You are on page 1of 7

Funding Conditions and Incentives for the Disclosure of Scientific Knowledge

Submission to the Royal Society: Science as a Public Enterprise: Opening up Scientific Information Authors: Professor Joshua S. Gans1 and Professor Fiona Murray2 Date: 29th July, 2011

Skoll Chair of Technical Innovation and Entrepreneurship and Professor of Strategic Management, Rotman School of Management, University of Toronto, Email: joshua.gans@rotman.toronto.ca
1

2

Associate Professor of Technological Innovation, Entrepreneurship & Strategic Management, Sloan School of Management, Massachusetts Institute of Technology. Email: fmurray@mit.edu

Rotman School of Management, University of Toronto 105 St George St, Toronto ON M5S 3E6

W www.joshuagans.com

About the Authors
Joshua Gans is an economist who holds the Skoll Chair of Innovation and Entrepreneurship at the University of Toronto, Canada. His research focuses on microeconomics, competition policy and innovation. He is the author of several textbooks and policy books, as well as numerous articles in economics journals. Gans received a Bachelor of Economics (Honours) from the University of Queensland before going to Stanford University to study for his Ph.D. in Economics. He graduated from Stanford in 1995 and was a full Professor at Melbourne Business School before moving to Toronto in 2011. In 2007, Gans received the inaugural young economist award from the Economic Society of Australia. This is an award given every two years to the best economist working in Australia, who is aged under 40. Fiona Murray is an Associate Professor of Management in the Technological Innovation and Entrepreneurship Group at the MIT Sloan School of Management and Faculty Director of the MIT Entrepreneurship Center. She received BA and MA degrees from the University of Oxford in Chemistry before receiving her doctoral degree from Harvard University’s School of Engineering and Applied Sciences. Her research interests moved away from the bench to the study of science commercialization, the organization of scientific research and the role of science in national competitiveness. After a short time on the faculty of Oxford’s Said Business School Fiona joined the MIT Sloan School of Management where she studies and teaches innovation and entrepreneurship with an emphasis on the life sciences and clean technology sector.

Disclosure of Scientific Knowledge

Submission

Table of Contents
Introduction Alternative modes of disclosure An open access requirement The role of intellectual property protection Communication and new media 1 1 2 3 4

i

Submission

Introduction
This submission has a narrow purpose: to highlight relevant research in economics and strategic management designed to understand the impact of various institutions and policies (most notably, policies around intellectual property protection) on the incentives of scientists and their funders to disclose scientific knowledge. It is important to recognise that many scientists, science-policy makers and observers are aware that the generation and application of knowledge is the well-spring of economic growth. Nonetheless, there is less recognition that the key mechanism generating the power of science to drive growth is the disclosure and sharing of the generated knowledge stock. If knowledge is kept secret or if it is locked down with long-lived intellectual property protection or patent thickets, the process of growth is disrupted with large cumulative effects on welfare. In light of this insight, The Royal Society’s deliberations should, in our opinion, consider how to change the default behaviors and conditions associated with research funding in order to ensure that disclosure and open access occurs rapidly and effectively.

Alternative modes of disclosure
In recent years there has been significant emphasis on research that can be described as existing in “Pasteur’s Quadrant” – that is research that simultaneously makes a contribution to fundamental scientific knowledge and to economically useful activities. 3 While Pasteur’s work on the germ theory of disease is the eponymous example of work that makes a fundamental and an economic contribution (and thus contributes to economic growth), more recent examples include the discovery of monoclonal antibodies, RNA interference research or advances in lasers. Work in Pasteur’s Quadrant can be disclosed through a variety of routes (although of course it can also be kept secret). Specifically, as a contribution to fundamental knowledge it could be the basis of a scientific publication. Alternatively, it could form the basis of a patent due to the usefulness of the new idea (assuming it meets non-obviousness and novelty requirements). However, knowledge in Pasteur’s quadrant can, with careful management of timing, also be disclosed in patent-paper pairs – in both patents and publications. As long as the publication arises after the filing of a patent application, the publication does not constitute prior art thus allowing the scientist both to generate potential kudos or reputational benefits from the publication and the property rights from the patent (assuming the patent is granted). Examples of such pairs are widespread. In a study of almost 200 publications from the British journal Nature Biotechnology content from 51% of the publications was also filed in patents (with the same inventors/authors and closely linked filing dates).4 Likewise, a recent study of over 2000 patents filed

3

Stokes, D. 1997. Pasteur’s Quadrant: basic science and technological innovation. Washington D.C.: The Brookings Institution; and Murray, F., & Stern, S. 2007. Do formal intellectual property rights hinder the free flow of scientific knowledge? An empirical test of the anti-commons hypothesis. Journal of Economic Behavior and Organization, 63(4), 648-687
4

Murray & Stern, op.cit.

Disclosure of Scientific Knowledge

1

Submission

by researchers in the University of California system linked all these patents to academic publications.5 This is not to say that patents have come to substitute for publications. A study of researchers at MIT over a 25 year period suggests a ratio of publications to patents of about 10:1, however that knowledge that is patented is linked to “paired” publications.6 Moreover, the prevalence of such patenting activity is linked to researchers who are the most productive in terms of their research publication outputs.7 Nonetheless, the decision to file for patents seems to remain a choice for researchers with few if any restrictions placed on them by public funders (although private funders are more explicit in their requirements for invention disclosure).

An open access requirement
Given the potential for much research to be disclosed through publications or patents, it is useful to analyse the different choices and the implications of different disclosure rules. Perhaps an appropriate starting point is to consider what happens if open access to research results is a default requirement. This is an approach that might be considered the norm for most publicly funded research and yet in reality, until recently, few grant-making agencies have explicitly required disclosure and open access. Instead they assume that researchers will indeed make their research available (at some point in the future) but do not mandate such behavior. In the event of such a requirement for publicly funded research, this will result in more disclosure and benefits that flow from that. Thus, on the face of it, this would appear to be a sensible move. However, as with all analytical endeavours, we need to consider how this requirement might change the behaviour of scientists. In situations where scientists have no alternative to public funding, then a requirement to disclose will not likely prevent them from accepting such funding. What happens, however, when scientists have the potential to tap into commercial sources of funding? In this situation, if the scientist is adverse to disclosure as a default (or at least to rapid disclosure), then they may move to accept commercial sources of funding instead. In this case, there is a benefit in that the disclosure requirement frees up some public funds for other scientists as those scientists with the potential to use other funds move elsewhere. But, we have to ask ourselves: which scientific activities will likely make this switch? It is likely to be made by those scientists who were most adverse to disclosure. If such a switch occurs, the public pursue will not be contributing towards more disclosure overall.8 This leads to the natural question - why might scientists not wish to be open in their research as a matter of course? After all, norms of scientific kudos and status lend themselves towards publication
5

Thompson, Neil, D. Mowery, A. Ziedonis, 2011. The Impact of Patents, Licensing and Materials Transfer Agreements (MTA) On the Flow of Knowledge. Working Paper. Oregon.
6

Bikard M and F Murray 2011. "How Creative is Collaboration? Examining the Organization of Knowledge Work," Working Paper, MIT.

7

Azoulay, P., Ding, W. & Stuart, T. 2009. The Impact of Academic Patenting on the Rate, Quality and Direction of (Public) Research Output. Journal of Industrial Economics, 75: 637-676
8

These effects are documented in Joshua S Gans and Fiona Murray, "Funding Scientific Knowledge: Selection, Disclosure and the PublicPrivate Portfolio," Rate and Direction of Inventive Activity, J. Lerner and S. Stern (eds), NBER, 2011 (forthcoming).

Disclosure of Scientific Knowledge

2

Submission

of results. One reason could be that the scientist has some expectation that they will earn commercial profits from their research -- even publicly funded research -- and would not want to jeopardise those commercial returns with full disclosure of research outcomes and data (particularly for research not fully covered by patents). Another reason is that the scientist would like to develop further results in the future and has no incentive to release results quickly – for example, some of the debate around disclosure of results from the Human Genome Project rested on the speed of disclosure of gene sequences with the Bermuda Rules shifting disclosure from the discretion of the researcher and typically 1-2 years to immediate disclosure. In other areas, more rapid and fuller data disclosure could be ‘forced’ by a requirement for disclosure and open access but this also suggests that such disclosure might have adverse consequences. For instance, it may reduce incentives for the scientist to build on their own research. It also suggests that rather than requiring disclosure, we need to investigate why it is that scientists do not seem to be rewarded (with kudos) for the earlier release of results.

The role of intellectual property protection
It is instructive here to also discuss the role of intellectual property protection in this context. As already noted, in many scientific disciplines, knowledge is disclosed simultaneously through both a patent and a publication. And while there are concerns that patent protection may raise the costs for follow-on researchers in their activities, the potential for patent-paper pairs suggests some complementarity between the two forms: that is, it is possible that stronger intellectual property protection (at least in some important instances) can promote greater scientific disclosure. The rationale for this is twofold.9 First, if commercial motives (from a scientist, their funder or both) lead to some reticence towards disclosure, patent protection can alleviate this reticence. Specifically, the patent already involves disclosures that may assist competitors and so the incremental assistance coming from publication may be relatively low. Second, the patent protects commercial interests from entry/competition and so, in the process, protects them from the consequences of disclosures in assisting competitors. This suggests that The Royal Society needs to carefully consider the role of IP protection in potentially promoting disclosure and the impact of prohibitions on the use of such protection attached to sources of funding. That said, we are not advocating blanket IP protection and assertion of rights. Instead, we are pointing out that the relationship is more nuanced and there are aspects of IP protection that may promote scholarly disclosure, particularly for private funders. While outside the scope of this brief submission, we would suggest that the challenges of IP rights for follow-on research seem to specifically be linked to the terms of licensing agreements related to these rights rather than IP per se.

9

These are discussed in Joshua S Gans, Fiona Murray and Scott Stern, “Contracting Over the Disclosure of Scientific Knowledge: Intellec-

tual Property and Academic Publication” mimeo., MIT, April 2001.

Disclosure of Scientific Knowledge

3

Submission

Communication and new media
The Royal Society has asked whether new media should change how scientists discuss research and also communicate about it. As a baseline communication options provide for more rapid disclosure and widespread communication and sharing. However, in the light of our prior discussions of the drivers of disclosure we want to note several additional issues to consider: • Attribution. How do scientists receive attribution for the content of a blog post or even for datasets online? How are these to be included in metrics of scientific performance that are now based on citation counts? And when online tools facilitate collaboration (e.g., the Polymath projects), how do individual contributions receive attribution? The point here is that the tools can be provided but getting the incentives aligned with traditional scientific norms and institutions is a greater challenge. • Timing. As noted above, the degree to which scientists can file patents and submit publications for a particular research project depends upon careful management of the timing of publication disclosure. Any disclosure in an informal, within university, setting does not constitute a disclosure that is prior art for patent filing. Nor is disclosure in a grant application. However it is likely that online disclosure would constitute prior art and thus requirements to rapidly share results, data etc. will likely change the options for paired patenting. As a consequence adoption of rapid dissemination modes may be more muted. • Materials. One additional consequence is to examine an additional output of knowledge production – materials. Biological materials as well as physical materials are an important output of research and input into follow-on research. As a consequence, it is important to recall the need to manage the physical flows of knowledge as well as and in conjunction with data and information flows. This suggests that it is imperative to continue to emphasize the linking of different flows to underlying research projects so that those using online data to replicate or build upon a finding can also easily link to the physical materials. • Link to Patents. The online infrastructure provides an excellent opportunity to link publications, patents, licenses and related data and materials. At present, the scientific information infrastructure effectively connects scientific papers to their citations and (to a lesser extent to their materials). A clear link to patents and more importantly to associated licenses would enable researchers to more carefully navigate intellectual property requirements, access issues etc. That said there are opportunities to expand the accessibility of published scientific knowledge for the broader community. It is easier for published scholarly research to be associated with commentary and other explanations from scientists themselves. These links promote a number of ends including an appreciation for research, communication of results and cautious translation of knowledge for broader consumption.

Disclosure of Scientific Knowledge

4