Bell Labs
in the 1950s (Holbrook et al., 2000). In their book, Scherer and co-authors (1959) describe
the patent landscape during the early phase of the industry’s development as follows:
During the past two decades a pronounced change has taken place in the policies of govern-
mental bodies towards patents owned by corporations . . . The courts have become increasingly
critical of patent validity, and cases in which the exercise of patent rights conflicted with antitrust
statutes have been prosecuted by denying the exclusiveness of the patent grant. Since 1941,
more than 100 judgments have been entered that required corporations to license their patents
to all applicants at reasonably royalties or no royalties at all. This trend was brought sharply to
the public’s attention in January 1956 when two of the nation’s foremost leaders in industrial
technology, the American Telephone and Telegraph Co and International Business Machines,
Inc., entered into decrees requiring them to license all of their more than 9,000 patents, in most
cases without receiving royalties in return.
From the 1950s through the early 1980s, AT&T and IBM—technological pioneers with
large portfolios of semiconductor-related patents—licensed their inventions widely in
return for access to subsequent inventions by licensees. Levin (1982) and Grindley and
Teece (1997) credit this liberal licensing scheme with the early growth and development of
an industry of firms specializing in semiconductor design and manufacturing. Based on
an analysis of citations to Bell Labs inventions, Watzinger et al. (2017) similarly conclude
that the consent decree stimulated follow-on innovation during this vital stage of the
industry’s development, particularly by young and small firms. While intuitive, determin-
ing whether the stimulus is explained by the consent decree alone is challenging due to the
interplay between antitrust and patent policies. When describing AT&T’s open licensing
of Bell Lab inventions, for example, Tilton (1971) observes:
Certainly the great probability that other firms were going to use the new technology with or
without licenses is another reason for the liberal licensing policy. Secrecy is difficult to maintain
in the semiconductor industry because of the great mobility of scientists and engineers and
their desire to publish. Moreover, semiconductor firms, particularly the new, small ones, have
demonstrated over and over again their disposition to infringe on patents. The prospect of
1
This section draws on material in Ziedonis (2003).

lengthy and costly litigation in which its patents might be overturned could not have been very
attractive.
Von Hippel (1988) makes a similar observation when describing competitive dynamics
within the industry in the early 1980s:
Since patents challenged in court are unlikely to be held valid, the result of high likelihood of
infringement accompanying use of one’s own patented—or unpatented—technology is not
paralysis of the field. Rather, firms in most instances simply ignore the possibility that their
activities might be infringing the patents of others. The result is what Taylor and Silberston’s
interviewees in the electronics components field termed “a jungle” and what one of my inter-
viewees termed a “Mexican standoff ” . . . The usual result is cross-licensing, with a modest fee
possibly being paid by one side or the other.2
As numerous scholars in law and economics have observed, the legal environment in the
United States changed dramatically in the 1980s, both overall and for firms in the U.S.
semiconductor industry (Merges, 1997; Jaffe, 2000; Gallini, 2002). Driven by concerns
about increased international competition and growing belief that stronger intellectual
property rights were needed to stimulate innovation investments, Congress passed a series
of laws that improved the functioning of the U.S. patent system and relaxed antitrust
constraints on firms. The National Cooperative Research Act (1984), for example,
reduced the antitrust penalties for collaboration among firms in “pre-commercial”
research, paving the way for the formation of research consortia such as SEMATECH
in the semiconductor industry (Grindley et al., 1994; Ham et al., 1998). The 1984
Semiconductor Chip Protection Act provided legal protection for the layout of chip
designs, although technological changes soon eroded the value of this sui generis form of
protection (Samuelson and Scotchmer, 2001).
The 1982 formation of the Court of Appeals for the Federal Circuit (CAFC) was
particularly important (Jaffe, 2000; Gallini, 2002).3 Although the driving force of this
centralized appellate court was a unification of U.S. patent doctrine, the CAFC put
in place a number of procedural and substantive rules that collectively favored patent
owners. As Merges (1997) explains, the new court interpreted patent claims more broadly,
increased evidentiary standards to make it more difficult to invalidate the rights of patent
owners, and was more willing to sustain large damage awards, thereby penalizing infring-
ing parties more severely. The plaintiff success rates in infringement cases also increased
substantially during this period (Lerner, 1995).
2
To reduce the risk of disruptions in supply, large customers of chips (e.g., IBM and the U.S.
government) typically required a semiconductor supplier to transfer to a competing firm know-
how and patent rights. These “second source” agreements further promoted cross-licensing in the
industry, but declined in use over the decade of the 1980s as the industry built up capacity and
matured (Grindley and Teece, 1997).
3
Until 1982, patent appeals were primarily heard in the court of appeals of the district in which
the case was tried, leading to “forum shopping” among firms (Jaffe, 2000). Adelman (1987) argues
that the CAFC’s establishment also represented “dissatisfaction with the functioning of both the
Supreme Court and the federal appellate courts” and a “realization by Congress that a uniform
and more reliable patent system was necessary for sustained economic growth and to rise to the
challenge of Japanese and German industrial competition.”

The CAFC also induced a dramatic increase in the use of preliminary injunctive relief
(Merges, 1997). Prior to 1982, such injunctions were only granted when the validity of
the patent was “beyond question”, the infringement was clear, and recovery through ex
post award of damages was not possible (“irreparable harm”). The CAFC adopted a less
stringent standard requiring the patentee to prove only that there was a “reasonable likeli-
hood” of prevailing at trial (Atlas Powder Co. v. Ireco Chemicals, 773 F.2d 1230 (1985)),
and it held that irreparable injury can be presumed (Smith International v. Hughes Tool
Co., 718 F.2d 1573 (1983)). Galasso and Schankerman (2010) document how the effect of
this change is confirmed by data on the use of preliminary injunctions by lower courts.
The proportion of requests for preliminary injunctions that were granted by the district
courts increased from about 32 percent before 1982 to 53 percent after the establishment
of the CAFC.
Not surprisingly, firms with important portfolios of semiconductor patents responded
to the “pro-patent” shift by adopting more aggressive licensing and litigation strategies—
both by seeking licenses from a larger number of firms and by charging higher royalty
rates for rights to use their inventions (Ziedonis, 2003). Soon after the CAFC’s formation,
for example, Texas Instruments (TI) launched an assertive licensing program targeted
against Japanese and Korean competitors in memory chip markets. After litigating its
patents successfully, TI earned almost $2 billion between 1986–93 from licensing rights
to its semiconductor patents and used the funds to re-invest in the development of new
product markets (Grindley and Teece, 1997).
Based on archival evidence and interviews with former IBM executives, Bhaskarabhatla
and Hegde (2014) provide an in-depth case study of the internal management practices
and rewards systems that IBM put in place in the late 1980s to seize opportunities in the
new regime. The authors document that, reversing its open-science norm of publishing
unpatented discoveries in IBM’s Technical Disclosure Bulletins (TDBs), IBM executives
tied employee bonuses more tightly to patent filings, established an internal “patent
factory,” and created a new business development group to increase revenues from out-
licensing IBM know-how and inventions.
Jack Kuehler, IBM’s president between 1989–93, explained the shift in internal practices
to IBM employees as follows:
A series of new laws in the [United States]—plus a much-improved court system for handling
disputes [the CAFC]—are helping patent holders protect their rights better than before . . .
From a simple means of protecting inventions patents have evolved into competitive weapons.
Recent cases before the courts have resulted in multimillion dollar settlements affecting product
line and even corporate profitability in what are now high-stakes battles…Being a world-class
manufacturer and marketer is not enough. You need to own the right to compete. That’s why
IBM is encouraging more patenting of inventions (Bhaskarabhatla and Hegde, 2014).
Bhaskarabhatla and Hegde (2014) further show that in the decade following the adoption
of its pro-patent management practices, IBM increased its successful U.S. patent applica-
tions by 492 percent—from 637 in 1988 to 3777 in 1998. The share of IBM’s technical
disclosures unprotected by patents fell sharply in turn, from 82 percent in 1988 to 15
percent by 1998, the last year that the TDBs were published. The authors find that these
pro-patent practices improved IBM’s financial bottom line, but led to a gradual reduction
in others’ citations to IBM inventions.

Hall and Ziedonis (2001) investigate the effects of the pro-patent shift on 95 U.S. semi-
conductor firms using a combination of quantitative and qualitative research methods.4
Consistent with the case study of IBM by Bhaskarabhatla and Hegde (2014), Hall and
Ziedonis (2001) find that semiconductor manufacturers started “ramping up” their patent
portfolios in the mid-to-late 1980s. As one interviewee noted, there were “a lot of patent-
able inventions around,” but the firm had “not taken the time and incurred the cost” to
patent those inventions in the past (Hall and Ziedonis, 2001).
Managers interviewed by Hall and Ziedonis (2001) suggested that the upsurge in
patenting was a strategic response to the more favorable judicial treatment of U.S. patents.
Although the CAFC was established in 1982, several interviewees commented that CEOs
at their companies first “woke up” to the new power of patents in 1986, when Kodak was
forced to halt production of instant cameras and pay almost $1 billion in damages to
Polaroid due to a patent infringement lawsuit. As a licensing director from a semiconduc-
tor manufacturer noted, the threat of injunction is a powerful lever when it is costly to halt
production and difficult to invent around inventions. Others noted the cascading effects
of TI’s successful litigation strategy and the more assertive stance of industry leaders (e.g.,
AT&T, IBM, and Motorola) in earning licensing revenues from their inventions.
Overall, the evidence suggests that patents became more valuable for use as bargaining
chips in negotiations with other patent owners. As Hall and Ziedonis (2001) conclude:
“a firm lacking a strong patent portfolio of its own with which to negotiate licensing or
cross-licensing agreements could face a more rapid erosion of profits in an era when the
costs and risks associated with infringement had increased.” Consistent with this view,
they document a dramatic upsurge in patenting between 1979–95 that outpaced R&D
spending in the industry and was unexplained by other factors. Hall (2005) reports a
similar upsurge in the propensities of firms to patent in other segments of the information
computing and technology sector after the CAFC’s formation.
In addition to triggering “patent portfolio racing” by capital-intensive firms, Hall and
Ziedonis (2001) find that the pro-patent shift had a distinctive—and separable—effect
on firms that specialize in the design of semiconductor devices, but contract out the
manufacture of their products to others. Managers from these semiconductor design (or
“fabless”) firms emphasized the importance of obtaining strong “bulletproof ” patents
to protect proprietary technologies against competitors in niche product markets and to
attract venture capitalist funding. Indeed, when asked to consider a hypothetical abolish-
ment of the patent system, representatives from both design firms and manufacturers in
the Hall and Ziedonis (2001) study voiced concerns about potential chilling effects on
entry by innovative design firms that relied on external sources of financing.
These latter insights are consistent with theoretical models linking stronger patent
rights to increased specialization in the organization of R&D (e.g., Arora, Fosfuri, and
Gambardella, 2001). In line with that view, Hall and Ziedonis (2001) provide quantitative
4
For the quantitative analysis, it was important to track annual R&D investments related to
semiconductors. The sample in Hall and Ziedonis (2001) therefore is based on publicly traded U.S.
firms whose primary business is semiconductors and related devices. The approach unfortunately
excludes from the estimation sample large US “systems” manufacturers such as IBM, AT&T, and
Motorola, and non-US firms (e.g., Toshiba, Samsung, and Siemens) that are important owners and
users of semiconductor-related patents.

evidence that the rate of entry by semiconductor design firms increased sharply following
the pro-patent shift. They were unable, however, to specifically pinpoint the increased
entry to the strengthened patent regime due to simultaneous advances in technological
platforms that facilitated vertical disintegration within the industry. Put differently, the
counterfactual world— that of entry rates and organization forms in the industry had the
legal environment remained unchanged—was unobservable.
In summary, there is considerable evidence that the shift in the U.S. legal landscape in
the 1980s significantly affected incentives to patent within the semiconductor industry,
and that specialized design firms relied heavily on strong patents both to secure financ-
ing from venture capitalists and to safeguard proprietary technologies in niche product
markets. Even within one industry, the effects of the pro-patent shift were multifaceted
and wide-ranging.
III. SURVEY EVIDENCE ON THE VALUE AND USE OF

PATENTS
Additional insights about the value and use of patents in the U.S. semiconductor industry
can be gleaned from surveys of practicing managers. Such surveys provide insights about
managerial perceptions, and often allow for useful cross-industry comparisons. The main
drawback is that most surveys seek to obtain representative views from many industries,
which can limit intra-industry coverage. Large-scale surveys can also be costly to admin-
ister, making it difficult to track changes in viewpoints and practices over time.
Of particular importance, the 1983 “Yale” and 1994 “Carnegie Mellon” Appropriability
Surveys (Levin et al., 1987; Cohen, Nelson, and Walsh, 2000) asked managers of R&D
labs in U.S. manufacturing firms to rate the relative effectiveness of patents in appropriat-
ing the returns to R&D investments. Despite the fact that the Carnegie Mellon Survey was
administered in 1994, well into the pro-patent regime, the findings for the semiconductor
industry were consistent with those reported in the earlier Yale Survey: R&D lab manag-
ers ranked patents among the least effective mechanisms for profiting from innovation
overall and relative to other industries. Instead, semiconductor R&D lab managers
emphasized the importance of lead time, secrecy, and superior manufacturing and/or
design capabilities as means for recouping returns to R&D investments.
Importantly, the Carnegie Mellon Survey asked follow-up questions about the motives
for patenting. Similar to interview evidence in Grindley and Teece (1997), Hall and
Ziedonis (2001), and Bhaskarabhatla and Hegde (2014), R&D lab managers from semi-
conductor firms emphasized the strategic value of patenting for both defensive (reducing
litigation risks) and offensive (gaining superior access to external rights and know-how)
reasons. As Cohen et al. (2000) report, similar views were espoused in other “complex
products” industries where the value of a single patent is inherently tied to that of other
patented and unpatented technologies.
Contributing more systematic evidence from entrepreneurial firms, the 2008 Berkeley
Patent Survey asked CEOs of technology-oriented startups a series of questions about
patents (Graham, Merges, Samuelson, and Sichelman, 2009). The authors received
responses from CEOs of 1332 companies in the biotechnology, medical devices, computer
software, and information technology (IT) hardware industries. Graham et al. (2009)

subsume semiconductor startups in the broader “IT hardware” industry, which also
includes communications and computer hardware companies. Overall, CEOs from
health-related and IT hardware companies ranked patents (along with first-mover advan-
tage) to be among the most important means for “capturing competitive advantage”.
This emphasis on the importance of patents as means for profiting from innovation for
IT hardware start-ups contrasts sharply with the more lackluster views expressed by larger
IT companies represented in the earlier Yale and Carnegie Mellon Surveys.
The Berkeley Patent Survey also asked the CEOs of start-ups to rank on a 1 (“not at
all important”) to 4 (“very important”) scale which factors drove the decision to seek U.S.
patent protection (Graham et al., 2009). The top-ranked reason (at 3.59 on average) was
to prevent others from unauthorized use of the invention, which is not surprising. The
next-highest-ranked reasons were arguably finance related, including “to improve chances
of securing investment” (at 3.30), “to improve chances/quality of liquidity” (3.24), and
“to enhance the company’s reputation” (3.13). These findings call into question whether
patents serve a meaningful signaling function in the market for entrepreneurial financing,
as suggested by Long (2002) and tested in Hsu and Ziedonis (2013).
In combination, this survey evidence suggests that patent rights are important for the
financing and development of early-staged companies, including semiconductor start-
ups. The survey findings in Graham et al. (2009) suggest, however, that software start-ups
may be an exception to this more general rule.
IV. PATENT HOLD-UP AND ITS RESOLUTION
In policy debates, the semiconductor industry is often heralded as a context rife with
patent “hold-up” problems (U.S. Federal Trade Commission (FTC), 2003, 2011). Given
the ongoing controversy surrounding this topic, we elaborate on the fundamental patent
“hold-up” problem below and discuss the challenges of measuring such problems directly.
A. The Hold-up Problem
In the economics literature, “hold-up” refers to a situation where one party is able to
expropriate rents from another, typically due to investments specific to a relationship.
Hold-up can generate economic inefficiencies that both courts and contracting parties
often try to avoid. “Bad” behavior (e.g., deception) is not required to generate losses of
economic surplus, but deception and concealment may magnify the impact of hold-up
(Farrell et al., 2007).
The basic idea is that simple market contracts insufficiently safeguard against expro-
priation when assets cannot be redeployed to the next best use (or user) without signifi-
cant loss of value (Klein et al., 1978). Asset-specific investments are pervasive. A famous
example is the supply of automobile body parts by Fisher to General Motors at the
beginning of the 20th century. Fisher Body made investments in machines highly specific
to General Motors and of little use to other car manufacturers. These specific investments
allowed General Motors to hold-up Fisher by threatening to change suppliers unless
Fisher reduced its prices (Klein, 1998). A natural way to minimize the hold-up problem is
to design contracts prior to making the investment. Even then, “ex-ante” contracts are an

imperfect solution given the difficult and cost of covering every contingency. Moreover,
some elements of performance cannot be measured or described unambiguously (e.g., the
taste of a soft drink or the effort a contractor puts toward a task).
A large body of literature in “transactions cost economics” shows how this imperfect
nature of contracts induces agents to underinvest in areas with high risks of expropriation
(Williamson, 1985). In these environments, there is also a strong incentive to internalize
transactions involving highly specific assets by performing economic activities within a
firm instead of relying on the market. In other words, severe risk of hold-up may lead to
integration and mergers among firms.
Patent licensing negotiations are a natural setting where the hold-up problem may arise.
A valid patent grants a patentee the right to exclude others from using the patented inven-
tion. This exclusionary right thus allows the patentee to extract rents through licensing
deals from firms using the patented technology. The magnitude of these rents depends on
the bargaining power of the patentee, which in turn depends on the timing of investments
and the feasibility and costs of ex ante contracting.
The issue of patent hold-up has recently appeared in prominent public debates on
patent policy in the U.S. and Europe (National Research Council, 2004; U.S. FTC, 2011;
European Commission, 2011). There is substantial agreement among academic scholars
and policy makers that the patent system can be a powerful policy tool to increase research
and development incentives and to promote follow-on innovation. However, there are
growing concerns that patent rights may also be an impediment, rather than an incentive,
to innovation if the increasing proliferation of patents and the fragmentation of owner-
ship rights among ﬁrms have sufficiently raised transaction costs and exposed ﬁrms to ex
post hold-up through patent litigation (Heller and Eisenberg, 1998).
B. Hold-up in the Semiconductor Industry
The semiconductor vertical chain includes a large number of production steps. While
a variety of companies (e.g., Intel) complete the entire process in-house, other industry
players frequently outsource part of their production activities (Turley, 2005). At the
extremes of the vertical chain spectrum, the “fabless” companies discussed earlier special-
ize in chip design and “pure-play” foundries manufacture chips without designing them.
While most of the semiconductor companies position themselves at intermediate levels
of vertical specialization, the industry has moved toward organizational separation of
semiconductor design from chip manufacturing (Monteverde, 1995). There are also cases
of firms that specialize in the acquisition and licensing of IP rights, such as Rambus and
ARM. Other companies such as Qualcomm and Xilinx combine licensing with specialized
component sales (Simcoe et al., 2009; Serrano, 2010; Galasso et al., 2013).
A key feature of the manufacture of semiconductor devices is the use of “clean rooms,”
where airborne particles are minimized and temperature, humidity, and pressure are
strictly monitored. Manufacturing equipment consists of two broad kinds: front-end
equipment—installed in clean rooms and used to produce silicon wafers and semicon-
ductor chips—and back-end equipment, used to assemble, package, and test the devices
(U.S. International Trade Commission (USITC), 2006). Typically, front-end equipment is
installed and arranged to accommodate the production of a specific semiconductor device.
Because minor changes in clean room operations may impact the sensitive equipment and

result in product damage, reconfiguring front-end equipment to accommodate changes in

product manufacturing is very expensive, to the extent that construction of a new fabrica-
tion facility is often a cheaper alternative (Turley, 2005; Theron et al., 1999).
Fabrication facilities not only are expensive, but also have a short lifespan. By 2006,
for example, a typical new manufacturing facility had an economic life of about three
years and cost $3 billion (SIA, 2006). In addition, there is a steep learning curve associ-
ated with the production of semiconductor devices (Macher and Mowery, 2003). Most
of the chips produced during the first weeks of operation of a new fabrication facility
are usually damaged. Over time the quality of the production improves, but it may take
more than 12 months of non-stop operations to have production yields above 90 percent
(Turley, 2005).
In short, clean rooms and front-end manufacturing equipment are difficult to redeploy
and expensive, and involve steep learning curves. These characteristics of the semiconduc-
tor industry imply that halting or altering production processes will be more costly for
firms that own and operate manufacturing facilities compared to firms that do not own
such facilities. In turn, this explains why patent litigation and preliminary injunctions are
viewed as more costly by firms with large investments in fabrication facilities. This idea
is supported by the findings of Hall and Ziedonis (2001), Ziedonis (2004), and Galasso
(2012) that document how capital-intensive semiconductor firms tend to react to the risk
of patent litigation by amassing large patent portfolios to improve their ex post bargaining
position in the event of settlement negotiation.
These features of the semiconductor manufacturing environment exacerbate the risk
of hold-up for firms working in the industry. To see this, consider—as in Ziedonis (2004)
—the problem of a semiconductor firm. Suppose the manufacturer could easily invent
around a patent at the initial stages of the production process (e.g., while designing new
chips or when specifying the layout of new fabrication facilities). In this case, the licens-
ing fees the patentee could obtain from the firm would be small, ex ante, because of the
manufacturer’s ability to invent around the patent (Levin et al., 1987; Teece, 1986). The
bargaining power of the manufacturer would be far weaker, however, if it learns about
the patent after embedding the technology in designs or production processes that are
expensive to redeploy. At this stage, these investments specific to the patented technology
allow the patentee to extract more rents from the firm since the cost to invent around the
patent is much larger at this point.
To illustrate more formally how the hold-up problem can affect licensing negotiations
for semiconductor firms, consider the following stylized model. There is one firm with
production facilities generating product market profits equal to V. A patentee holds a
patent that the firm is allegedly infringing. In the absence of a license agreement, the
parties resolve the dispute in court, each incurring a litigation cost of L. We assume that
the court will find the patent infringed with probability r. If the patent is found infringed,
the infringing firm sustains a loss equal to K to halt the production and design around the
infringed patent. Let us indicate the profits of the firm in case of litigation as:
pLit = V(1 − r) − rK − L = V − r(K + V) − L.
Now consider the licensing option. If we indicate with f the (fixed) licensing fee requested
by the patentee, the profits of the licensee will be:

pLic = V − f.
From the above formulas it follows that the maximum license fee that will be accepted by
the licensee—the fee that renders the licensee indifferent between accepting or rejecting
the offer—is equal to:
f = r(K + V) + L.(1)
Equation (1) highlights a variety of features that determine the outcomes of licensing
negotiations in the semiconductor industry. First, it shows that the rents that can be
extracted by patentees are larger, the larger the cost of litigation of the alleged infringer,
L. Intuitively, this suggests that firms that have a comparative advantage in litigation (e.g.
because of “deeper pockets” or access to superior legal counsel) will be able to access tech-
nologies at better terms. One may expect small firms to have a larger value for L. Lerner
(1995) provides evidence that small firms avoid investing in technology fields where the
threat of litigation from large firms is high. Lanjouw and Lerner (2001) show that the use
of preliminary injunctions by large firms can discourage small firm innovation. Simcoe et
al. (2009) illustrate how small private firms are particularly litigious after disclosing their
patents to SSOs. Finally, Lanjouw and Schankerman (2004) show that the probability of
being involved in a suit is higher for patents owned by small firms than for those owned
by large firms.
A second intuitive insight from equation (1) is that the licensing fee depends on the
strength of the patent, r. Such parameter implies that a strengthening of patent protec-
tion, such as that experienced in the 1980s with the establishment of the CAFC, is likely
to have increased the cost of accessing patented technology and the rent extracted by
patentees. Third, the fee is larger for technologies generating greater profits in the product
market, V.
Finally, equation (1) shows that the licensing fee is likely to be larger the greater the cost
of halting production, K. This idea is consistent with the interviews summarized in Hall
and Ziedonis (2001) and Ziedonis (2004): losses incurred when halting or altering produc-
tion processes are more detrimental for firms investing intensively in product-specific
manufacturing facilities. Notice that when investments in clean rooms and front-end
manufacturing equipment are substantial (i.e. K is very large), the patentees can extract a
large rent holding-up the alleged infringer even when the patent protecting the technology
is weak (i.e. r is small).
A related stream of transactions-costs research emphasizes that hold-up problems
may also surface due to fragmented ownership rights (Heller and Eisenberg, 1998).
This literature suggests that granting too many exclusionary rights to too many parties
can reduce the use of economic resources. In the patent context, this view implies that
bargaining failure can arise when a technology user requires licenses from numerous
disparate patentees. In this case, uncoordinated negotiations among the parties can
generate “royalty stacking” that reduces the licensee’s profit and, in extreme cases, can
prevent downstream development (Heller and Eisenberg, 1998; Lemley and Shapiro,
2006; Galasso and Schankerman, 2010).
To clarify the point, consider a hypothetical example of a semiconductor firm that
is considering an investment of $10 billion in a new fabrication facility. Assume the

manufacturer identifies 1500 patents potentially infringed in the design or manufacture

of its products and is still unsure of the effective scope and validity of those patents. Will
the investment in fabrication facility take place?
For simplicity, consider two extreme scenarios. In scenario 1, one firm owns all the
patents. In scenario 2, the patents are assigned to 1500 different patentees. If patent
negotiations are costless, as in a Coasian setting with zero transactions costs, ownership
of patent rights will not matter and the fabrication facility investment will take place as
long as it is efficiency-enhancing (Coase, 1960). But once we assume non-trivial transac-
tions costs, there are substantial differences in the bargaining environment across the two
scenarios that may influence the investment decision of the firm. As explained in Ziedonis
(2004), in scenario 2 the costs and potential delays involved in bargaining sequentially with
a large number of fragmented rights holders may render patent negotiations infeasible for
the manufacturer. In other words, the costs and potential delays associated with patent
negotiations depend on the concentration of ownership of rights and the “industrial
organization” of the technology field (Galasso and Schankerman, 2015a, 2015b).
C. Empirical Evidence of Hold-up
Despite the substantial theoretical work developing the idea of patent hold-up and char-
acterizing settings where the problem is likely to be more pervasive, an empirical analysis
of patent hold-up is inherently challenging. Ideally, the researcher would like to observe
the details of patent licensing negotiations, including the ex ante alternatives, the rents
extracted by the patentee, and the specific investments made by licensees/infringers. These
data are typically unavailable, since both the terms of patent licensing agreements and the
details of negotiations leading to a contract are kept confidential.5
Moreover, the fear of patent hold-up may lead firms to vertically integrate, avoid
innovation investments, or design institutions such as patent pools or fair, reasonable,
and non-discriminatory (FRAND) commitments to mitigate the problem. These strategic
responses in turn affect the licensing data available to researchers.
D. Patent Litigation
The challenges involved in collecting direct evidence of patent hold-up lead the law
and economics literature to study indirect evidence of the problem. The most common
approach has been to collect data on patent disputes and their resolution. Litigation and
delays in the settlement process are typically associated with hold-up and high transaction
costs for the negotiating parties.
Ziedonis (2003) provides the first comprehensive analysis of patent cases filed in U.S.
District Courts and the USITC from January 1, 1973, through June 30, 2001, that involve
136 dedicated U.S. semiconductor firms as defendant or plaintiff. Sample firms include
the universe of publicly traded U.S. firms during 1973–2000 that either listed semiconduc-
tors and related devices (SIC3674) as their primary line of business or were identified by
5
Anand and Khanna (2000) discuss the scarcity of licensing data and contrast it with data
availability in other areas of economics.

industry sources as dedicated U.S. semiconductor firms. In 2000, sample firms collectively
generated over $88 billion in revenues, spent $12 billion in R&D, and were awarded
roughly 31 000 U.S. patents. The sample does not include non-U.S. firms (e.g., Samsung
or Siemens) and large U.S. “systems” manufacturers (e.g., IBM or AT&T) because it is not
possible to identify R&D investments targeted to semiconductor technologies for these
large diversified companies.
The empirical analysis follows this sample of semiconductor firms over time and
focuses on the patent acquisition and enforcement histories at the level of individual
firms. This approach allows examining changes in the litigation propensity of firms
over time. The sample includes both semiconductor “manufacturers” (i.e., firms such as
Intel, Texas Instruments, and Micron Technologies, which design and manufacture the
majority of their products in-house) and “design” firms (i.e., firms such as Altera, Xilinx,
and SonicBlue). Even though most of the design firms in the sample commercialize and
sell products of their own, they are typically much smaller in size (in terms of number
of employees and sales revenues) than manufacturing companies in the sample and they
invest more heavily in R&D.
Several interesting, albeit descriptive, trends emerge from this study. First, the analysis
shows that roughly 56 percent of the sample firms were involved in at least one reported
patent case filed in U.S. District Courts and the USITC between January 1, 1973, and
June 30, 2001. On average, semiconductor firms involved in patent cases tend to be larger
(in terms of sales or number of employees), invest more in R&D (in absolute terms and
per employee), and own larger patent portfolios than semiconductor firms not involved
in patent litigation during the sample period.
Second, the data show a sharp increase in the number of annual cases filed involving
semiconductor firms around the mid-1980s that continued throughout the 1990s. This
trend suggests that legal disputes over intellectual property became more common in
semiconductors—despite the widespread use of cross-licenses in this industry. More
importantly, while the litigation rate per R&D dollar increased dramatically in the
semiconductor industry during 1986–2000 from that in the preceding decade (by as much
as 93 percent), the number of cases filed per 1000 patents slightly declined between the
two periods.
This asymmetric effect is driven by the dramatic rise in patenting by semiconductor
firms since the mid-1980s because of the “patent portfolio races” of U.S. semiconductor
manufacturers documented in Hall and Ziedonis (2001). Figure 18.1 (reproduced from
Figure 3 in Ziedonis 2003) provides evidence of this impressive growth in patenting.
Finally, the paper shows that the average litigation rate of specialized design firms in
the sample is high and is more than twice that of manufacturers in the sample. On aver-
age, semiconductor manufacturers litigate with a more diverse set of parties and enforce
patents that are almost four years older than the average patent in their portfolios. In
contrast, design firms typically litigate against other design firms and enforce patents that
are roughly the same age as the average patent in their portfolios.
Somaya (2003) studies the determinants of patent litigation and settlement in a sample
including firms from two distinct industries: research medicine (which includes biotech-
nology, drug delivery systems, assays, and dental innovations) and technology research
(which includes semiconductors, data storage, computer systems, I/O devices, computer
applications, and networking technologies). His analysis provides evidence of substantial

100
6000 # Patents Awarded
90
# Patent Cases
80
5000
Number Patent Awards
70
Number Cases
4000 60
50
3000
40
2000 30
20
1000
10
0 0
1975 1980 1985 1990 1995 2000
Year
Note: This figure is reproduced with permission from Figure 3 in Ziedonis (2003).
Figure 18.1 Semiconductor patent awards v. case filings, 1973–2001 (Ziedonis, 2003)
differences between litigation in the technology sector and medical research in the period
1983–93. The paper shows that the litigation propensity is consistently lower for research
medicine patents compared to technology patents. In the technology sector, patent
cases are more likely to include more than one patent, and countersuits are much more
frequent. Finally, litigation involving individual inventors is more likely in the technology
sector, whereas litigation by universities is less likely.
Taken together, the results in Somaya (2003) are consistent with the idea that patents
have different uses across industries. The evidence for the semiconductor field, in which
the propensity of countersuits is very high, suggests that mutual hold-up between firms
is quite common in this sector.
Galasso and Schankerman (2015a) exploit patent litigation at the CAFC to quantify
the extent to which patent rights impede the cumulative innovation process, and to iden-
tify whether their impact differs across technology fields and firms. There are two reasons
why it is challenging to study empirically the impact of patent protection on innovation
incentives. The first problem is that comparable technologies with and without patent
protection need to be identified. The second issue is that it is hard to measure innovation
activity related to a specific patent. To address these problems, Galasso and Schankerman
(2015a) exploit patent invalidation cases litigated at the CAFC. Their dataset comprises
1357 CAFC decisions from 1983–2008, with information on whether each patent was
invalidated. About 40 percent of the decisions in their sample are associated with loss of
patent protection for the technology.
To estimate the effect of patent rights on follow-on innovation, Galasso and
Schankerman (2015a) look at citations received by patents in a five-year window fol-
lowing the CAFC decision and compare those invalidated with those which are upheld.
Measuring cumulative innovation with citations by later patents is a common practice in
the economics of innovation literature where a large number of empirical studies exploit
citations as a way to trace knowledge spillovers (Griliches, 1992). It is important to note

that citations can either under- or overestimate the extent of follow-on innovation. This
happens when inventors develop improvements that are not patented (or patentable), or
when a citing inventor does not actually build on a cited patent. While using product level
information is clearly desirable, citations are often the only practical measure for studies
that cover a wide range of technology fields.
A fundamental challenge with litigation data is that invalidated patents may differ
from not-invalidated patents and that such differences may also affect patent citations.
For example, a positive shock to the commercial value of the underlying technology may
increase citations to a patent and, at the same time, induce the patentee to invest heavily
in the case to avoid invalidation. It is crucial to address this “endogeneity” issue in order
to test whether the impact of patent protection on cumulative innovation is causal.
The empirical strategy followed by Galasso and Schankerman (2015a) exploits the fact
that judges are assigned to patent cases through a computer program which randomly
generates three-judge panels. Essentially, their empirical methodology is equivalent to
comparing citations received after CAFC decisions by patents that are invalidated because
they were randomly assigned to judge panels with high propensity to invalidate with
citations received by similar patents that are not invalidated because they were randomly
assigned to judge panels with low propensity to invalidate. In conducting this exercise,
they also control for a number of confounding factors such as the age of the patent, the
technology field, and the number of citations received before the CAFC decision.
Galasso and Schankerman (2015a) find that patent invalidation is followed by a 50
percent increase in subsequent citations to the litigated patent, on average. This evidence
suggests that, on average, patents block follow-on innovation. More importantly, they also
show that the impact of patent invalidation differs substantially across broad technology
areas. Their empirical analysis shows that patent invalidation has a large and statistically
significant impact on cumulative innovation in the fields of semiconductors, computers
and communications, electronics, and medical instruments. However, they find only a
small and statistically insignificant effect in the chemical, pharmaceutical, and mechanical
technology fields. Moreover, they find that the impact of invalidation is predominantly
driven by the invalidation of patents owned by large firms, which increases the number of
small innovators subsequently citing the focal patent. This suggests that invalidation of
a particular patent is unlikely to affect strategic interaction between large semiconductor
firms, but it may affect innovative investment by small entrepreneurial firms.
In a companion study, Galasso and Schankerman (2015b) study the impact of patent
invalidation on subsequent innovation and exit by the patent holder. They show that
patent invalidation leads to a 50 percent decrease in patenting by the patent holder, on
average, but the effect is entirely driven by small innovative firms in technology fields
where they face many large incumbents. In addition, the loss of patent rights significantly
increases the likelihood of exit for small firms.
The findings in Galasso and Schankerman (2015a, 2015b) help in understanding the
role of patents in the semiconductor industry. They support the idea that semiconduc-
tor firms face patent hold-up and that the economic trade-offs generated by patents in
this industry are different from those faced by firms in other innovation environments.
Moreover, these two studies show that patent rights affect innovation by small and large
firms very differently.

E. Hold-up in Semiconductor Standard Setting
In a variety of industries compatibility standards outline key technical criteria for shared
technology platforms. Compatibility standards define key formats and interfaces for
shared technology platforms (Simcoe, 2012). By coordinating firms toward technological
specifications, standards generate a variety of economic benefits (Shapiro, 2001). First,
by reducing incompatibility problems, there is an increase in users’ willingness to pay for
a technology. Second, by attracting a greater number of users, compatibility standards
generate incentives to entry and innovate in the technology space.
SSOs are forums where interested companies seek consensus and endorse specific
technologies to develop industry standards (Simcoe, 2012). SSOs play an important
role in semiconductors where compatibility and interoperability induce coordination in
choice of components, packaging, test methods, and materials. The Joint Electron Device
Engineering Council (JEDEC), founded in 1958, is the most prominent organization
developing standards for semiconductor devices.
In principle, the risk of hold-up can be substantial when SSOs include patented
technologies in a standard. Before an industry standard is chosen, the bargaining
power of a patentee is quite weak because SSOs compare alternative technologies that
have the potential to become a standard. After the adoption of a standard, alternative
technologies become less attractive and patents covering the standard become much
more valuable. The change in patent value may induce members of standard committees
to withhold information about their patent applications when they vote for the inclusion
of technologies in a standard. The magnitude of these inefficiencies is at the center
of a recent growing literature on patent hold-up in standard settings. Research on the
topic has been primarily theoretical and provides contrasting views on the severity of
the problem. At the heart of these divergent opinions, there is the inherent difficulty of
conducting empirical research on patent hold-up and the confidentiality of licensing
deals.
Epstein, Kieff, and Spulber (2012) argue that private solutions implemented by SSOs
(e.g., licenses, reputation through repeat play, and RAND commitments) are likely to curb
substantially the severity of patent hold-up problems. Intuitively, because the objective of
SSOs is to maximize the market success of commercial standards, they are likely to take
actions to limit potential threats of hold-up. For example, the Institute of Electrical and
Electronics Engineers Standards has been very proactive by clarifying and strengthening
the FRAND licensing commitments it requires from participants.
Other scholars have come to different conclusions. Among others, Farrell et al. (2007),
Lemley and Shapiro (2013), and Lerner and Tirole (2014) propose regulatory actions,
arguing that market mechanisms are not sufficient to alleviate the negative impact
of patent hold-up on innovation incentives. These papers are motivated by anecdotal
evidence from companies such as Wang Laboratories and Rambus that have been found
guilty of deceptive actions as members of JEDEC. In Wang Labs, Inc. v. Mitsubishi, 103
F.3d 1571 (1997), the CAFC asserted that Wang, by failing to disclose to JEDEC its
patent applications, created an implied license from Wang to Mitsubishi to practice the
invention. In Rambus v. FTC, Docket No. 9302, Opinion of the Commission 3 (2006),
the FTC found Rambus guilty of monopolization for not revealing its patent applications
while a member of JEDEC.

These studies show that strategic deception by patent holders not only impacts the
licensing process, but can also generate antitrust concerns and harm competition. In
discussing remedies to restore competition and compensate injured parties, these papers
emphasize that the focus should be the increment to market power compared to the
competitive environment that would have appeared with an open and well-informed
technology competition. This line of research spurred various actions taken by courts and
policy makers to mitigate the concern of patent hold-up in complex industries (Shapiro,
2016). First, in a 2006 landmark case, the Supreme Court greatly reduced the threat of
patent hold-up by limiting the availability of injunctions to patent holders (eBay Inc. v.
MercExchange, L.L.C., 547 U.S. 388 (2006)). Second, in Ericsson v. D-Link, 773 F.3d.
1201 (2014), the CAFC clarified that the determination of “reasonable royalties” for pat-
ents essential to a standard should reflect the ex-ante incremental value of the technology
(i.e., the value prior to its inclusion in the standard).
F. Hold-up Resolution: Licensing and Cross-licensing
The high risk of unintentional infringement in the semiconductor industry induces firms
to develop IP strategies aimed at mitigating the risk of being held-up and halting produc-
tion. One way to reduce hold-up risk is to enter a license agreement ex ante, or before
investments in new manufacturing facilities or product designs take place.
Siebert and von Graevenitz (2010) provide empirical evidence of ex ante licensing
strategies in the semiconductor industry. The authors compile a dataset of licensing
contracts involving semiconductor companies signed between 1989–99. They obtain this
information from Thompson Financial and other publicly available sources such as busi-
ness reports, filings published in the National Cooperative Research Act, and announce-
ments made in the public press. The data show that licensing remains widespread in the
semiconductor industry, with 847 transactions reported between 1989–99.
After examining the details of each licensing contract, the authors manually classify
each deal in an ex ante or ex post contract. Their data show that ex ante agreements
are much more common than ex post deals, with 549 ex ante licenses versus 298 ex
post licenses during the 1989–99 period. Moreover, while the frequency of ex post deals
appears constant during the sample period, ex ante contracts follow an inverse-U shape
over time with a peak in the mid-1990s. Siebert and von Graevenitz (2010) also find that ex
ante licensing deals between two firms are more likely when firms have high technological
similarity (measured by exploiting cross-citations of patents of the two firms).
A special case of ex ante deals are cross-license contracts, which are bilateral agreements
in which two firms choose not to enforce intellectual property rights against each other.
Galasso (2012) studies broad cross-licensing deals that are agreements covering the entire
patent portfolios or patents in some extensive technology class. Broad cross-licensing is a
common IP strategy in industries such as computers and semiconductors, where products
combine many patentable technologies and where it is easy to unintentionally infringe a
patent (Grindley and Teece, 1997; Shapiro, 2001). Press releases and companies’ annual
reports clearly reveal that broad cross-license deals are indeed widespread in the industry.
Galasso (2012) shows that, of the ten semiconductor companies with the largest R&D
expenditure in the period 1990–95 (carrying out more than 60 percent of the R&D expendi-
ture of the industry), eight entered into at least one cross-licensing deal between 1990–2000.

To illustrate the economic incentives that lead to broad cross-licensing, Galasso (2012)
develops a game theoretical model in which two firms are involved in a series of infringe-
ment disputes. In the absence of a broad cross-license agreement, these disputes are
litigated and the infringing firm stops producing if the court finds it liable. The cost that a
firm sustains when halting production increases with its capital intensity. Firms negotiate
cross-licensing contracts through a bargaining procedure in which they learn the value of
each other’s patent portfolios.
The Galasso (2012) model shows that both the decision to cross-license and the timing
of the agreement depend crucially on firms’ capital intensities. Specifically, it shows that
two firms will sign a cross-license agreement only if their capital intensities are large
enough. The intuition for this result is the following. A broad cross-license agreement is
costly to a firm because it involves sharing patented technologies with a rival firm. On the
other hand, a firm benefits from such a cross-license because it avoids the loss associated
with discontinuing production. Because this benefit increases with firms’ capital intensi-
ties, a cross-license agreement is profitable only for firms with high capital intensities.
In addition, the model predicts that broad cross-license negotiations will have shorter
duration for firms with high capital intensities and for firms facing a low frequency of
infringements. This occurs because firms have an incentive to delay the agreement in order
to obtain additional information on the value of the rival’s patent portfolio. However,
waiting is costly because it involves litigation that is particularly detrimental for firms with
high capital intensities. Finally, the model predicts greater broad cross-license agreements
between firms with complementary technologies in their patent portfolios.
Galasso (2012) tests the predictions of the model using a unique dataset that combines
information on broad cross-license agreements and patent litigation in the semiconductor
industry. Specifically, the sample includes 218 publicly traded U.S. firms whose principal
line of business is semiconductors and related devices (SIC 3674), for which Compustat
has data for at least three years between 1985–2005. Broad cross-license contracts were
identified from the SEC annual filings (10-K) and company press releases.6
The empirical findings are consistent with the theoretical model and can be summarized
as follows. First, high capital intensity increases the likelihood that firms will sign broad
cross-license agreements and decreases the duration of licensing negotiations. Second,
broad cross-license agreements take longer to negotiate when firms patent in similar
technology areas. Finally, the analysis shows that firms are more likely to enter broad
cross-license agreements when their patent portfolios are complementary, as indicated by
a high frequency of cross-citations between the patents of the two firms.
Harhoff, von Graevenitz, and Wagner (2015) study challenges to patent validity as a
strategy to mitigate the risk of hold-up. Their analysis uncovers the public good nature of
patent invalidation. That is, invalidation reduces the risk of hold-up not only for the firm
that litigated the patent, but also for all firms operating in the technology space covered
by the invalidated patent. This positive effect on competing firms reduces the incentives
6
For example, Micrel Inc.’s 10-K filing for 2003 states that “on May 23, 2002, the Company
entered into a Patent Cross License and Settlement Agreement with National Semiconductor
which settled all outstanding patent disputes between the companies and cross licensed the entire
patent portfolio of each company.”

to challenge patent validity, especially in a crowded technology area like semiconductors.

Using data on opposition against patents at the European Patent Office, Harhoff, von
Graevenitz, and Wagner (2015) show that in fields with a large number of mutually
blocking patents, the incidence of opposition is sharply reduced, particularly among
large firms.
V. CONCLUSION
The past 50 years have witnessed remarkable improvements in the speed, size, and power
of semiconductor devices, providing a vital underpinning for the modern information
economy. Both before and after fundamental shifts in U.S. patent and antitrust policies
during the 1980s, semiconductor companies devised ways to move the technological
frontier forward. It might be tempting to conclude from this fact that patent rights fail
to shape technological progress in this industry. Overall, the empirical evidence is at odds
with this view. Instead, the evidence calls for a more nuanced interpretation: even within a
single industry, patent rights play multifaceted roles that can alter the innovative activities
of firms. Within this industry, the ability to secure strong and enforceable protection for
patented inventions seems particularly important for new companies that rely on external
sources of financing and for smaller firms. Although patent hold-up is difficult to observe
directly, indirect evidence and interview insights suggest that capital-intensive firms in
the industry are particularly vulnerable to such problems. However, these companies also
find ways to help mitigate such problems, whether through improving their bargaining
positions with larger portfolios of patents, through ex ante licensing and cross-licensing
agreements, or through participation in SSOs.
There are several useful directions for further research on patent rights and innovation
in the semiconductor industry. First, more empirical and survey evidence on the actual
timing and structure of patent licensing negotiations would be extremely useful in assess-
ing the impact of patent thickets, royalty stacking, and hold-up on technology diffusion
and innovation incentives. Second, there is the need for an investigation of patent asser-
tion entities operating in the semiconductor space. Specifically, it would be very valuable
to understand whether non-practicing firms play a key role as patent intermediaries or,
instead, whether they exacerbate hold-up problems through patent-trolling behavior.
Third, the recent decades have witnessed an internationalization of the semiconductor
production process that currently spans many countries (Breznitz, 2005). Exploring the
link between international patent law and global fragmentation of the semiconductor
manufacturing and commercialization process is a promising avenue for future research.
REFERENCES
Adelman MJ. 1987. “The New World of Patents Created by the Court of Appeals for the Federal Circuit”, 20(4)
Journal of Law Reform 979–1007.
Anand BN and Khanna T. 2000. “The Structure of Licensing Contracts”, 48(1) Journal of Industrial Economics
103–35.
Arora A, Fosfuri A, and Gambardella A. 2001. Markets for Technology: The Economics of Innovation and
Corporate Strategy. Cambridge, MA: MIT Press.

Bhaskarabhatla A and Hegde D. 2014. “An Organizational Perspective on Patenting and Open Innovation”,
25(6) Organization Science 1744–63.
Breznitz D. 2005. “Development, Flexibility and R&D Performance in the Taiwanese IT Industry: Capability
Creation and the Effects of State–Industry Coevolution”, 14(1) Industrial and Corporate Change 153–87.
Coase R. 1960. “The Problem of Social Cost”, 3 The Journal of Law and Economics 1–44.
Cohen WM, Nelson RR, and Walsh J. 2000. “Protecting Their Intellectual Assets: Appropriability Conditions
and Why U.S. Manufacturing Firms Patent (or Not)”, NBER Working Paper #7552.
Contreras JL. 2017. “Patents, Technical Standards, and Standards-Setting Organizations: A Survey of the
Empirical, Legal, and Economics Literature”, in P Mennell, D Schwartz, eds., Research on the Economics of
Intellectual Property Law Volume 2—Analytical Methods. Cheltenham, UK: Edward Elgar.
Epstein RA, Kieff FS, and Spulber DF. 2012. “The FTC, IP, and SSOs: Government Hold-Up Replacing
Private Coordination”, 8(1) Journal of Competition Law and Economics 1–46.
European Commission. 2011. “Implementing Enhanced Cooperation in the Area of the Creation of Unitary
Patent Protection”, Commission Staff Working Paper.
Farrell J, Hayes J, Shapiro C, and Sullivan T. 2007. “Standard Setting, Patents and Hold-up”, 3 Antitrust Law
Journal 603–70.
Galasso A. 2012. “Broad Cross-License Negotiations”, 21 Journal of Economics and Management Strategy
873–911.
Galasso A and Schankerman M. 2010. “Patent Thickets, Courts and the Market for Innovation”, 41(3) RAND
Galasso A and Schankerman M. 2015a. “Patents and Cumulative Innovation: Causal Evidence from the
Courts”, 130(1) Quarterly Journal of Economics 317–69.
Galasso A and Schankerman M. 2015b. “Patent Rights, Innovation and Firm Exit”, NBER Working Paper
No. 21769.
Galasso A, Schankerman M, and Serrano CJ. 2013. “Trading and Enforcing Patent Rights”, 44(2) Rand Journal
Gallini N. 2002. “The Economics of Patents: Lessons from Recent U.S. Patent Reform”, 16(2) Journal of
Economic Perspectives 131–54.
Graham SJH and Mowery DC. 2003. “Intellectual property protection in the US software industry”, Patents in
the Knowledge-based Economy 219–58, p. 231.
Graham SJH, Merges RP, Samuelson P, and Sichelman T. 2009. “High Technology Entrepreneurs and the
Patent System: Results of the 2008 Berkeley Patent Survey”, 24 Berkeley Technology Law Journal 1255–327.
Griliches Z. 1992. “The Search for R&D Spillovers”, 94 Scandinavian Journal of Economics 29–47.
Grindley PC and Teece DJ. 1997. “Managing Intellectual Capital: Licensing and Cross-Licensing in
Semiconductors and Electronics”, 39(2) California Management Review 101–28.
Hall BH. 2005. “Exploring the Patent Explosion”, 30(1) Journal of Technology Transfer 35–48.
Hall BH and Lerner J. 2010. “The Financing of R&D and Innovation”, in BH Hall and N Rosenberg, eds.,
Elsevier Handbook of the Economics of Innovation. Elsevier-North Holland.
Hall BH and Ziedonis RH. 2001. “The Patent Paradox Revisited: An Empirical Study of Patenting in the U.S.
Semiconductor Industry, 1979–95”, 32(1) Rand Journal of Economics 101–28.
Harhoff D, Von Graevenitz G, and Wagner S. 2015. “Conflict Resolution, Public Goods, and Patent Thickets”,
62(3) Management Science 704–21.
Heller MA and Eisenberg RS. 1998. “Can Patents Deter Innovation? The Anticommons in Biomedical
Research”, 280(5362) Science 698–701.
Holbrook D, Cohen WM, Hounshell DA, and Klepper S. 2000. “The Nature, Sources, and Consequences of
Firm Differences in the Early History of the Semiconductor Industry”, 21 Strategic Management Journal
1017–41.
Hsu D and Ziedonis RH. 2013. “Resources as Dual Sources of Advantage: Implications for Valuing
Entrepreneurial-Firm Patents”, 34(7) Strategic Management Journal 761–81.
Jaffe A. 2000. “The U.S. Patent System in Transition: Policy Innovation and the Innovation Process”, 29
Klein B, Crawford RG, and Alchian AA. 1978. “Vertical Integration, Appropriable Rents, and the Competitive
Contracting Process”, 21(2) The Journal of Law and Economics 297–326.
Klein B. 1998. “Hold-up Problem”, in P Newman ed., The New Palgrave Dictionary of Economics and the Law.
Palgrave Macmillan UK.
Lanjouw JO and Lerner J. 2001. “Tilting the Table? The Use of Preliminary Injunctions”, 44(2) Journal of Law
and Economics 573–603.
Lanjouw JO and Schankerman M. 2001. “Characteristics of Patent Litigation: A Window on Competition”,
32(1) RAND Journal of Economics 129–51.
Lanjouw JO and Schankerman M. 2004. “Protecting Intellectual Property Rights: Are Small Firms
Handicapped?”, 47(1) Journal of Law and Economics 45–74.

Lemley MA and Shapiro C. 2006. “Patent Hold-up and Royalty Stacking”, 85(2006) Texas Law Review
1991–2049.
Lemley MA and Shapiro C. 2013. “A Simple Approach to Setting Reasonable Royalties for Standard-Essential
Patents”, 28 Berkeley Technology Law Journal 1135–66.
Lerner J. 1995. “Patenting in the Shadow of Competitors”, 38 Journal of Law and Economics 563–95.
Lerner J and Tirole J. 2014. “A Better Route to Tech Standards”, 343(6174) Science 972–73.
Levin RC. 1982. “The Semiconductor Industry”, in RR Nelson, ed., Government and Technical Progress: A
Cross-Industry Analysis. Oxford UK: Pergamon Press.
Levin RC, Klevorick AK, Nelson RR, and Winter SG. 1987. “Appropriating the Returns from Industrial
Research and Development”, 3 Brookings Papers on Economic Activity 783–820.
Long C. 2002. “Patent Signals”, 69 University of Chicago Law Review 625–79.
Macher JT and Mowery DC. 2003. “Managing Learning by Doing: An Empirical Study in Semiconductor
Manufacturing”, 20 Journal of Product Innovation Management 391–410.
Macher JT, Mowery DC, and Hodges D. 1998. “Reversal of Fortune? The Recovery of the US Semiconductor
Industry”, 41(1) California Management Review 107–36.
Markoff J. 2016. “Moore’s Law Running Out of Room, Tech Looks for a Successor”, May 4, New York Times.
Merges RP. 1997. Patent Law and Policy: Cases and Materials, 2nd edn. Charlottesville VA: The Mitchie Company.
Monteverde K. 1995. “Technical Dialog As an Incentive for Vertical Integration in the Semiconductor
Industry”, 41 Management Science 1624–38.
National Research Council. 2004. A Patent System for the 21st Century. Washington DC: National Academies
Press.
Rysman M and Simcoe TS. 2008. “Patents and the Performance of Voluntary Standards-Setting Organizations”,
Samuelson P and Scotchmer S. 2002. “The Law and Economics of Reverse Engineering”, 111(7) Yale Law
Journal 1575–663.
Scherer FM, Herzstein SE, Dreyfoos AW, Whitney WG, Bachman OJ, Pesek CP, Scott CJ, Kelly TG, and Galvin
JJ. 1959. Patents and the Corporation: A Report on Industrial Technology Under Changing Public Policy, 2nd
edn. Boston MA: Harvard Business School.
Semiconductor Industry Association (SIA). 2006. Semiconductor Industry Association, 2006 Association Annual
Report, available at www.siaonline.org/cs/annual_report.
Semiconductor Industry Association (SIA). 2016. Factbook, available at http://go.semiconductors.org/2016-sia-
factbook-0-0.
Serrano C. 2010. “The Dynamics of the Transfer and Renewal of Patents”, 41(4) RAND Journal of Economics
686–708.
Shapiro C. 2001. “Navigating the Patent Thickets: Cross-Licenses, Patent Pools, and Standard-Setting”, in AB
Jaffe, J Lerner, and S Stern, eds., Innovation Policy and the Economy vol 1. Cambridge MA: National Bureau
of Economic Research.
Shapiro C. 2003. “Technology Cross-Licensing Practices: FTC vs Intel”, in J Kwoka and L White eds., The
Antitrust Revolution. Oxford UK: Oxford University Press.
Shapiro C. 2016. “Patent Remedies”, forthcoming American Economic Review Papers & Proceedings.
Siebert R and von Graevenitz G. 2010. “Jostling for Advantage or Not: Choosing Between Patent Portfolio
Races and Ex Ante Licensing”, 73(2) Journal of Economic Behavior and Organization 225–245.
Simcoe TS. 2012. “Standard Setting Committees: Consensus Governance for Shared Technology Platforms”,
Simcoe TS, Graham SJH, and Feldman MP. 2009. “Competing on Standards? Entrepreneurship, Intellectual
Property, and Platform Technologies”, 42 Journal of Economic Management and Strategy 775–816.
Somaya D. 2003. “Strategic Determinants of Decisions Not to Settle Patent Litigation”, 24(1) Strategic
Management Journal 17–38.
Tansey R, Neal M, and Carroll R. 2005. “Get Rich, or Die Trying: Lessons from Rambus’ High-Risk Predatory
Litigation in the Semiconductor Industry”, 12(1) Industry and Innovation 93–115.
Teece DJ. 1986. “Profiting from Technological Innovation: Implications for Integration, Collaboration,
Licensing and Public Policy”, 15(6) Research Policy 285–305.
Theron DC, Rayter J, Paprotny I, and Mackulak GT. 1999. A Quantitative Approach for Minimizing Fab Cost
of Ownership through Systems Integration, in the Proceedings of the 10th Annual IEEE/SEMI Advanced
Semiconductor Manufacturing Conference and Workshop, Boston, MA.
Tilton JE. 1971. International Diffusion of Technology: The Case of Semiconductors. Washington DC: Brookings
Institution.
Turley J. 2005. The Essential Guide to Semiconductors. Prentice Hall Professional Technical Reference.
US Federal Trade Commission (FTC). 2003. To Promote Innovation—The Proper Balance of Competition and
Patent Law and Policy. Washington, DC: FTC. Available at /www.ftc.gov/opp/intellect/.
US Federal Trade Commission (FTC). 2011. The Evolving IP Marketplace: Aligning Patent Notice and Remedies

with Competition, available at www.ftc.gov/sites/default/files/documents/reports/evolving-ip-marketplace-

aligning-patent-notice-and-remedies-competition-report-federal-trade/110307patentreport.pdf.
US International Trade Commission. 2006. Semiconductor Manufacturing Equipment. USITC Publication 3868.
US Patent and Trademark Office (USPTO). 1995. “Technology Profile Report: Semiconductor Devices and
Manufacture”, 1/1969–12/1994. Washington DC: US Department of Commerce.
US White House. 2013. Patent Assertion and US Innovation. Executive Office of the President. Washington DC:
US White House.
Von Graevenitz G, Wagner S, and Harhoff D. 2011. “How to Measure Patent Thickets—A Novel Approach”,
111(1) Economic Letters 6–9.
Von Hippel E. 1988. The Sources of Innovation. Oxford UK: Oxford University Press.
Watzinger M, Fackler T, Nagler M, and Schnitzer M. 2017. How Antitrust Enforcement Can Spur Innovation:
Bell Labs and the 1956 Consent Decree. University of Munich Working Paper.
Weise E. 2015. “How Computers Got Amazing: Moore’s Law at 50”, 15 April, USA Today.
Williamson O. 1985 The Economic Institutions of Capitalism. Simon and Schuster.
Ziedonis RH. 2003. “Patent Litigation in the U.S. Semiconductor Industry”, in WA Cohen and SA Merrill, eds.,
Patents in the Knowledge-Based Economy. Washington DC: National Academies Press.
Ziedonis RH. 2004. “Don’t Fence Me In: Fragmented Markets for Technology and the Patent Acquisition
Strategies of Firms”, 50(6) Management Science 804–20.
Cases
Atlas Powder Co. v. Ireco Chemicals (1985), 773 F.2d 1230.

eBay Inc. v. MercExchange, L.L.C. (2006), 547 U.S. 388.
Ericsson, Inc. v. D-Link Systems, Inc. (2014), 773 F.3d. 1201.
Smith International v. Hughes Tool Co. (1983), 718 F.2d 1573.
Wang Labs., Inc. v. Mitsubishi Elecs. Am. Inc. (1997), 103 F.3d 1571.
National Cooperative Research Act, P.L. 98-462 (1984).

Rambus Inc., FTC (2006) Docket No. 9302, Opinion of the Commission 3, available at
www.ftc.gov/os/adjpro/d9302/060802commissionopinion.pdf.

19. Patent trolls
Jay P. Kesan* 7
Contents
I. Introduction
II. The Definition and History of Non-practicing Entities
III. Distinguishing Between the Types of Non-practicing Entities
A. Universities
B. Individual Inventors
C. Large Patent Aggregators
D. Failed Operating/Start-up Company
E. Patent Holding Company
F. Operating Company
G. Technology Development Company
H. Other Unconventional Non-practicing Entities
IV. Examining the Benefits and Disadvantages of Non-practicing Entities
A. Benefits of Non-practicing Entities
B. Disadvantages of Non-practicing Entities
V. Current Proposed Solutions to Non-practicing Entities’ Abusive Behavior
A. Recent Efforts to Reduce Abusive Non-practicing Entities’ Behavior
1. The American Invents Act of 2011
2. Elimination of Federal Rules of Civil Procedure Form 18
3. Proposed Congressional legislation
4. Patent insurance coverage
VI. Conclusion
References
I. INTRODUCTION
Newspapers have covered how nearly 1000 farmers were sued for patent infringement in
just one year in the same court by the same patent-owning plaintiff who neither farmed
nor manufactured any farming products himself (Mossoff, 2014). However, you may not
recall reading or hearing about this because it happened over 125 years ago in the 1880s.
Non-practicing entities (NPEs) have existed for centuries, often playing integral roles in
innovation and some of society’s most well-known technological advancements.
A patent troll—also commonly referred to as an NPE, patent assertion entity (PAE),
* Professor of Law and Workman Research Scholar at the University of Illinois at

Urbana-Champaign.
445

or patent monetization entity—is an individual or company that owns one or more

patents but does not actually make, use, or practice the product or process that the patent
covers (Cotropia, Kesan, and Schwartz, 2014; Jeruss, Feldman, and Walker, 2012; Chien,
2010). This chapter is meant to serve as a review of studies of NPEs and their effect
on the United States patent system. The information presented is not intended to pass
judgment or persuade readers that NPEs are definitively beneficial or detrimental to the
patent system, but rather seeks to inspire readers to assess the issue of NPEs from a more
detailed and informed perspective. Ultimately, it is hoped that this material will serve as
an encouragement to carefully examine the effects of NPEs on the patent system.
There are varying opinions and interpretations on NPEs, their behavior, and the effects
of their behavior. Those who study NPEs each offer their own estimation of the effects of
NPEs on the patent system, with generalized claims such as “NPEs cause an increase in
litigation,” “on average NPEs cost defendants X amount of dollars,” and “NPEs siphon
Y amount of dollars from legitimate business practices.” Yet the accuracy of this data is
problematic. First, the alleged “increase in litigation” sometimes may be attributable to
factors other than NPEs’ behavior. For example, the recent increase in number of patent
cases may be attributable, at least partially, to the passage of the Leahy-Smith American
Invents Act (AIA) (Pub. L. No. 112-29, 125 Stat 284, 2011). The AIA eliminated the
practice of joinder of defendants based on mere allegations that each defendant infringed
the same patent (35 U.S.C. § 299, 2012; Taylor, 2013). This forced many patent plaintiffs
to file separate suits against different alleged infringers, inflating the total number of
patent cases (Cotropia, Kesan, and Schwartz, 2017). Second, accurate and meaningful
information on the influence of NPEs on the patent system is not readily available because
of the prevalence of privately-owned parties disinclined to publicize delicate information,
confidential settlement agreements, and a general unwillingness to “put their cards on
the table” to show any competitive advantage they have discovered. Finally, opinions on
NPEs may be inaccurate because of the tendency to misclassify a party as an NPE and/or
mistakenly lump the behavior of extremely different types of NPEs into one homogenous
category.
The patent system is a complex being, like the human body; the minor effects of an
activity can often been felt in other unrelated areas. For instance, changes in procedural
patent litigation can affect how a company manages its operations and patent portfolio,
and this in turn ultimately affects the economy (Laird, 2015). While examining the patent
system at a granular level can provide key insight into a particular area, overly specific
examinations lead to inaccurate over-generalizations when such conclusions are extrapo-
lated. It is important to remember that the circumstances vary on a case-by-case and
fact-by-fact basis. For example, a pharmaceutical company may wish to make sure that it
preserves its hold on the market for the full term of its patent, whereas many information
technology patent owners abandon their rights in the middle of a patent’s term because
the underlying technology is not relevant to their future products/services.
Current empirical studies offer mixed evidence and conclusions regarding the effect of
NPEs on the patent system. Ideally, after reading this chapter, a researcher will be capable
of determining substantial differences in behavior and characteristics among the various
types of NPEs. This would further help the researcher to differentiate which kinds of
NPEs behavior are detrimental and which are favorable to the patent system.
Section II of this chapter provides a brief historical context of NPEs in the United

Patent trolls 447
States. In section III, the various types of NPEs are described and differentiated in some
depth. Section IV examines the benefits and disadvantages of NPEs. Finally, section V
explores current reforms addressing NPEs and their behavior. Section VI concludes.
II. THE DEFINITION AND HISTORY OF NON-PRACTICING

ENTITIES
To those unfamiliar with patent law, the modern idea of a patent troll appears to be a
relatively new concept that has developed as a result of twentieth-century advancements
in channels of communication. The mass media portrays patent trolls like the trolls of
mythical folktales that guard a bridge and demand payment from travelers to pass. Like
folktale trolls, patent trolls demand payments from practicing entities for permission to
continue making and/or using an invention (EconoSTATS, 2015). Thus, patent trolls are
pictured as profiting by overwhelmingly intimidating alleged infringers so that they have
no choice but to “pay the troll” (EconoSTATS, 2015; La Belle, 2014).
Mass media coverage on the issue of patent trolls has increased in the last several years,
thereby furthering the notion that this is a novel problem that threatens to undermine the
patent system, stifle innovation, and disrupt the economy (Economist, 2015). In 2013,
President Obama commented on patent trolls, “They don’t actually produce anything
themselves. They’re just trying to essentially leverage and hijack somebody else’s idea and
see if they can extort some money out of them” (Sternburg, 2013). Supreme Court Justice
Anthony Kennedy noted in eBay v. MercExchange, 547 U.S. 388 (2006) that “[a]n industry
has developed in which firms use patents not as a basis for producing and selling goods
but, instead, primarily for obtaining licensing fees.”
The patent troll business model is commonly portrayed as seeking financial reward
through demanding that other companies, which are allegedly infringing the invention
covered by the patent troll’s patent, pay a licensing fee to use the patent. If they refuse to
pay, alleged infringers face the threat of a patent infringement lawsuit. There is a chance
an alleged patent infringer could prevail in court by proving it did not infringe the patent
or that the patent is invalid, and thus unenforceable. However, in such a case, the defend-
ant faces the substantial risk of having to pay exorbitant damages and/or having its entire
operation shut down through a permanent injunction (Cotropia, Kesan, and Schwartz,
2014). Additionally, litigating a patent infringement lawsuit is expensive, with the average
NPE-related lawsuit estimated to cost about $1.5 million to defend (Pham, 2014). It is
estimated that patent troll-related litigation has cost the U.S. economy $320 billion over
the last five years (Cotropia, Kesan, and Schwartz, 2014). Patent lawsuits can often drag
on for years, and thus can prevent accused infringing companies from operating and/or
growing their business (Pogue, 2015). Once litigation progresses, alleged infringer defend-
ants must undergo burdensome discovery, and thereby become susceptible to having
private information and secret competitive advantages disclosed. In this way, patent trolls
heavily incentivize alleged infringers to pay a licensing fee to avoid litigation that is likely
to cost ten to 100 times as much as the proposed licensing fee (Burt, 2015). This model
is not necessarily an accurate depiction of NPEs’ behavior. Although studies do show
that over 80 percent of all defendants exit NPE-initiated lawsuits because of voluntary
settlements (Cotropia, Kesan, and Schwartz, 2014), this figure is no different from the rate

of voluntary settlements in patent cases brought by operating company patent plaintiffs

(Cotropia, Kesan, and Schwartz, 2017).
Further, history suggests that profiting from patents that patent holders do not them-
selves practice is not a new idea. Thomas Alva Edison, one of America’s most recognized
and renowned inventors, both sold and licensed his patented innovations (Mossoff, 2014).
Early in his career, Edison sold his patents to finance his research and development.
Even after becoming widely successful, Edison still employed the secondary market,
most notably selling his patented incandescent light bulb to General Electric. Ironically,
by today’s standards, Edison’s routine conduct of inventing and then selling or licensing
his discoveries would earn him the moniker of “Patent Troll of the Golden Gate Bridge”
instead of “The Wizard of Menlo Park.”
In 1895, George Baldwin Selden, an American attorney famous for representing Susan
B. Anthony, was granted a patent for an “improved road-engine” (History.com, 2017).
Despite copying the idea from an exhibit at the 1872 Centennial Convention, Selden
profited immensely through royalty payments achieved from suing automobile manu-
facturers and from eventually selling the patent to a group of investors which continued
to sue automobile manufacturers (Economist, 2015). Charles Goodyear, best known for
inventing a process to create vulcanized rubber commonly used today in car tires, never
manufactured or sold his products (Mossoff, 2015). Rather, Goodyear sold the patent he
obtained in 1844 for his invention to other inventors and firms, which used the patented
process to manufacture, sell, and use rubber products that were vital in the success of the
industrial revolution.
Thus, NPEs have long been an integral part of the U.S. patent system, and the U.S.
economy as a whole. Technological advancement in information technology and the
globalization of intellectual property are likely among the reasons why NPEs appear to
be a novel enterprise. With the mechanical age and industrial revolution giving way to the
electronic age and advancements in biotechnology, pharmaceuticals, and communication
technologies, the only increasing development of the twenty-first century has been renam-
ing these parties “patent trolls” (EconoSTATS, 2015). It may be near impossible to define
the term “patent troll” without unintentionally conveying a bias toward the activity, but
a more accurate and encompassing definition of an NPE is an individual or company
whose business model is to license patented innovations to others instead of practicing,
manufacturing, using, or selling the innovation itself (Mossoff, 2014).
Finally, it is still unclear whether there has been a recent substantial rise in activity by
NPEs (Electronic Frontier Foundation, 2017). Other factors could be contributing to the
increase, or perceived increase, in patent trolling, such as more patents being issued, those
patents that are issued being more overly broad and vague than in the past, or shifts in
certain sectors of technology. For instance, profound advancements in mobile technology
have created a market for the development of mobile phone applications (Rice, 2013).
Unlike other types of technology, building a mobile app requires limited programming
skill and minimal financial capital, and has very few front-end barriers to entry. A recent
survey of 100 mobile designers reported that it takes just 18 weeks on average to build a
mobile application from the idea phase until it is available for download and widespread
use. Consequently, mobile application developers have been the targets of NPEs (which
own patents over the technology utilized by the applications) demanding the app develop-
ers pay for a license or face a lawsuit (Electronic Frontier Foundation, 2017).

Patent trolls 449
III. DISTINGUISHING BETWEEN THE TYPES OF NON-

PRACTICING ENTITIES
NPEs come in many shapes and sizes. NPEs run the gamut from universities and
individual inventors trying to monetize their scientific discoveries to venture capitalists
seeking to exploit other’s inventions for financial gain (Cotropia, Kesan, and Schwartz,
2014). NPEs’ motivations and objectives vary greatly (Schwartz and Kesan, 2014; Lemley
and Melamed, 2017; Lee, 2015). Consequently, the litigation outcome also varies depend-
ing on the type of NPE (Allison, Lemley, and Schwartz, 2017). While NPEs may not
practice the inventions that they hold patents on, some NPEs employ a sizable staff of
engineers to conduct research and development in pursuit of patentable advancements,
and thus are legitimately pursuing technological advancements (Maurer and Haber,
2017). In some instances, companies such as Rambus have evolved from NPE patent
licensing companies to operating companies. Thus, to truly understand the effect NPEs
have on the patent system, it is imperative to differentiate and define the various types of
NPEs (Cotropia, Kesan, and Schwartz, 2014; Risch, 2012; Ball and Kesan, 2009). Patent
law scholars recognize the following types of entities as potentially NPEs: universities,
individual inventors, patent aggregators, failed operating or start-up companies, patent
holding companies, operating companies, technology development companies, and other
unconventional entities (Cotropia, Kesan, and Schwartz, 2014).
A. Universities
A university is a public or private institution of higher education that conceives patent-

able discoveries and inventions through its institution’s research (Cotropia, Kesan, and
Schwartz, 2014; Lemley, 2008). An example of an NPE university is the University of
Wisconsin (Nocera, 2015). Universities license their patents to other companies because
universities are concerned with enhancing their reputation by conducting research to
expand the fields of science and promote their students’ education (Pérez-Peña, 2013).
Additionally, universities lack the resources necessary to operate, and/or are disinterested
in assuming the risks associated with, a full-fledged business related to their inventions.
While some of these licensing deals can be incredibly lucrative, such as the University of
Wisconsin’s recent $234 million judgment against Apple, many university licensed patents
yield paltry financial returns in comparison to the large amount of time and money spent
researching and developing the invention (Nocera, 2015; Pérez-Peña, 2013; Krogstad,
2012). Royalties from university-developed patents are typically distributed among the
inventor, the university, and the university’s technology transfer office. The university-
employed inventor’s portion of the royalty functions as a method of compensation; the
university reinvests its share in more research, university-funded programs, and operating
costs; and the technology transfer office’s share of the royalties goes towards the office’s
operating costs. Nonetheless, all patents asserted by universities emanate from the uni-
versity’s own research efforts and development of the underlying technology (Cotropia,
Kesan, and Schwartz, 2014).

B. Individual Inventors
Individual inventors may own one or more patents solely or jointly with other inventors
(Cotropia, Kesan, and Schwartz, 2014). Like universities, individual inventors have
developed the underlying technology of their patent. Some individual inventors lack
the financial capital and/or access to the industrial infrastructure necessary to practice,
develop, or maintain their patented inventions, and thus licensing deals are the only way
to monetize their inventions (Quinn, 2015a; Cotropia, 2009). These individual inventors
then seek more established operating companies to jointly benefit from their invention
(Quinn, 2015a). Individual inventors are susceptible to having their patented inventions
infringed by large operating companies unwilling to pay for a license, and in such instances
the individual inventor’s only way to enforce his or her intellectual property rights is to
sue the operating company for patent infringement. An example of an individual inven-
tor is Thomas Edison (Mossoff, 2014). These are also the types of inventors commonly
portrayed in movies, such as Joy and Flash of Genius, who are working against the odds
to achieve the “American dream” by providing society with an amazing technological
breakthrough (Cotropia, Kesan, and Schwartz, 2014).
However, individual inventors can also be individuals, like George Selden, who recog-
nize openings and trends in existing technology, and then shrewdly attain dubious patents
with the sole purpose of asserting the patent against a practicing entity for financial gain
(Economist, 2015). Thus, there is great variety in terms of interests and behaviors even
among individual inventor NPEs.
C. Large Patent Aggregators
A large patent aggregator is a company with a large patent portfolio whose primary busi-
ness is enforcing the patents of numerous other individuals and entities (Cotropia, Kesan,
and Schwartz, 2014; Schwartz, 2014). Patent aggregators do not develop the underlying
technology of the patents they own, but instead purchase the patents of others (Cotropia,
Kesan, and Schwartz, 2014). Because large patent aggregators own thousands of patents,
they can pursue a patent monetization strategy that is highly diversified, with reduced
risk, and involving cumulative assimilation of specialized knowledge over time. In many
instances, large patent aggregators only assert their strongest and most valuable patents.
Such patents are usually very difficult to invalidate, because they cover truly novel and
non-obvious inventions. Further, infringement of such strong patents is hard to avoid,
because they cover innovations that are core to the relevant technology or industry.
Because of the small chance of showing that the patent is not being infringed or that the
patent is invalid, an accused infringer faces higher pressure to settle. Examples of patent
aggregators include Acacia companies, Wi-Lan, and Intellectual Ventures.
D. Failed Operating/Start-up Company
A failed operating or start-up company is an entity that owns one or more patents cover-
ing a technology that it originally invented, although was unsuccessful in its attempt to
practice the invention (Cotropia, Kesan, and Schwartz, 2014; Osenga, 2015). These com-
panies have retained their original patents and later seek to enforce them as the method

Patent trolls 451
for generating revenue (Cotropia, Kesan, and Schwartz, 2014). The revenue earned from
enforcing the patents enables original corporate backers and investors, which otherwise
would receive no money, to recoup and/or achieve a return on their original investment.
An example of a failed operating or start-up company is Broadband Graphics LLC.
E. Patent Holding Company
Patent holding companies are entities, such as limited liability companies, that appear
to have been formed solely to hold and enforce a patent or small portfolio of patents
(Cotropia, Kesan, and Schwartz, 2014). Typically, the original inventor of the patent does
not own the patent holding company. Frequently, these companies are formed shortly
before litigation is commenced in order to insulate the current owner of the patent.
Patent holding companies can also operate as “patent pools” to provide a single source
for licensing of a certain technology. When a patent holding company operates as a pool,
it creates market efficiency by lowering the barriers to entry.
F. Operating Company
Operating companies are entities that manufacture products and deliver services (Cotropia,
Kesan, and Schwartz, 2014). Operating companies also frequently license their patents
(Self, 2015). The revenues yielded from an operating company licensing its patents are
often reinvested in the research and development of new inventions. While operating
companies are not typically considered NPEs, they can still exhibit “patent troll” type
behavior. For instance, an operating company can own a patent covering an invention
and, despite not practicing the underlying technology of the patent, still assert the patent
against competitors through licensing agreements and litigation. Additionally, an operat-
ing company can transfer its patents to a patent holding company in order to protect itself
and its assets (Cotropia, Kesan, and Schwartz, 2014). Even if not practicing the technol-
ogy underlying the patent being asserted, an operating company may—unlike most other
NPEs—be exposed to patent infringement counterclaims, because the defendant may
use its patent portfolio to assert one of its own patents against the technologies that are
actually being employed by the operating company (Cotropia, Kesan, and Schwartz,
2017; Love, 2013; Chien, 2009). An example of an operating company is Hewlett Packard.
G. Technology Development Company
A technology development company invests in the invention of technology, sometimes

with the intention of licensing rather than practicing the patent (Cotropia, Kesan, and
Schwartz, 2014). While a technology development company is the original owner of the
patent, it does not manufacture products covered by the patents. Examples of technology
development companies are Walker Digital LLC and Tessera Technologies.
H. Other Unconventional Non-practicing Entities
While the aforementioned categories describe entities that own patents, there are also
NPEs that do not own any patents but still utilize the patent system for financial gain

(Quinn, 2015b). In 2015, Hayman Capital Management, a successful hedge fund that
made a fortune shorting subprime mortgages during the recent American housing market
downturn, introduced a new strategy (Fisher, 2015). Hayman Capital, led by its founder
Kyle Bass and infamous patent asserter Erich Spangenberg, began filing Inter Partes
Review (IPR) petitions within the U.S. Patent Trial and Appeal Board (PTAB) to chal-
lenge the validity of the patents covering biotechnology and pharmaceutical companies
most successful products. Concurrent with its IPR filings, Hayman Capital would short
the operating company’s stock whose patents it owned and/or buy the stock of the com-
pany’s competitors (Fisher, 2015; Quinn, 2015). Hayman Capital’s expectation was that
the patent owning company’s stock would plummet, and/or its competitors’ stock would
skyrocket, as a result of (the fear of) the patent being invalidated through the PTAB’s post-
grant proceedings, and the company would lose its market share to competitors—thereby
enabling Hayman Capital and its clients to make millions.
Additionally, because anyone can file a PTAB post-grant proceeding, regardless of
whether they are a real party of interest to the patent, NPEs can file post-grant proceed-
ings against a patent holder and then profit by agreeing to a settlement in exchange for
terminating the PTAB petition (Gulfo, 2015). Unlike patent-owning NPEs that risk
having the validity of their patent challenged by an alleged infringer, there is no form
of recourse against NPEs that do not own any patents (Quinn, 2015b). Further, this
“reverse trolling” conduct is currently legal (Gulfo, 2015). While this behavior seems to
be especially dastardly, it can be partially motivated by, for example, trying to reduce the
price of pharmaceuticals for consumers by having more suppliers in the marketplace—
accomplished by eliminating the “monopoly” created by the company’s patent.
Other types of NPE that do not own any patents include entities that fraudulently
misrepresent ownership in a patent or that try to enforce an expired patent in order to
extract money from an operating company by sending vague demand letters (H.R. 2045,
114th Cong. (2015)). This behavior is, of course, illegal.
IV. EXAMINING THE BENEFITS AND DISADVANTAGES OF

NON-PRACTICING ENTITIES
On the one hand, an NPE may be asserting a truly novel and well-deserved patent to help
regulate the market for such inventions by restricting incremental inventions, deterring
infringement, and fostering efficiency by allowing the designers to just pay the licensing
fee and focus on app development. On the other hand, an NPE may be using its patent
purely to prey on unsuspecting developers for financial gain. Therefore, detailed empirical
studies are needed to ascertain how specific NPE behavior is beneficial and/or detrimental
to the patent system.
Many allege that NPEs assert frivolous lawsuits using dubious patents. However, the
litigation behavior of NPEs varies greatly and is specific to the type of NPE (Cotropia,
Kesan, and Schwartz, 2014; Hagiu and Yoffie, 2013).1 A recent empirical study on the
1
Notably, the role of NPEs is not limited to litigation. See Andrei Hagiu and David B. Yoffie,
“The New Patent Intermediaries: Platforms, Defensive Aggregators, and Super-Aggregators”,

Patent trolls 453
litigation behavior of NPEs revealed that defendants in NPE-initiated patent litigation

lawsuits settle about 80 percent of the time, as do defendants in operating company-
initiated patent litigation suits (Cotropia, Kesan, and Schwartz, 2014). In fact, when
comparing NPE litigation behavior and outcomes with operating companies, there is
very little difference in their settlement rate, in the rate of adjudications on the merits
(i.e., grant of summary judgment or judgment after trial), or in the rate of procedural
dispositions. Additionally, this study also showed there is little support for the “hit and
run” (i.e., sue and promptly settle) anecdotes, as NPEs did not settle patent lawsuits in
60 or 120 days after filing a complaint any more than operating companies. In fact, in
cases involving a judgment after trial which were brought by individuals, the alleged
infringer prevailed 94 percent of the time. In cases involving a judgment after trial which
were brought by failed operating companies, the alleged infringer prevailed 60 percent of
the time. In cases involving a judgment after trial which were brought by patent holding
companies, the alleged infringer prevailed 79.6 percent of the time. In cases brought by
operating companies, the alleged infringer prevailed 52.9 percent of the time. While the
results of this data can be interpreted in a variety of ways, it is clear that settlement is
the most likely outcome of NPE-initiated litigation and operating company-initiated
litigation, and an alleged infringer typically prevails in those cases that proceed to final
judgment. Most importantly, however, the data shows that different types of NPEs behave
differently when asserting their patent rights and, thus, overgeneralizations with regard to
NPEs are not warranted (Cotropia, Kesan, and Schwartz, 2017).
A. Benefits of Non-practicing Entities
NPEs provide many benefits to the patent system and society as a whole. Most notable is
that some NPEs do actually make groundbreaking discoveries that spur innovation, and
the current system allows these NPEs to be justifiably compensated for their contributions
(Stoll, 2014). In turn, this innovation drives economic growth, creates jobs, brings scien-
tific and technological advancements to society, and accomplishes the exact objectives for
which patents were enumerated in the constitution, namely, “[t]o promote the progress of
science and useful arts” (U.S. Const. art. 1, sec. 8, clause 8).
NPEs create a useful and more efficient intermediary for the exchange of technology
(Cotropia, Kesan, and Schwartz, 2014). Not all inventors possess the requisite business
skills or savvy to monetize their inventions, and similarly not all salespeople possess the
technical or creative ingenuity to invent. Inventors are often unable to accurately and
fairly assess the value of their inventions because they are very passionate about their
work. Thus, NPEs foster the shift in paradigm from closed innovation to open innovation.
Licensing deals permit practicing companies to bring innovation to the marketplace with-
out having to devote capital for years of research and development, yet still compensate
the inventor. By allowing the inventors to invent, the manufacturers to manufacture, the
investors to invest, the marketers to market, and the sellers to sell, licensing enables all of
the involved actors to concentrate on their strengths; and thus facilitates the market to
27 J. Econ. Perspectives 45 (2013) (providing an overview of NPE-related transactions outside of

litigation).

operate at maximum efficiency (Stoll, 2014). In turn, this theoretically allows inventions
to be more readily available and at more reasonable prices for consumers (Cotropia,
Kesan, and Schwartz, 2014). Therefore, arguably, the existence of NPEs helps facilitate
and regulate supply and demand. Further, any amount of money lost to licensing fees
and legal costs associated with NPEs is minuscule compared to the overall scheme of the
American technology sector (Maurer and Haber, 2017).
Without some of the current dynamics of NPEs, many useful discoveries would not
necessarily be realized or available because of a lack of funding, lack of optimized
production practices, and/or ineffective distribution channels (Cotropia, Kesan, and
Schwartz, 2014). Additionally, NPEs can play an important role in protecting inven-
tors (Parker, 2015). NPEs allow inventors to manage and shift their individual risk by
providing a financial safety net, a sword against misappropriation, and a shield against
any attackers.
While the aforementioned perspective represents an ideal model, the reality is that many
NPEs do not actually return significant money to inventors or investors (Maurer and
Haber, 2017). Instead, the NPEs and their lawyers retain the majority of the revenues. A
study on the profitability of publicly traded NPEs found that only a handful of NPEs are
profitable, with the most successful earning a mere 12 percent profit per year. In actuality,
the majority of NPEs actually lose money. Thus, despite the perceived low risks and high
rewards associated with NPEs, it is not as profitable a business as one would suspect.
B. Disadvantages of Non-practicing Entities
The main problem that some NPEs impose on the patent system is that they practice a
legalized form of “extortion” and create inefficiencies within the market (Pegoraro, 2015).
NPEs assert patents—often overly broad “wrongly issued patents for ‘inventions’ that
are neither new nor non-obvious” —against companies that are allegedly infringing these
patents. NPEs then send demand letters or file lawsuits against these alleged infringers
seeking compensation in the form of a license (Stoll, 2014). Some patent enforcement
activities by NPEs against alleged infringers are warranted, because they represent
legitimate legal actions against misappropriation of technology that was developed
after several years and millions of dollars of research and development (Nocera, 2015).
However, other NPEs do not even own the patent they seek to enforce, are asserting an
invalid or expired patent, and/or deliberately use other deceptive practices in order to
exploit other entities for financial gain (Pogue, 2015). While some NPEs offer the benefits
of a more efficient marketplace for the development and use of technology, other NPEs
damage the marketplace by creating inefficiencies through barriers to entry and increased
transaction costs (Parker, 2015). Ultimately, this creates disincentives to invent and invest
in new technology (Pogue, 2015). However, the issues of NPEs are not exclusive to the
technology sector (Chien, 2013). As a whole, NPEs sue more non-tech companies, such
as retailers and service providers, than tech companies.
Some NPEs’ enforcement strategy begins by sending an alleged infringer a demand
letter requesting it to pay a licensing fee or else be sued in federal court for patent
infringement (EconoSTATS, 2015). These demand letters can often be deliberately vague
in order to prevent the alleged infringer from being able to ascertain exactly which patent
its invention infringes or how it infringes. NPEs use this strategy to encourage alleged

Patent trolls 455
infringers to “just pay the troll,” rather than engage in even more costly and lengthy
litigation.2 Further, reaching a settlement protects the NPE’s patent from having its
validity challenged in counterclaims or in a declaratory judgment lawsuit.
NPEs often abuse the federal court system in order to threaten alleged infringers into
paying for a license (Parker, 2015). Prior to 2008, there were four times as many patent
cases filed by operating companies as those filed by NPEs (Gupta and Kesan, 2016). In
2012, PAEs initiated 62 percent of all patent litigation lawsuits (Chien, 2013). In the event
that an alleged infringer refuses to pay a licensing fee, prior to the elimination of Form 18
of the Federal Rules of Civil Procedure in December 2015, an NPE needed only to name
the patent and claim the patent had been infringed to satisfy the standard for bringing its
lawsuit (EconoSTATS, 2015). If the pleading standard is satisfied, the burden shifts to
the alleged infringing defendant to prove it does not infringe the NPE plaintiff’s patent.
As a result, the defendant may incur substantial discovery costs to determine which of
its products or services are being accused of infringement and how those product or
services infringe the patent (Parker, 2015). To make matters worse, the discovery process
is inexpensive for most NPEs because they typically possess very few documents, because
they do not practice their inventions and do not interact with clients and suppliers with
regard to them (Parker, 2015; Bessen and Meurer, 2014). In addition, the remedy that a
NPE plaintiff must prove is limited, because it most often comprises a mere estimate of
the reasonable royalty for past and future sales (Cotropia, Kesan, and Schwartz, 2017;
Seaman, 2016). Lastly, attorneys on a contingency basis often represent NPEs, and thus
NPEs accrue no legal fees unless they settle or prevail. Patent litigation’s high costs and
elongated timeline incentivize alleged infringers to settle and pay for a license rather than
risk losing their business and paying exorbitant damages.
Arguably, some NPEs contribute nothing to the advancement of society and instead
impede legitimate efforts of innovation by increasing the costs of doing business (Parker,
2015). For alleged infringers, having to pay for licensing royalties and/or litigation costs
increases production costs (Maurer and Haber, 2017). These increases in production
costs are then passed on to consumers in the form of higher prices. This can create a
deadweight loss (i.e., a loss in economic efficiency) in the marketplace. Findings have
indicated that markets reacted negatively to NPE-initiated lawsuits during the period
spanning 1990–2010 (Bessen, Ford, and Meurer, 2012). However, a more recent study
indicates that the market has grown to better understand and anticipate NPE behavior,
and consequently today NPE-initiated lawsuits have very minor effects on the market
(Giudici and Blount, 2017). It is not necessarily NPEs themselves that are troublesome,
but rather the spillover effects their behavior has on business, innovation, and the
economy (Chien, 2013).
2
But see Jason Rantanen, “Slaying the Troll: Litigation As an Effective Strategy Against
Patent Threats”, 23 Santa Clara Computer & High Tech. L.J. 159 (2006) (discussing reasons why an
alleged infringer might choose litigation over settlement).

V. CURRENT PROPOSED SOLUTIONS TO NON-PRACTICING

ENTITIES’ ABUSIVE BEHAVIOR
Taking these facts into account, it is clear that some degree of reform is needed to address
NPE behavior that is detrimental to the patent system as a whole. While changes are most
likely to come in the form of legislation or judicial decisions, the private sector may also
offer viable solutions.
A. Recent Efforts to Reduce Abusive Non-practicing Entities’ Behavior
1. The American Invents Act of 2011

Part of the reason why there has been delayed action toward addressing NPEs’ abusive
behavior emanates from the unforeseen negative effects that have arisen out the passing of
the AIA in 2011 (Howell, 2015). For instance, the PTAB post-grant proceedings of IPR,
Covered Business Method Review, and Post Grant Review, established under the AIA,
were meant to serve as a faster and less expensive alternative to federal court litigation that
would allow anyone, including cash-strapped individual inventors and small businesses, to
challenge the validity of a patent’s claims. However, allowing anyone to challenge a pat-
ent’s claims has unintentionally opened the door to abuse of the patent system, and thus
led to further instances of concerning behavior by NPEs such as Hayman Capital (Laird,
2015). Additionally, because the standards of review in the PTAB are more favorable to
petitioners than the standards applied in district court, filing a petition for one of the
post-grant proceedings creates a substantial incentive for settlement among parties.
The AIA also implemented a change in joinder rules for patent litigation by preventing
patent plaintiffs from joining multiple defendants in a single litigation based solely on
the allegation that all of the defendants infringe the same patent (35 U.S.C. §299). This
change was meant to prevent NPEs from filing baseless lawsuits by forcing them to incur
the significant cost of litigating multiple suits, instead of just one lawsuit lumping all
possible infringers together (Shen, 2014). However, the change does not appear to have
achieved its intended result because NPEs have filed individual suits against each separate
defendant and then sought to have all of the cases consolidated back into one lawsuit
(Cotropia, Kesan, and Schwartz, 2017). Thus, the change in joinder rules seems to have
had minimal success in inhibiting abusive litigation by NPEs, and instead has created
unintended consequences such as increased forum shopping and inefficiencies in pre-trial
activity (Quinn Emanuel, 2017). Therefore, the history of implementation of the AIA
suggests that any changes made in the patent system to address NPEs must be carefully
tailored to only eliminate the detrimental behavior of NPEs, and not inadvertently burden
all patent owners or stifle innovation.
2. Elimination of Federal Rules of Civil Procedure Form 18

Prior to December 1, 2015, it was tremendously easy to file a patent complaint in federal
court (Harmon, 2015). While all other litigants were bound by the 2007 Supreme Court
decision in Bell Atlantic Corp. v. Twombly, 550 U.S. 544 (2007), which raised the standard
for pleading in civil cases to contain “enough facts to state a claim for relief that is
plausible on its face,” this precedent was mostly ignored in patent litigation. Instead, the
Federal Circuit continued to require that a plaintiff only fill out Form 18, a template

Patent trolls 457
originally written in 1938 for patent complaints that is included in the appendix to the
Federal Rules of Civil Procedure, in order to satisfy the pleading standards. Form 18 only
required the plaintiff to list the patent(s) that had allegedly been infringed and the defend-
ants it alleged had infringed these patents. The plaintiff was not required to specify which
specific claims of the patent had allegedly been infringed or which of the defendant’s
products or services were the subject(s) of the suit. Thus, a plaintiff needed only name the
defendant and a patent number in the complaint in order to “get in the courtroom door.”
This exception was incredibly valuable for NPEs because they were able to inexpensively
file barebones complaints and then shift the expenses and burdens onto defendants. The
defendants would then have to go through discovery in order to determine the substance
of the allegations. Therefore, this created a considerable incentive for defendants to settle
instead of trying to disprove infringement or challenge the validity of the asserted patent.
However, on December 1, 2015, Form 18 was eliminated when the Judicial Conference
of the United States instituted its changes to the Federal Rules of Civil Procedure, notably
removing the appendix of forms (Harmon, 2015). Consequently, plaintiffs suing for
patent infringement will now likely have to adhere to Twombly and be held to the same
standard as litigants from other areas of law. Apparently attributable to this change taking
effect, there was a large increase in patent lawsuits filed in November 2015, many of which
were filed by NPEs. On November 30, 2015, the last day before Form 18 was eliminated,
a new record was established for the most suits filed in a single day by NPEs. While this
change may cause a decrease in abusive litigation practices by NPEs, it will be interesting
to observe how federal district court judges interpret these changes. It is plausible that the
elimination of Form 18 may not resolve the issue entirely, and instead will contribute to
forum shopping for district courts that construe these changes in a NPE-friendly manner.
3. Proposed Congressional legislation

The legislation aimed at curtailing the abusive behavior of NPEs must be enacted with
careful consideration and precision. It is imperative that Congress not shortsightedly
enact legislation that will have an adverse effect on the entire patent system, thereby
unintentionally frustrating innovation. There have been a number of recent legislative
proposals brought to Congress to address NPEs, including the Innovation Act (H.R. 9,
114th Cong. (2015)), the Targeting Rogue and Opaque Letters (TROL) Act (H.R. 2045,
114th Cong. (2015)), the Support Technology and Research for Our Nations Growth
Patents (STRONG Patents) Act (S. 632, 114th Cong. (2015)), the Support Technology
and Research for Our Nation’s Growth and Economic Resilience Patents (STRONGER
Patents) Act (S. 1390, 115th Cong. (2017)), and the Protecting American Talent and
Entrepreneurship (PATENT) Act (S. 1137, 114th Cong. (2015–2016)). Among the sug-
gested mechanisms of regulating NPEs’ conduct were: (1) heightening the standard of
specificity needed in patent infringement complaints (the Innovation Act, the STRONG
Act, the PATENT Act); (2) shifting litigation costs to the plaintiff if the suit is unreason-
able (the Innovation Act, the PATENT Act); (3) regulating patent-related demand letters
(the Innovation Act, the TROL Act, the STRONG Act); (4) revising the PTAB post-grant
proceedings (the STRONG Act, the STRONGER Act); and (5) other means (e.g., the
TROL Act and the STRONGER Act propose to preempt all state laws relating to the
assertion of patent rights and authorize the Federal Trade Commission and/or state
attorney generals to bring actions to stop any abusive behavior (H.R. 2045 §§3–4)). The

bills intend to limit the abusive behavior of NPEs, but they may also impose barriers for
all patent litigants (Iancu, and Wells, 2015).
Some legislators question whether all of the proposed changes are needed in light of
recent Supreme Court cases such as Alice Corp. v. CLS Bank, 134 S Ct. 2347 (2014),
which “raised the bar for patentability and enforcement of software patents.” The effects
of the Alice decision have apparently already begun to be experienced, with a recent study
suggesting that the case was the reason for a 13 percent reduction in the number of patent
lawsuit filings from 2014–15. Further, the Supreme Court’s decision in TC Heartland v.
Kraft Foods Group Brands, 137 S. Ct. 1514 (2017) may have a substantial effect on the
behavior of many NPEs. There, the Court interpreted the term “residence” as used in
the patent venue statute, 28 U.S.C. § 1400(b), to mean the state of incorporation of a
domestic corporation, reversing the Federal Circuit’s interpretation encompassing “any
district in which the corporation is subject to personal jurisdiction.” This decision may
prevent NPEs from filing their suits in the district courts of their preference—for example,
the U.S. District Court for the Eastern District of Texas (Chien and Risch, 2018; Marino
and Nguyen, 2014). Prior to the decision in TC Heartland, Eastern District of Texas had
been “wildly popular with patent plaintiffs,” accounting for over one-third of all patent
suits in recent years (Love and Yoon, 2017). The apparent reason for this was a number
of procedural advantages that this particular district court offered, including a fast docket
that “allow[ed] plaintiffs to recover damages faster while placing greater pressure on
defendants to settle.”
4. Patent insurance coverage

Despite congressional legislation likely being the most effective way to resolve detrimental
NPE practices, the private sector also presents viable solutions. In recent years, some
patent infringement risk has been mitigated and managed by insurance coverage (Loeber,
Saka, and Shimshoni, 2015). The policies can vary greatly in scope and price, and—like
any form of insurance—the products are only as good as the policy forms on which they
are written. For instance, some policies insure the claims themselves; others cover the costs
of patent litigation; and some provide indemnification against settlements and verdicts.
However, patent insurance is still in the early stages of development, so the market of
carriers is very limited and the cost of premiums is incredibly high. Nonetheless, one of
the driving forces behind the emergence of patent insurance is to offer a focused approach
to handling NPEs. Companies such as Intellectual Property Insurance Services Corp. and
RPX Corp. already recognize the growing risk and financial toll associated with NPEs,
and offer insurance policies specifically designed to protect against NPE threats and
litigation.
Even though the policy premiums are expensive, patent insurance offers a potentially
valuable solution to entities with limited resources because it counteracts abusive NPEs
using the threat of costly litigation to extort alleged infringers. Additionally, the growth
of patent insurance may lead to a decrease in the number of “detrimental” NPE practices
because “beneficial” NPEs, such as individual inventors and start-up companies, will no
longer have to sell their patents to large patent aggregators or patent holding companies in
order to shift the risk of having their patents challenged by wealthier and more established
entities. Further, prospective policyholders must undergo a long and thorough application
and underwriting process. Therefore, NPEs with overly broad patents will likely be denied

Patent trolls 459
coverage themselves. Ultimately, increased adoption of patent infringement insurance

could completely neutralize the threat and cost of litigation, thereby eliminating arguably
the strongest and most abusive tactical weapon that NPEs possess.
VI. CONCLUSION
It is incorrect to conclude that NPEs are, across the board, detrimental to the patent
system. The presence and prevalence of NPEs are not recent developments. The techno-
logical contributions of NPEs have enabled some of society’s greatest inventions to reach
the marketplace and be readily available, thereby accomplishing the underlying principles
for which the patent system was established. However, there are types of NPE behaviors,
such as the extensive use of cease and desist letters, that raise significant concerns. Thus,
there exists a need for reform in the patent system that will deter disruptive and abusive
NPE behavior without also unintentionally burdening all patent owners. Further, any
changes adapted must be careful not to eliminate those NPEs that provide legitimate and
measurable benefits to the patent system. In addressing the problems made by NPEs,
a careful balance must be struck in order to enforce patent rights commensurate with
innovation.
REFERENCES
Allison, John R., Lemley, Mark A., and Schwartz, David L. 2018. “How Often Do Non-Practicing Entities Win
Patent Suits?”, 32 Berkeley Tech. L.J. 237.
Ball, Gwendolyn G. and Kesan, Jay P. 2009. “Transaction Costs and Trolls: Strategic Behavior by Individual
Inventors, Small Firms and Entrepreneurs in Patent Litigation”, University of Illinois Law & Economics
Research Paper No. LE09-005. Available at http://ssrn.com/abstract=1337166.
Bessen, James E., and Meurer, Michael J. 2014. “The Direct Costs from NPE Disputes”, 99 Cornell Law Review
387–424.
Bessen, James E., Ford, Jennifer Laurissa, and Meurer, Michael J. 2012. “The Private and Social Costs of Patent
Trolls”, 34(4) Regulation, 26, Winter 2011–2012. Available at https://ssrn.com/abstract=1982139.
Burt, Scott. 2015. “It’s Time To Stand Up To Patent Trolls!”, World Intellectual Property Organization Magazine,
www.wipo.int/wipo_magazine/en/2015/01/article_0002.html.
Chien, Colleen V. 2009. “Of Trolls, Davids, Golaiths, and Kings: Narratives and Evidence in the Litigation of
High-Tech Patents”, 87 North Carolina Law Review 1571–615.
Chien, Colleen V. 2010. “From Arms Race to Marketplace: The Complex Patent Ecosystem and Its Implications
for the Patent System”, 62 Hastings Law Journal 297–355.
Chien, Colleen V. 2013. “Patent Trolls by the Numbers”, Santa Clara University Legal Studies Research Paper
No. 08-13. Available at https://ssrn.com/abstract=2233041.
Chien, Colleen V., and Risch, Michael. 2017. “Recalibrating Patent Venue”, 77 Maryland Law Review 47.
Cotropia, Christopher A. 2009. “Individual Inventor Motif in the Age of the Patent Troll”, 12 Yale Journal
Cotropia, Christopher A., Kesan, Jay P. and Schwartz, David L. 2014. “Unpacking Patent Assertion Entities
(PAEs)”, 99 Minnesota Law Review 649–703.
Cotropia, Christopher A., Kesan, Jay P. and Schwartz, David L. 2018. “Heterogeneity Among Patent Plaintiffs:
An Empirical Analysis of Patent Case Progression, Settlement, and Adjudication”, 15(1) Journal of Empirical
Economist. 2015. “Patent Trolls: Why no one likes them”, www.economist.com/news/business-and-finance/21645604.
EconoSTATS at George Mason University. 2015. “Who Are Patent Trolls and What Will H.R. 9 Do About
Them?”, Forbes, www.forbes.com/sites/econostats/2015/05/29/who-are-patent-trolls-and-what-will-h-r-9-do-ab
out-them/#635899f21d97.
Electronic Frontier Foundation. 2017. “Patent Trolls”, www.eff.org/issues/resources-patent-troll-victims.

Fisher, Daniel. 2015. “Are Short-Sellers Really making Money Off New Patent-Review Law? Not Yet”, Forbes,
www.forbes.com/sites/danielfisher/2015/10/24/are-short-sellers-really-making-money-off-new-patent-review-
law-not-yet/#5ea055ad5cd9.
Giudici, Emiliano, and Blount, Justin Robert. 2017. “Evaluating Market Reactions to Non-Practicing Entity
Litigation”. Available at https://ssrn.com/abstract=2928367.
Gugliuzza, Paul R. 2015. “Patent Litigation Reform: The Courts, Congress, and the Federal Rules of Civil
Procedure”, 95 Boston University Law Review 279–301.
Gulfo, Joseph. 2015. “Hedge Funds, ‘Reverse Trolls’ Crushing Biopharma Innovation”, CNBC Biotech and
Pharmaceuticals, www.cnbc.com/2015/07/22/biopharma-hammered-by-hedge-funds-reverse-trolls-commentar
y.html.
Gupta, Kirti, and Kesan, Jay P. 2016. “Studying the Impact of eBay v. MercExchange on Injunctive Relief in
Patent Cases”, fig. 5. University of Illinois College of Law Legal Studies Research Paper No. 17-03, http://
papers.ssrn.com/sol3/papers.cfm?abstract_id=2816701.
Hagiu, Andrei, and Yoffie, David. 2013. “The New Patent Intermediaries: Platforms, Defensive Aggregators,
and Super-Aggregators”, 27 Journal of Economic Perspectives 45–66.
Harmon, Elliot. 2015. “Form 18 is Dead. What’s Next for Patent Trolls?”, Electronic Frontier Foundation, www.
eff.org/deeplinks/2015/12/form-18-dead-whats-next-patent-trolls.
History.com. 2017. “November 05, 1895: George Selden Patents Gas-Powered Car”, www.history.com/this-day-
in-history/george-selden-patents-gas-powered-car/print.
Howell, Kellan. 2015. “Patent Law Reform Bill to Stop Trolls Could Also Stifle Innovation, Critics Say”, Washington
Times, www.washingtontimes.com/news/2015/feb/5/patent-law-reform-bill-to-stop-trolls-could-stifle/?page=all.
Iancu, Andrei, and Wells, Maclain. 2015. “2015 Patent Reform: Is It Happening?”, 22(13) Westlaw Journal
Intellectual Property 1–5.
Jeruss, Sara, Feldman, Robin, and Walker, Joshua. 2012. “The America Invents Act 500: Effects of Patent
Monetization Entities on US Litigation”, 11 Duke Law & Technology Review 357–89.
Krogstad, Jens. 2012. “Universities Struggle With Falling Invention Royalties”, USA Today, www.usatoday.
com/story/money/business/2012/11/25/universities-struggle-with-falling-invention-royalties/1725591/.
La Belle, Megan M. 2014. “Against Settlement of (Some) Patent Cases”, 67 Vanderbilt Law Review 375–441.
Laird, Lorelei. 2015. “Reverse Trolling: Patent Holders Allege Financial Companies Are Misusing New Post-
Grant Review Process for Profit”, December ABA Journal 2015.
Lee, Edward. 2015. “Patent Trolls: Moral Panics, Motions in Limine, and Patent Reform”, 19 Stanford
Lemley, Mark A. 2008. “Are Universities Patent Trolls?”, 18 Fordham Intellectual Property, Media and
Lemley, Mark A., and Melamed, A. Douglas. 2017. “Missing the Forest for the Trolls”, 113 Columbia Law Review
2117–90.
Loeber, Christopher C., Saka, Joseph, and Shimshoni, Hilla. 2015. “Beware of Patent Trolls: Insurance
Coverage For Infringement Lawsuits Brought By Non-Practicing Entities”, 26(5) Westlaw Journal Insurance
Coverage 2–14.
Love, Brian J. 2013. “An Empirical Study of Patent Litigation Timing: Could a Patent Term Reduction
Decimate Trolls Without Harming Innovators?”, 161 University of Pennsylvania Law Review 1309–59.
Love, Brian J., and Yoon, James. 2017. “Predictably Expensive: A Critical Look at Patent Litigation in the
Eastern District of Texas”, 20 Stanford Technology Law Review 1–37.
Marino, Fabio E., and Nguyen, Teri H.P. 2014. “Has Delaware Become the “New” Eastern District of Texas?
The Unforeseen Consequences of the AIA”, 30 Santa Clara High Technology Law Journal 527–53.
Maurer, Noel, and Haber, Stephen. 2017. “Patent Trolls or Patent Elves? Evidence from Publicly-Traded Patent
Assertion Entities”, Hoover Institute Working Group on Intellectual Property, Innovation, and Prosperity,
Stanford University. Working Paper No. 17003.
Mossoff, Adam. 2014. “Thomas Edison Was a ‘Patent Troll’: Patent litigation isn’t nearly as new as people think it
is”, SLATE MAG, www.slate.com/articles/technology/history_of_innovation/2014/05/thomas_edison_charles_
goodyear_and_elias_howe_jr_were_patent_trolls.html.
Mossoff, Adam. 2015. “Repetition of Junk Science & Epithets Does Not Make Them True”, IPWatchDog,
www.ipwatchdog.com/2015/11/19/repetition-of-make-them-true/id=63302/.
Nocera, Joe. 2015. “The Patent Troll Smokescreen”, New York Times, www.nytimes.com/2015/10/24/opinion/
the-patent-troll-smokescreen.html?_r=0.
Osenga, Kristen J. 2015. “Formerly Manufacturing Entities: Piercing the Patent Troll Rhetoric”, 47 Connecticut
Parker, Clifton B. 2015. “Patent Trolls Serve Valuable Role in Innovation, Stanford Expert Says”, Stanford News.
Available at http://news.stanford.edu/news/2015/february/haber-patent-trolls-022315.html.
Pegoraro, Rob. 2015. “Why the Tech Industry Hates Patent Trolls, and You Should Too”, Yahoo! Tech, www.
yahoo.com/tech/why-the-tech-industry-hates-patent-trolls-and-you-121628489339.html.

Patent trolls 461
Pérez-Peña, Richard. 2013. “Patenting Their Discoveries Does Not Pay Off for Most Universities, a Study
Says”, New York Times, www.nytimes.com/2013/11/21/education/patenting-their-discoveries-does-not-pay-
off-for-most-universities-a-study-says.html.
Pham, Michael. 2014. “The Cost of Defending Against Patent Trolls”, WinTech, www.wintechblog.com/2014/07/
the-cost-of-defending-against-patent-trolls/.
Pogue, Melissa. 2015. “Under the Bridge: Patent Trolls and the Toll on Innovation”, Martin Prosperity Institute,
http://martinprosperity.org/content/under-the-bridge-patent-trolls-and-the-toll-on-innovation/.
Quinn, Gene. 2015a. “Inventors are NOT Patent Trolls and They are NOT the Problem”, IPWatchDog, www.
ipwatchdog.com/2015/03/16/inventors-are-not-patent-trolls-and-they-are-not-the-problem/id=55752/.
Quinn, Gene. 2015b. “Patent Abuse or Genius? Is Kyle Bass Abusing the Patent System?”, IPWatchDog, www.
ipwatchdog.com/2015/04/08/is-kyle-bass-abusing-the-patent-system/id=56613/.
Quinn Emanuel Urquhart & Sullivan, LLP. 2017. “Multi-Defendant Joinder Under the America Invents Act:
Much Ado About Nothing?”, www.quinnemanuel.com/the-firm/news-events/article-december-2012-multi-de
fendant-joinder-under-the-america-invents-act-much-ado-about-nothing/.
Rantanen, Jason. 2006. “Slaying the Troll: Litigation as an Effective Strategy Against Patent Threats”, 23 Santa
Clara Computer & High Technology Law Journal 159–210.
Rice, Kelly. 2013. “How Long Does it Take to Build a Mobile App?”, Kinvey, www.kinvey.com/blog/2086/
how-long-does-it-take-to-build-a-mobile-app.
Risch, Michael. 2012. “Patent Troll Myths”, 42 Seton Hall Law Review 457–97.
Schwartz, David L. 2014. “On Mass Patent Aggregators”, 114 Columbia Law Review Sidebar 51–69.
Schwartz, David L., and Kesan, Jay P. 2014. “Analyzing the Role of Non-Practicing Entities in the Patent
System”, 99-2 Cornell Law Review 425–56.
Seaman, Christopher B. 2016. “Permanent Injunctions in Patent Litigation After eBay: An Empirical Study”,
Self, Laurie. 2015. “Misleading Patent Troll Narrative Driven by Anecdote, Not Facts”, IPWatchDog, www.
ipwatchdog.com/2015/11/12/misleading-patent-troll-narrative-driven-by-anecdote-not-facts/id=63122/.
Shen, Dongbiao. 2014. “Misjoinder or Mishap? The Consequences of the AIA Joinder Provision”, 29(4)
Sternburg, Ali. 2013. “Obama Acknowledges Patent Troll Problem [w/transcript]”, Patent Progress, www.patent
progress.org/2013/02/14/obama-acknowledges-patent-troll-problem-w-transcript/.
Stoll, Robert L. 2014. “Patent Trolls: Friend or Foe?”, World Intellectual Property Organization Magazine.
Available at www.wipo.int/wipo_magazine/en/2014/02/article_0007.html.
Taylor, David O. 2013. “Patent Misjoinder”, 88 New York University Law Review 652–728.
Case List
Alice Corp. v. CLS Bank (2014), 134 S Ct. 2347.

Bell Atlantic Corp. v. Twombly (2007), 550 U.S. 544.
eBay v. MercExchange (2006), 547 U.S. 388.
TC Heartland v. Kraft Foods Group Brands (2017), 137 S. Ct. 1514.
U.S. Const. art. 1, sec. 8, clause 8.

35 U.S.C. § 299 (2012).
The Leahy-Smith American Invents Act, Pub. L. No. 112-29, 125 Stat 284 (2011).
H.R. 9, 114th Cong. (2015).
H.R. 2045, 114th Cong. (2015).
S. 632, 114th Cong. (2015).
S. 1137, 114th Cong. (2015–2016).
S. 1390, 115th Cong. (2017).

20. Patents and innovation in economic history
Petra Moser* 3
Contents
I. Introduction
II. Intellectual Property Rights as a Source of Economic Development
III. A Continuing Search for Data on Innovation
IV. Observing Innovations With and Without Patents: Historical Technology Fairs
V. Many Important Innovations Occur Outside of the Patent System
VI. Patent Laws were not Necessary for Innovation
VII. Even with Low Fees, Most Innovations were not Patented
VIII. Similar to Today, Litigation Risks were Extremely Damaging
IX. Patent Pools Discouraged Innovation
X. Compulsory Licensing Encouraged Innovation
XI. Patents have been Useful in Some Industries – but not in Others
XII. Intellectual Property Rights for Living Organisms have not Encouraged
Innovation
XIII. Patent Laws have Changed the Direction of Innovation
XIV. Patent Laws have Changed the Geography of Innovations
XV. Narrow Intellectual Property Rights—Through Copyrights—as a Promising
Alternative
XVI. Conclusions and Directions for Future Research
References
I. INTRODUCTION
Innovation is commonly viewed as the primary driver of sustained improvements in

human welfare and economic growth. The creation of intellectual property institutions,
particularly patents, has been motivated by a desire to stimulate innovation, and support
economic growth. Yet the connections between patents and innovation are difficult to
identify empirically, and the predictions of economic theory are somewhat ambiguous.
For example, basic models of patent laws (in the spirit of Nordhaus, 1969) predict that
* Petra Moser (pmoser@stern.nyu.edu) is an economist who studies creativity and innovation

at New York University, and is a Research Associate at the National Bureau of Economic Research
and the Centre for Economic Policy Research. This chapter was originally prepared for the Annual
Review of Economics. Professor Moser’s research has been supported by the National Science
Foundation through CAREER Grant 1151180, through a Fellowship at the Center for Advanced
Study, and a grant from the Institute for New Economic Thinking Institute on “Copyright and
Creativity.”
462

Patents and innovation in economic history 463
extensions in the length of patents encourage innovation by granting inventors stronger

property rights in their ideas. Yet alternative models suggest that stronger patents can
discourage innovation if they reduce the payoffs to later innovators who rely on previous
inventions as an input for their work (Scotchmer, 1991; Bessen and Maskin, 2007).
A strong tradition in economic history, which primarily relies on qualitative evidence
and statistical correlations, has emphasized the importance of patents as a primary driver
of innovation. Recent improvements in empirical methodology—through the creation
of new data sets and advances in identification—have produced research that challenges
this traditional view. The findings of this literature provide a more nuanced view of the
effects of intellectual property, and suggest that when patent rights have been too broad
or strong, they have actually discouraged innovation. This chapter summarizes the major
results from this research and presents open questions.
Economic history plays an important role in research on patent policies because it
provides some of the best data, as well as credibly exogenous sources of policy variation.
In the nineteenth century, for example, several countries failed to adopt patent systems
for significant periods of time, or abolished existing patent laws for political reasons that
were unrelated to innovation. In addition, the world technology fairs of the nineteenth
century (Moser, 2005, 2011, and 2012) present opportunities to examine innovations that
were made within and outside of patent systems across countries. The catalogs of these
exhibitions enable researchers to observe the rate and direction of technical change in
countries without patents, as well as the rate at which innovators use patents when they
are available. Both are critical for understanding the effects of patents laws on innovation,
and represent a major advantage over much of the literature on modern innovation, which
relies primarily on patent data.
The lens of history also allows researchers to examine changes in policy from a distance,
which often leads to a more complete understanding of different aspects of intellectual
property institutions. This is particularly useful because the design of a well-functioning
patent system depends on understanding the trade-offs associated with many different
components of patent policy, including some that are exceedingly difficult to analyze
with modern data. For example, a key issue in international debates relates to compulsory
licensing, which allows developing countries to license foreign-owned patents to domestic
firms—without the consent of foreign patent owners. To devise effective policies to govern
compulsory licensing, policy makers must weigh the benefits of compulsory licensing
for consumers and firms in developing countries against the potential costs of reducing
innovation incentives for foreign firms whose intellectual property rights are violated.
Another example of such trade-offs are patent pools, which allow competing firms
to combine their patents and license them as a bundle to outside non-member firms.
Regulators who decide whether or not to allow these arrangements must evaluate the
benefits that pools provide for their members against the potentially negative and
dynamic effects of pools on future innovation. Pools benefit members and consumers by
minimizing litigation risks and facilitating the commercializations of inventions for which
competing firms own mutually overlapping patent rights. Yet they may also discourage the
creation of new technologies by allowing pool members to cooperate (instead of compet-
ing to improve). Moreover, the existence of a pool may increase litigation risks for outside
firms, which makes it harder for them to compete with the pool. These effects have thus
far been ignored by theoretical models, but they feature prominently in historical data.

More generally, analyses of intellectual property institutions must consider alternative

mechanisms, such as secrecy, which inventors can use instead of patents. With a few
notable exceptions (e.g., Anton and Yao, 2004) theories of intellectual property rights
have ignored these alternatives. Yet their existence allows inventors to opt out of the patent
system, which means that changes in patent policies will not have the predicted effects.
In part, economic analyses have ignored alternative mechanisms, such as secrecy, because
they are intrinsically more difficult to observe. This makes it difficult to evaluate their
importance, even though we know intuitively that they exist. Notably, data from world
technology fairs indicate that, historically, the great majority of innovations have been
created outside of the patent systems (e.g., more than 80 percent of exhibits at the 1851
Crystal Palace Great Exhibition).
II. INTELLECTUAL PROPERTY RIGHTS AS A SOURCE OF

ECONOMIC DEVELOPMENT
The origins of intellectual property rights reach back as far as 1474, when the Venetian
Republic began to offer exclusive rights to entrepreneurs who had invented or imported
a new technology. The Venetian law was purposefully agnostic about the source of an
innovation; Venetian rulers were happy to attract skilled artisans with new inventions,
whether they had created that invention or not. Other European countries introduced
modified versions of the Venetian system, hoping to promote economic development.
But from the beginning, this system proved vulnerable to exploitation (David, 1994).
For example, British monarchs up to King James I used the grant of royal monopolies
on technical inventions as a mechanism to extract rents that financed wars or courtly
estates. Britain’s Statute of Monopolies of 1624 replaced the royal monopolies with
a system of legal property rights in ideas. North and Thomas (1973) argue that the
creation of this intellectual property institution helped trigger Britain’s Industrial
Revolution. Well-functioning property rights in ideas allow inventors to keep a large
enough share of the fruits of their labor, motivating them to invest in creating socially
valuable inventions.
This ideal motivated the creation of “the world’s first modern patent institution”
(Sokoloff and Khan, 2001), the US patent law. Thus, Article 1 of the US Constitution
instructs Congress to “promote the Progress of Science and useful Arts, by securing for
limited Times to Authors and Inventors the exclusive Right to their respective Writings
and Discoveries.” In contrast with the early British system, the US patent system com-
bined the twin goals of creating inventions and encouraging their widespread adoption.
Importantly, the United States also introduced a system of patent examinations in 1836
(Khan, 2005). This system charged patent examiners (ideally technical experts in their
field) with the responsibility to establish the novelty of each invention and—at least in
principle—refuse patents for inventions covered by previous patents.
European countries began to copy features of the US patent system after they observed
American innovations at the Crystal Palace fair in London in 1851. Until the Great
Exhibition of the Works of Industry of all Nations (as the Crystal Palace fair was officially
called), European observers had ridiculed the United States as a technological backwater.
But soon after the fair had opened, droves of visitors began to admire US innovations

such as the McCormick grain reaper, which promised enormous productivity gains in
agricultural production. Another US invention, Samuel Colt’s revolver, demonstrated
the potential of the American system of manufacturing, in which the components were
made with such uniformity and precision that they could be treated as interchangeable
parts (Hounshell, 1985).
Soon afterwards, Britain’s Parliament established a commission to investigate the
US system of manufacturing (Rosenberg, 1969). In his 1855 report on New York’s own
Crystal Palace exhibition, which was held on the ground of the current Bryant Park
between July 1853 and November 1854, Commissioner Joseph Withworth (1854) included
a detailed description of the US patent system, emphasizing low fees:
It remains to say a few words, in conclusion, on the scale of fees. These are extremely moderate in
the case of citizens, and the privileges and protection afforded by letters patent are most properly
placed within the reach of all citizen inventors (Withworth, 1854, p. 67)
Since then, economic historians have credited low fees with encouraging broad-based
innovation in the United States (Khan and Sokoloff, 1993, 1998, 2001; Lamoreaux and
Sokoloff, 1999; Khan, 2005).
Notably, Withworth (1854) also describes the dramatically higher fees that the US
system levied on foreign, and especially British, inventors: “But foreigners labour under
great disadvantages, being required, if British subjects, to pay nearly seventeen times as
much as a citizen ($500), if any other foreigner, ten times as much as a citizen ($300).”
These fees are interesting for two reasons, which invite empirical analyses of their
effects. First, expensive fees probably discouraged foreigners from patenting their inven-
tions in the United States, which made it easier for US nationals to copy them. In spirit,
this is reminiscent of modern policies that weaken the patent rights of foreign inventors,
such as compulsory licensing (discussed below). Second, it is remarkable that the US
system reserved the largest fees for inventors from Britain, who were at the technology
frontier. Taken together these features suggest that the early US patent system may have
encouraged economic development by reducing the costs of copying foreign inventions.
III. A CONTINUING SEARCH FOR DATA ON INNOVATION
Following the pioneering analyses of hand-collected data in Schmookler (1962 and

1966), and Sokoloff (1988), patent counts have become the standard proxy for innova-
tion. In recent years, the availability of electronic data sets, such as the National Bureau
of Economic Research’s Patents and Citations dataset for 1976–2002 (Jaffe, Hall, and
Trajtenberg, 2001) and optical character recognition files of US patents since 1920, has
further contributed to the popularity of patents as a focus for research.1
Yet it is important to keep in mind that patents are an “imperfect, fallible measure”
of the “net accretion of economically valuable knowledge” (Griliches, 1990). This is
1
See Moser and Voena (2012), for a detailed description of Google/US Patent and Trademark
Office patent data since 1920, Moser et al. (2014) for inventor-level data, and Lampe and Moser
(2016) for historical data on patent citations since 1920.

articularly important for analyses of patent laws because we need to understand what
p
types of innovations—if any—would be made in the absence of patent laws. By construc-
tion, patent data fail to capture these innovations.
Studies of individual industries, such as Allen (1983) and Nuvolari (2004), have
addressed this issue by collecting independent measures of performance. That is, rather
than focusing on counts of patents or their citations, this research has attempted to
directly measure improvements of particular products. For example, Lampe and Moser
(2010) use data on sewing speeds to observe the rate of innovation in sewing machines.
Similarly, Moser, Ohmstedt, and Rhode (2014) compile data on advances in yields, and
other biological characteristics of patented corn hybrids, to measure the speed of biologi-
cal innovation. Yet such performance measures are impossible to construct in a manner
that makes them comparable across industries. This feature limits their use in empirical
analysis of the direction of technical change, and other important questions.
IV. OBSERVING INNOVATIONS WITH AND WITHOUT

PATENTS: HISTORICAL TECHNOLOGY FAIRS
Historical data on exhibits at world fairs make it possible to observe innovations—

irrespective of their patent status—across sectors, across countries, and over time.
Starting with the Crystal Palace fair in London in 1851, the organizers of these fairs
systematically collected and displayed technologies from across the world (Moser, 2003,
2005, 2011, and 2012). The scale of the fairs was unprecedented, and remains unparalleled
today. In 1851, Crystal Palace—a 1848-foot-long greenhouse of cast iron and glass—was
the largest enclosed space on earth; it housed 17 062 exhibitors from 40 countries; at a time
when London had fewer than 2 million inhabitants, more than 6 million tickets were sold.
The American Centennial Exhibition in Philadelphia in 1876 attracted more than 10 mil-
lion visitors (Kroker, 1975). To see all 10 863 exhibits from 35 countries at the Centennial
fair, a visitor had to walk the equivalent of a marathon. Another important fair, the
1893 World’s Columbian Exposition in Chicago, covered 717 acres in Jackson Park by
Lake Michigan; it attracted 27.5 million visitors. Finally, to host the Panama-Pacific
International Exposition in San Francisco in 1915, its organizers drained a swampy part
of the San Francisco Bay, which later became the Presidio. This fair welcomed 30 000
exhibitors from 32 countries and 19 million visitors.
Exhibits were recorded in catalogs that helped visitors navigate the fairs. Many of
these catalogs became significant reference works for the current state of technology.
For example, Britain’s Official Catalogue of the great exhibition of the work of industry
of all nations (1851) became a “book of reference to the philosopher, merchant, and
manufacturer” (Auerbach, 1999) to an exhibition that was a “veritable acting industrial
encyclopaedia” (Tallis, 1852; The Times, December 5, 1850). Catalog entries include the
exhibitor’s name, location, and a description of the exhibit: “406, Fourdrinier, E. N., 38
Barclay St., Sunderland, Patent safety apparatus, for preventing loss of life and destruc-
tion of property when a rope or chain breaks in shafts of mines and collieries.” In addition
to the prize-winning reaper and Colt’s revolver, examples of exhibits included power-loom
lathes, and the first sewing machines.
As a means of quality control, exhibits were subject to a systematic and careful process

of selection.2 Commissioners for the Crystal Palace fair visited cities and rural areas to set
up local selection committees and to ensure broad-based representation (Bericht, 1853;
Auerbach, 1999). National papers, such as the British Times, Economist, and Sun, and the
US Scientific American, advertised the fairs in regular features (e.g., The Times, November
14, 20, December 6, 9, and 30, 1850). In Britain, 65 local commissions formed more than
300 sub-commissions (Auerbach, 1999). Local committees selected exhibits in the first
round; their submissions competed at the national level. Among these selections, national
committees chose the most promising innovations to represent their country in a hard-
fought international “contest of industries” (Kretschmer, 1999; Auerbach, 1999). Fewer
than 30 percent of applicants were admitted to the fair (Bericht, 1853). After the fair,
universities, scientists, and museums competed to purchase exhibits (Auerbach, 1999).
A concern with exhibition data is that transportation costs may have influenced the
selection of exhibits, and that exhibits from host countries were over-represented. For
example, three-quarters of the 13 acres in Machinery Hall were taken up with American
machines in 1876, including the colossal Corliss steam engine, which provided power to
the fair (Andrews, 1894). The influence of transportation costs was mitigated because
heavy, large, or fragile inventions were presented by models. In 1851, for example, 88 of
194 British exhibits in “Civil Engineering, Architecture, and Building Contrivance” were
represented by models, including a 770-meter long chain suspension bridge across the
Dnepr in Kiev (Rolt, 1970). Similarly, T. Powell of Monmouthshire, Britain, exhibited a
“Model for apparatus used for shipment of coals from boats or wagons at Cardiff dock”
at Crystal Palace, and A. Watney of Llanelly, Wales, exhibited “Models of anthracite
blast furnaces.”3
V. MANY IMPORTANT INNOVATIONS OCCUR OUTSIDE OF

THE PATENT SYSTEM
Although prominent accounts have emphasized the importance of patents in stimulating

innovation (Sokoloff and Lamoreaux, 1999), exhibition data show that the majority
of nineteenth-century innovations occurred outside of the patent system. In 1851, for
example, only 11 percent of British exhibits were patented. Such low patenting rates are
more consistent with alternative explanations that emphasize a culture of entrepreneur-
ship (Landes 1969), the free exchange of knowledge (Allen, 1983; Nuvolari, 2004), and
2
By comparison, selection processes at smaller regional fairs in the US and Britain and even
at later world fairs were substantially more haphazard (Khan, 2014), and exhibits at these fairs may
not be a good proxy for innovation.
3
Another potential source of bias stems from restrictions on exhibition space, which were set
by the host country. Anecdotal evidence, however, indicates that such restrictions may not have
been binding. For example, when the US commission to the Crystal Palace fair thought that US
exhibitors would be short on exhibition space, it asked the British commission for more room and
was granted its request (Haltern, 1971). Floor plans for the Centennial exhibition show that 11
countries—including Germany, Britain, Spain, and Sweden—built additional exhibition space on
the Centennial grounds.

investments in science (Arora and Rosenberg, 1998; Mowery and Rosenberg, 1998) as
potential sources for innovation.4
In addition to capturing innovation without patents, exhibition data make it possible to
control for the quality of innovations through prizes that were granted to a small share of
exceptional inventions. These data raise a question for empirical and theoretical research:
are inventors more or less likely to patent high-quality inventions? Existing theories of
innovation predict that firms are more likely to use secrecy if innovations are important.
Patents require disclosure, which is risky if patents cannot effectively prevent competitors
from using a patented invention (Anton and Yao, 2004; Horstmann, MacDonald, and
Slivinski, 1985). Exhibition data, however, show that high-quality exhibits were slightly
more likely to be patented. In 1851, 15 percent of Britain’s prize-winning exhibits were
patented, compared with 11 percent of average-quality exhibits. Notably, the prize data
also show that the large majority of prize-winning inventions were not patented, which
calls into question the importance of patents as a driver of high-quality invention.
VI. PATENT LAWS WERE NOT NECESSARY FOR

INNOVATION
The nineteenth century offers a unique setting to examine the effects of patent laws on
innovation because no other period had a comparable amount of exogenous variation
in national laws. In 1851, the statutory length of a patent varied from no protection in
Denmark and Switzerland to 14 years in Britain and the United States, and 15 years
in Bavaria, Belgium, and the Netherlands. Much of this variation was in place because
countries had adopted patent laws according to idiosyncratic allegiances of national
rulers (Penrose, 1951). Another benefit of the empirical setting is that it precedes the
period of intense lobbying from interests that dominate debates today.
Exhibition data show that the existence of patent laws was not a necessary condition
for high levels of innovation (Moser, 2003). Crystal Palace data in Moser (2005, Table
1) include 11 610 exhibits from 12 Northern European countries in 1851. Two of these
countries—Switzerland and Denmark—offered no patent protection. Data for the
Philadelphia Centennial Exhibition cover 6482 exhibits from 10 Northern European
countries in 1876. Two of them—Switzerland and the Netherlands—offered no patent
protection, and three had short-lived patents. Notably, the absence of patent protection
did not prevent innovation in countries without patent laws. In 1851, for example,
Switzerland contributed 110 exhibits per 1 million Swiss citizens, which was twice the
mean of the average country, and three times compared with the median of 36 exhibits
4
Low patenting rates are particularly remarkable considering that exhibition data may be biased
toward recording patented inventions, if inventors were reluctant to exhibit inventors that were pro-
tected by secrecy. To address this problem, the UK government created a registration system to offer
“Protection from Piracy to Persons exhibiting new Inventions.” Only 600 exhibitors (less than 3.3 per-
cent) took advantage of this system. Instead, exhibitors advertised the finished product, rather than
the machines that had made it, to maintain secrecy surrounding the machine. For example, Drewsen
& Sons of Silkeborg, Jutland, exhibited “Specimens of paper, glazed by a machine constructed by the
exhibitor,” instead of their paper-making machine (Official Catalogue 1851, p. 210).

per 1 million people. Switzerland also won a disproportionate number of prizes, with 47
per 1 million people—more than twice the mean value of 20 and more than four times
the median of 11. Switzerland continued to lead innovation in food processing (e.g.,
by improving milk chocolate and ready-made meals), all before adopting a full-fledged
patent system in 1907.
VII. EVEN WITH LOW FEES, MOST INNOVATIONS WERE NOT

PATENTED
Most economists understand intuitively that many important innovations—ranging

from nineteenth-century dyes and scientific instruments to Ford’s conveyer belt and the
famously secretive Coca-Cola—were not patented. Yet it is difficult to determine how
important non-patenting is quantitatively, and whether it varies across sectors and over
time.5 Exhibition data allow researchers to measure such variation across industries, over
time, and—importantly—across patent systems.
In 1851, patent fees in Britain ranged between £100 and £120—“approximately 4 times
per capita income in 1860” (Khan, 2005). In addition, inventors faced a cumbersome
process with legal fees and bribes (MacLeod, 1988; Dutton, 1984). In Charles Dickens’ “A
Poor Man’s Tale of the Patent System” (1868), the fictional inventor Old John complains:
Look at the Home Secretary, the Attorney-General, the Patent Office, the Engrossing Clerk,
the Lord Chancellor, the Privy Seal, the Clerk of Patents, the Lord Chancellor’s Purse-bearer,
the Clerk of the Hanaper, the Deputy Clerk of the Hanaper, the Deputy Sealer, and the Deputy
Chaff-was. No man in England could get a patent for an Indian-rubber band or an iron hopp,
without feeing all of them. Some of them, over and over again.
By comparison, US patent fees were substantially lower. Moreover, inventors could

submit their patent applications by mail, which should have made it much easier for
rural inventors to apply for a patent. Khan and Sokoloff (1998 and 2001) have credited
the superior design of the US patent system with encouraging the “democratization” of
invention in the United States.
The British system reflected its origin in royal privilege and effectively limited access to a select
class of patentees. In contrast, the United States established the world’s first modern patent
institution, one that was consciously designed to stimulate participation in invention across a
wide spectrum of the population (Khan and Sokoloff, 1998)
Yet exhibition data show that US patenting rates were only slightly higher than British
patenting rates; the difference was less than five percentage points (Moser, 2012). These
results challenge the view that the design of the US patent system played an important
role in stimulating high levels of innovation (Lamoreaux and Sokoloff, 1998; Sokoloff
and Khan, 1998, 2001).
5
The most complete evidence in modern settings come from survey data (Levin, Klevorick,
Nelson, and Winter, 1987; Cohen, Nelson, and Walsh, 2000), which suggest that chemicals and
pharmaceuticals relied strongly on patents, while most other industries relied heavily on secrecy.

Location data for inventors, however, provide some suggestive evidence in favor of the
hypothesis that the US patent system helped democratize innovation. A first test compares
variation in patenting rates across urban and rural areas in Britain and the United States.
With the caveat that the data include only 574 observations across the United States, there
are no noticeable systematic differences in patenting rates across urban and rural areas in
the United States (Moser, 2012). By comparison, British patenting rates (for a sample of
more than 6000 exhibits) were significantly higher in London, compared with rural rates.
These findings highlight the need for further analyses to examine the role of access costs
in determining the effectiveness of patents.
VIII. SIMILAR TO TODAY, LITIGATION RISKS WERE

EXTREMELY DAMAGING
A potential explanation for the low rates at which valuable innovations in the United Sates
were patented is that even though the costs of patent applications were low, the costs of
patent enforcement may have been substantial. Dutton (1984) observes that US courts
were more likely to uphold the patent rights of original inventors, while British courts
tended to be more anti-patent. Khan (2005) argues that this pro-patent attitude encour-
aged invention in the United States. The US Patent Act of 1793 specified that “simply
changing the form or proportion of any machine . . . shall not be deemed a discovery.” It
was, however, often left to the courts to determine the boundaries between the original
inventor’s patent and the next invention that was sufficiently novel to merit a new patent.
Eli Whitney, for example, patented a saw gin in 1793 that could be used to remove seeds
from short staple upland varieties of cotton (Olmstedt and Rhode, 2008; Lakwete, 2003).
Whitney’s patent remained subject to litigation until December 1807, one month after it
had expired. The knowledge of such cases may have led inventors to incorporate litigation
risks into the expected costs of patenting, and avoid patents whenever possible.
Litigation risks were particularly high when US courts enforced overly broad patents
that had been issued by the US Patent and Trademark Office (USPTO). In 1846, for
example, the USPTO issued Patent 4750 to Elias Howe for an “Improvement in Sewing
Machines.” Howe’s patent was broad enough to cover most commercially viable sewing
machines at the time. Like a twenty-first century “patent troll,” Howe used his patent to
threaten litigation, instead of commercializing his invention. In 1852, a District Court
upheld Howe’s patent, and he began to collect license fees of $25 per machine—roughly
one-fifth the average price of a sewing machine (Lampe and Moser, 2014). When other
firms sued based on their own patents, production came to a halt during the 1851–56
“sewing machine wars” (Bissell, 1999). By 1867, Howe had received $2 million in license
fees (Parton, 1867) —roughly $30 million in 2014 dollars.
IX. PATENT POOLS DISCOURAGED INNOVATION
To resolve litigation, Elias Howe, the Singer Company, and two other manufacturers
formed the first patent pool, the Sewing Machine Combination, in 1856. Like contempo-
rary pools, the Combination allowed its members to use each others’ patents, and license

all pool patents together as a bundle to outside firms. Litigation data confirm that the
creation of a pool reduced litigation risks for members (Lampe and Moser, 2010). The
pool also reduced license fees from $25 per machine for Howe’s patent to $5 for the bundle
of patents for members and $15 for outside firms. Patenting, however, declined after the
pool formed and increased again only after it dissolved in 1877, with the expiration of the
last broad patent (Lampe and Moser, 2010). A comparison with Britain’s sewing machine
industry, which had no patent pool, suggests that this decline in innovation was limited to
the United States. In Britain, sewing machine patents continued to increase gradually as a
share of all patents until the early 1870s and experienced no increase after 1877.
Did the decline in US patenting represent a decline in innovation? Here again the
difference between patents and innovation is important: patenting may decline because
inventors no longer have to create “patent thickets” (Shapiro, 2001) to protect themselves
from litigation. With a little bit of detective work, economic history can answer this
question. Articles on sewing machines in nineteenth-century magazines, such as the
Scientific American and the Lady’s Home Journal, suggest that contemporary customers
valued sewing machines that were light, quiet, and, most importantly, fast. Sewing speed
is measured as the number of stitches per minute that a machine could perform, and is
reported on contemporary trade cards and catalogs, which we were able to access in the
records of the Smithsonian Institution. These data indicate that improvements in sewing
speed slowed soon after the pool had been established and increased again after it had
dissolved (Lampe and Moser, 2010).
The experience of the sewing machine industry also suggests another important policy
implication of patent pools for the direction of technical change. Data on patents and firm
entry indicate that the creation of a patent pool diverted innovation away from the pool
technology toward a substitute technology that was already understood to be inferior
(Lampe and Moser, 2014). As long as the pool created a significant wedge between license
fees for members and outsiders, outsiders entered with the substitute because they needed
to avoid licensing and litigation. Data on firm survivals show that firms that entered with
the substitute were significantly more likely to fail compared with other entrants, suggest-
ing that the shift toward the substitute technology was socially wasteful.
Whether these results are generalizable to other industries and modern pools is an open
question. A companion study of pools that formed across a broad range of 20 industries
in the 1930s, however, unambiguously confirms the decline in patenting (Lampe and
Moser, 2014). In the absence of effective antitrust, dominant firms used a pool to further
limit and discourage competition by outside firms. We find that this reduction in competi-
tion was associated with a significant decline in innovation.
These results highlight the need for additional empirical and theoretical analyses of
patent pools. Theoretical models of the price effects of pools are well developed (Shapiro,
2001; Lerner and Tirole, 2004), but a pool’s effect on innovation is less well understood.
Importantly, existing theoretical models focus almost exclusively on pool members and
ignore effects on outside firms. Yet outside firms account for more than 90 percent of all
patents in industries with patent pools (Lampe and Moser, 2014), which suggests that
ignoring them can bias theoretical predictions.

X. COMPULSORY LICENSING ENCOURAGED INNOVATION
Another prominent mechanism of intellectual property policies is “compulsory licens-

ing,” which allows government agencies to license foreign-owned patents inventions to
domestic firms—without the consent of foreign patent owners. This policy is a central
concern in international debates. World Trade Organization rules allow developing
countries to use it if negotiations with foreign patent owners fail, and countries such as
Brazil and Thailand have used compulsory licensing to improve access to foreign-owned
pharmaceuticals, including drugs to treat HIV/AIDS.
While it is impossible to predict the effects of such interventions on innovation, histori-
cal episodes of compulsory licensing create opportunities for empirical analysis. In 1917,
the US Trading with the Enemy Act (TWEA) empowered government agencies, such
as the Chemical Foundation, to license German-owned patents to US firms. The Act’s
primary goal was to place all enemy property, including patents, “beyond the control or
influence of its former owners” (Alien Property Custodian, 1919). By February 22, 1919,
the U.S. Alien Property Custodian (1919), A. Mitchell Palmer, announced that “practi-
cally all known enemy property in the United States has been taken over by me.” In 1919,
the US Chemical Foundation began to issue non-exclusive licenses for enemy-owned
patents to US firms.
In one of the first analyses of the USPTO/Google bulk data patent files, Moser and
Voena (2012) show that compulsory licensing of German-owned patents in organic
chemistry led to a 20 percent increase in US invention for affected technologies. Their
baseline estimates compare changes in patent grants per year after 1918 for 336 technolo-
gies with compulsory licensing with changes in patent grants for a control group of 7248
technologies without licensing in the same industry. Lee and Moser (2015) extend these
results to an economy-wide analysis during World War II.
Methodologically, the analysis exploits the USPTO’s detailed classification system to
distinguish narrowly defined technologies (measured at the level of USPTO subclasses)
that were differentially affected by compulsory licensing. Technology fixed effects (at
the level of USPTO subclasses) and year fixed effects, as well as technology-specific
trends, control for variation in the use of patents across technologies and over time. This
approach makes it possible to control for unobservable factors, such as improvements in
education or protectionist tariffs, which may have encouraged US innovation in chemicals
regardless of compulsory licensing.
Firms under threat of compulsory licensing today argue that it will weaken their
incentives to invest in R&D. The historical records, however, suggest the opposite. Baten,
Moser, and Bianchi (2015) collect and analyze firm-level data on German patents to
examine invention by German firms that were differentially affected by compulsory
licensing under the TWEA. This analysis indicates that compulsory licensing was
associated with a 28 percent increase in patenting by German inventors. Controlling
for variation in the quality of patents (through renewal data) indicates that some of the
observed increase was due to an increase in lower-quality patents, possibly as firms built a
“thicket” of strategic patents around technologies that were now threatened by entry. But
even with quality controls, we find that compulsory licensing led to a 17 percent increase
in high-value German patents.

XI. PATENTS HAVE BEEN USEFUL IN SOME INDUSTRIES –

BUT NOT IN OTHERS
Does this mean that intellectual property rights always hold back innovation? Probably
not, but to answer the question economists must understand when inventors use and don’t
use patents. Exhibition data reveal a great deal of variation in the use of patents across
industries and over time. In 1851, for example, fewer than 5 percent of Britain’s chemical
exhibits were patented, 10 percent of scientific instruments, and 8 percent of exhibits in
food processing, compared with 20 percent of manufacturing machinery (Moser, 2012).
The data also show that—despite the most substantial differences in national patent
systems—US inventors appear to have relied on patents, and avoided them in the same
industries as British inventors (Moser, 2012). Echoing results on low overall patenting
rates in the United States, these findings suggest that the institutional features of the US
patent system did not significantly affect patenting behavior.
What determines variation in the use of patents across industries? Qualitative evidence
points to variation in the effectiveness of secrecy, as an alternative to patents. With crude
tools of chemical analysis, secrecy was an effective mechanism to protect mid-nineteenth-
century improvements in chemicals, because competitors could not yet reverse-engineer
improvements. Commercially valuable dyes, such as indigo, proved immune to reverse-
engineering, despite considerable efforts in industrial espionage until the late nineteenth
century (Haber, 1958). Secrecy was also effective in protecting improvements in scientific
instruments, such as the rectangular prisms of Swiss glassmakers and the optical instru-
ments of Danish makers (Bericht I, 1852). Scientific breakthroughs, such as Dmitrii
Mendeleev’s publication of the periodic table in 1869, increased the risks associated
with the use of secrecy, as they transformed chemical analysis in the second half of the
nineteenth century (Haber, 1958).
Exhibition data show that advances in methods of scientific analysis, which made
it possible for the first time to reverse-engineer many types of innovations, intensified
inventors’ dependency on patents. At the world fairs of 1851 and 1876, respectively 0
and 5 percent of US chemical innovations were patented. By 1893 and 1915, respectively
19 and 20 percent of US chemical innovations were patented. At the same time, innova-
tions in manufacturing machinery remained unaffected by improvements in scientific
methods, because they had always been easy to reverse-engineer. Patenting rates for these
innovations stayed roughly constant between 44–49 percent for manufacturing machinery
(Moser, 2012). This suggests that the shift toward patenting was due to the reduced
effectiveness of secrecy in chemicals, rather than an economy-wide increase in the use of
patents.
XII. INTELLECTUAL PROPERTY RIGHTS FOR LIVING

ORGANISMS HAVE NOT ENCOURAGED INNOVATION
A contentious question is whether animals and plants should be subject to patent protec-
tion. Until 1930, living organisms could not be patented. In the aftermath of World War
I, however, concerns about food security motivated the passage of the 1930 US Plant
Patent Act, which created intellectual property rights for plants that propagate (asexually)

through roots rather than seeds. In support of the Act, plant breeders had argued that
they needed intellectual property rights to protect their investments in R&D because com-
petitors could easily reproduce an improved plant—which had taken years to develop—by
taking a cutting. Stark Brothers Nursery for example, had built a cage with a burglar
alarm to prevent competitors from taking cuttings of its improved Golden Delicious apple
(Moser and Rhode, 2012). Responding to these arguments, Congress hoped to encourage
the development of a domestic US plant breeding industry by creating plant patent rights.
Data on US plant patent grants, however, show that nearly half of all US plant patent
grants between 1930–70 were for roses (Moser and Rhode, 2012). Had the provision
of intellectual property helped to encourage the creation of a research-based US rose
breeding industry? Until World War II, US nurseries were dependent on imported nursery
stock from Europe. In the 1940s commercial nurseries began to build their own mass
hybridization programs. In a striking parallel to pharmaceutical R&D, it takes up to 12
years to develop a new rose, and most attempts to create new roses fail; fewer than 1 in
1000 seedlings typically lead to a commercially successful new rose (Robb, 1964). Yet as
soon as a breeder succeeds in developing a new rose, competitors can propagate it by
stealing cuttings.
To examine whether the creation of plant patents encouraged innovation, we again con-
struct an alternative (no-patent) measure of innovation in roses. This measure draws on
registration records for newly created rose varieties between 1916–70. Matching registra-
tion records with plant patents shows that fewer than 20 percent of newly registered roses
after 1930 were protected by a plant patent (Moser and Rhode, 2012), again highlighting
the fact that most innovations are not patented.
Registration data also indicate that US breeders created fewer new roses after 1931.
In fact, lineage data show that most roses that are commercially successful today are
descendants of roses that public sector, rather than private sector, plant scientists devel-
oped before the creation of plant patents. These results indicate that plant patent played
no significant role in encouraging innovation in rose breeding, and they highlight the need
for analyses of other sectors.
XIII. PATENT LAWS HAVE CHANGED THE DIRECTION OF

INNOVATION
If patents are important in some industries (e.g., manufacturing machinery), but not in
others (e.g., scientific instruments or chemicals in the nineteenth century), changes in
patent laws may influence the direction if not the level of technical change (Moser, 2005).
These patterns are borne out in exhibition data: countries without patent laws contribute
as many exhibits and prize winners as countries with patent law. But their innovations
are disproportionately focused on industries in which secrecy is effective, so that inventors
are less dependent on patents. At Crystal Palace, one-quarter of exhibits from countries
without patent laws were scientific instruments, compared with one-seventh of exhibits
from other countries (Moser, 2005).
After the Netherlands abolished its patent protection—as a result of a political victory
of the free-trade party in 1869—Dutch innovations shifted toward food processing,
another industry in which secrecy was effective. Between 1851–76, the share of Dutch

innovations in food processing increased from 11 to 37 percent (Moser, 2005). Many

other innovations in this industry, including milk chocolate, baby foods, and ready-made
soups, were made in Switzerland and the Netherlands when neither country had a patent
law (Schiff, 1971).
The case of margarine illustrates the effectiveness of secrecy compared with patents.
The French chemist Mège Mouriès disclosed the process of producing margarine to the
Dutch entrepreneurs Jurgens and van den Bergh, believing that he was protected by a
patent. In 1871, however, while the Netherlands did not have patent laws, Jurgens and van
den Bergh began to manufacture margarine there. After the partners argued and split, van
den Bergh independently improved the taste of margarine, and Jurgens failed to reverse-
engineer van den Bergh’s improved margarine until 1905 (Schiff, 1971).
XIV. PATENT LAWS HAVE CHANGED THE GEOGRAPHY OF

INNOVATIONS
One important area for future analyses relates to the role that intellectual property rights
play in determining the geography of innovation. Can variation in the use of patents help
to explain why some regions—within the same country and the same patent system—are
more innovative than others?
Accounts of innovation in Britain and the United States indicate that knowledge shar-
ing without patents laid the foundations for innovation and technical change (Allen, 1983;
Nuvolari, 2004; Thomson, 2009). For example, technical performance in the Cleveland
(UK) iron-smelting district improved rapidly between 1850–75, as iron producers shared
technical details about the construction and performances of blast furnaces (Allen, 1983).
Yet these benefits appear to be extremely localized.
Anecdotal evidence suggests that the ability to protect ideas through intellectual
property rights encourages inventors to talk about them, and allows future generations
of inventors to build on their ideas. Many inventors, such as Mège Mouriès, were more
willing to disclose technical information when they felt protected by patents. The British
inventor Robert Ransome, for example, only began to advertise his plough-shares in
Norwich and East Anglia after he received a patent in 1803 (MacLeod, 1988). Moreover,
historical records suggest that, in the absence of patent protection, artisans took fierce
measures to prevent the knowledge of new technologies from spreading to outsiders. Thus
silk weavers in seventeenth-century Bologna hanged Ugolino Menzani because he had
shared the knowledge of a new silk twisting machine with Venetian weavers (Belfanti,
2004).
Thus the availability of intellectual property rights may encourage the diffusion of new
ideas, and thereby enable follow-on invention. Although encouraging the diffusion of new
technologies is one of the two key goals of the US patent system, there is little systematic
empirical evidence on the effectiveness of patents in encouraging diffusion.
The shift toward patenting in chemicals (documented in Moser, 2012), however,
creates an opportunity to examine whether a shift toward patenting can encourage the
geographic diffusion of innovative activity. In chemistry, inventors began to rely more on
patents after a series of scientific breakthroughs (including the creation of the periodic
table) reduced the effectiveness of secrecy. Geographic data on the location of origins of

exhibits show that chemical innovations became significantly more geographically diffuse
after 1851. By comparison, the localization of innovation in manufacturing machinery
(which did not experience a shift toward patenting) remained unchanged. Data from the
US Decennial Census between 1840–1920 also show that the geographic decentralization
of chemical innovations cannot be explained by changes in the localization of production.
While chemical exhibits became less localized after 1876, chemical production remained
relatively stable.
These results are consistent with the hypothesis that intellectual property rights can
encourage the diffusion of innovation, but more empirical evidence is needed to identify
these effects.
XV. NARROW INTELLECTUAL PROPERTY RIGHTS—

THROUGH COPYRIGHTS—AS A PROMISING
ALTERNATIVE
Can economists help to identify alternative mechanisms to protect intellectual property

without the shortcomings of patents? One promising alternative are copyrights, which
create intellectual property rights that are substantially narrower than patents. An
extremely stylized example helps to illustrate the difference in scope: a copyright for
Giacomo Puccini’s Tosca (1900) protects the specific score and libretto, whereas a patent
could in principle protect the idea of an opera about two unlucky artists in love. Because
copyrights are so much narrower than patents, they avoid many pitfalls of broad patents,
which makes them an appealing alternative to patents for software and many other fields.
Moreover, the growing importance of digital content, which is primarily protected through
copyrights, highlights the need for understanding the effects of copyrights. Yet—with a
few notable exceptions, such as Oberholzer-Gee and Strumpf (2007), Reimers (2014), and
Waldfogel (2012) —copyrights are understudied by economists.
In the case of copyrights, the key trade-off is encouraging the creation of new cultural
goods and allowing for their use by consumers and other artists. Initial evidence suggests
that basic levels of copyright protection can encourage creativity and innovation. Starting
from low levels of existing protection, extensions in copyright length for Romantic Period
authors raised the price of books (Li, MacGarvie, and Moser, 2014) and increased pay-
ments to authors, even though most of the gains accrued to a small number of exception-
ally popular authors, such as Sir Walter Scott (MacGarvie and Moser, 2014).
The introduction of copyrights also appears to have caused an increase in both the
quantity and the quality of operas in nineteenth-century Italy. Italian states that came
under Napoleon’s rule before 1804 (when France adopted the code civil) adopted the
French copyright law of 1793; whereas Italian states that came under Napoleon’s rule
after 1804 did not get copyrights. Data on the quantity, popularity, and durability of
new operas indicate that the introduction of copyrights led to an increase in the quantity
and quality of operas (Giorcelli and Moser, 2015). These data also show that copyright
extensions beyond the life of the composer had no effects on creativity.
Copyrights, do, however, come at the cost of restricting access to existing works. These
costs may be especially high when copyrights extend beyond the life of the composer and
are extremely long, as they are in the United States and Europe today, where copyright

lengths approach 100 years. Broadly speaking, restrictions on access (and higher prices
as a result of copyrights) increase the costs of access to existing works, which build the
foundation for future generations of what Suzanne Scotchmer (1991) called “cumulative
research.”
To investigate the effects of copyrights on “cumulative” or “follow-on” science, Biasi
and Moser (2015) exploit a historical event of copyright violations under the World War
II Book Replication Program (BRP). This program issued temporary six-month licenses
for German-owned copyrights to US publishers, which then reprinted exact copies of
German science books and sold them for a lower price to US libraries and firms.
Citations data indicate that the BRP resulted in a significant increase in new books and
publications that built on German science books. A simple model of knowledge produc-
tion predicts a differential effect of copyrights on human-capital intensive industries,
and a comparison of changes for mathematics and chemistry confirms this prediction.
Notably, the increase in citations was much more pronounced in mathematics—which
was more dependent on human capital—than in chemistry. Intuitively, US mathemati-
cians who now had access to a copy of Courant and Hilbert’s Methods of Mathematical
Physics at their local university library relied only on their human capital to use methods
of mathematical physics in their own research. Chemists, however, were more likely to also
require access to physical capital, in the form of laboratory space, to use the knowledge in
a BRP book in their own research.
XVI. CONCLUSIONS AND DIRECTIONS FOR FUTURE

RESEARCH
Economic history offers unique data sets and a wealth of exogenous variation to
analyze the effects of intellectual property rights to innovation. For example, exhibits at
nineteenth-century world fairs offer a rare alternative (non-patent) measure of innovation
to examine the effects of patent laws at a time when several countries had not yet adopted
patent laws or abolished them for political reasons. These data indicate that the large
majority of historical innovations occurred outside of the patent system. Moreover, they
reveal that countries without patent laws, such as Switzerland and the Netherlands, were
at least as innovative as countries with patent laws. Comparisons across industries also
show that the need for patents varies significantly across industries and over time. This
need for patents has been strongest when innovations were easy to reverse-engineer, so
that inventors could not (or no longer) rely on secrecy to protect their innovations. At an
economy-wide level, such differences have influenced the direction of technical change,
even when they failed to increase levels of innovation.
More generally, intellectual property rights appear to have been most beneficial when
they were narrow and short-lived. Compulsory licensing, for example, forces patent
owners to license their inventions to new firms, including potential competitors. This
policy has encouraged innovation—not only by the firms that benefit from improved
access, but by the original owners of licensed patents, which now face more competition.
Similarly, patent pools that allow competitors to combine their patents have discouraged
innovation, in part by making it more difficult for entrants to compete with incumbents
that have formed a pool.

Interestingly, initial results on the effectiveness of copyrights suggest that basic levels of
copyrights (which are significantly narrower than patents) can encourage creativity. But
these results also suggest that extensions in the length of copyrights trigger no additional
boost in creativity. Such extensions may in fact create significant welfare losses by discour-
aging the diffusion of cultural goods. These effects are particularly severe for science, a
field in which the creation of new knowledge depends critically on access to existing work.
Many important questions and research opportunities remain. For example, what are
the effects of intellectual property rights on the diffusion of innovation? Can basic levels
of property rights help to ensure a broad-based participation in processes of creativity
and innovation? And are copyrights a useful alternative to patents? Historical analyses
can shed light on these questions—with the benefit of distance, new data sets, and careful
empirical analyses.
Although the focus of this chapter has been on intellectual property rights, the
creation of “economically valuable knowledge” (Griliches, 1990) is also vitally dependent
on human capital: inventors, scientists, and entrepreneurs. To encourage innovation,
economic policies must enable and encourage these individuals to invest in creativity and
innovation. A small number of existing studies suggest that basic levels of intellectual
property rights can encourage broad-based participation—for example,

Ben Depoorter, Peter Menell, David Schwartz - Research Handbook On The Economics of Intellectual Property Law - Vol 1 - Theory and Vol 2 - Ana

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Ben Depoorter, Peter Menell, David Schwartz - Research Handbook On The Economics of Intellectual Property Law - Vol 1 - Theory and Vol 2 - Ana

Uploaded by

Copyright:

Available Formats

RESEARCH HANDBOOK ON THE ECONOMICS OF

INTELLECTUAL PROPERTY LAW

DEPOORTER_V1_9781848445369_t.indd 1 30/07/2019 15:48

Research Handbook on the Economics of Property Law

Research Handbook on the Economics of Family Law

Research Handbook on the Economics of Antitrust Law

Research Handbook on the Economics of Corporate Law

Research Handbook on the Economics of European Union Law

Research Handbook on the Economics of Criminal Law

Research Handbook on the Economics of Labor and Employment Law

Research Handbook on Austrian Law and Economics

Research Handbook on Behavioral Law and Economics

Research Handbook on the Economics of Intellectual Property Law

Research Handbook on the Economics of Intellectual Property Law

DEPOORTER_V1_9781848445369_t.indd 2 30/07/2019 15:48

RESEARCH HANDBOOKS IN LAW AND ECONOMICS

Cheltenham, UK • Northampton, MA, USA

DEPOORTER_V1_9781848445369_t.indd 3 30/07/2019 15:48

Edward Elgar Publishing, Inc.

A catalogue record for this book

Library of Congress Control Number: 2019945636

This book is available electronically in the

ISBN 978 1 84844 536 9 (2 volume set)

Typeset by Servis Filmsetting Ltd, Stockport, Cheshire

DEPOORTER_V1_9781848445369_t.indd 4 30/07/2019 15:48

PART I INTELLECTUAL PROPERTY AS PROPERTY

1 Intellectual property as property 2

PART II IP AND INCENTIVES

4 Philosophical foundations of IP law: the law and economics paradigm 72

DEPOORTER_V1_9781848445369_t.indd 5 30/07/2019 15:48

PART III IP COSTS

14 Tailoring intellectual property rights to reduce uniformity cost 377

PART IV IP AND INSTITUTIONS

17 Patent institutions: shifting interactions between legal actors 473

PART V IP, DEVELOPMENT, AND INTERNATIONAL TRADE

22 Creative development: copyright and emerging creative industries 582

DEPOORTER_V1_9781848445369_t.indd 6 30/07/2019 15:48

Michael Abramowicz, Professor of Law, George Washington University, USA

DEPOORTER_V1_9781848445369_t.indd 7 30/07/2019 15:48

DEPOORTER_V1_9781848445369_t.indd 8 30/07/2019 15:48

DEPOORTER_V1_9781848445369_t.indd 1 30/07/2019 15:48

* Harold C. Hohbach Distinguished Professor of Patent Law and Intellectual Property;

DEPOORTER_V1_9781848445369_t.indd 2 30/07/2019 15:48

II. ‘PROPERTY’ AND IP

The US Constitution authorizes Congress to ‘promote the Progress of Science and

DEPOORTER_V1_9781848445369_t.indd 3 30/07/2019 15:48

III. DISTINGUISHING TANGIBLE AND INTELLECTUAL

DEPOORTER_V1_9781848445369_t.indd 4 30/07/2019 15:48

DEPOORTER_V1_9781848445369_t.indd 5 30/07/2019 15:48

DEPOORTER_V1_9781848445369_t.indd 6 30/07/2019 15:48

IV. THREE CROSS-CUTTING THEMES

A. Property and Possession

1. Possession, property origins, and the public domain

DEPOORTER_V1_9781848445369_t.indd 7 30/07/2019 15:48

DEPOORTER_V1_9781848445369_t.indd 8 30/07/2019 15:48

DEPOORTER_V1_9781848445369_t.indd 9 30/07/2019 15:48

DEPOORTER_V1_9781848445369_t.indd 10 30/07/2019 15:48

DEPOORTER_V1_9781848445369_t.indd 11 30/07/2019 15:48

2. The challenges of non-possessory property

DEPOORTER_V1_9781848445369_t.indd 12 30/07/2019 15:48

Justice Holmes alludes to another apparent copyright anomaly: although copyright

DEPOORTER_V1_9781848445369_t.indd 13 30/07/2019 15:48

B. Property and Information

1 Intellectual property as property 2

4 Philosophical foundations of IP law: the law and economics paradigm 72

14 Tailoring intellectual property rights to reduce uniformity cost 377

17 Patent institutions: shifting interactions between legal actors 473

22 Creative development: copyright and emerging creative industries 582

III. DISTINGUISHING TANGIBLE AND INTELLECTUAL

Williams compared subsequent publications and diagnostic applications of genes that