The1bst

Topics in Regulatory Economics and Policy
Michael A. Crew
Timothy J. Brennan
Editors
The Future
of the Postal
Sector in a
Digital World
Series Editor:
Michael A. Crew
More information about this series at http://www.springer.com/series/6646

Michael A. Crew • Timothy J. Brennan
Editors
The Future of the Postal

Sector in a Digital World
Editors
Michael A. Crew Timothy J. Brennan
Rutgers Business School School of Public Policy
Rutgers University University of Maryland Baltimore
Newark, NJ, USA County (UMBC)
Baltimore, MD, USA

ISBN 978-3-319-24452-5 ISBN 978-3-319-24454-9 (eBook)
DOI 10.1007/978-3-319-24454-9
Library of Congress Control Number: 2015956529
Springer Cham Heidelberg New York Dordrecht London

© Springer International Publishing Switzerland 2016
This work is subject to copyright. All rights are reserved by the Publisher, whether the whole or part of
the material is concerned, specifically the rights of translation, reprinting, reuse of illustrations,
recitation, broadcasting, reproduction on microfilms or in any other physical way, and transmission
or information storage and retrieval, electronic adaptation, computer software, or by similar or
dissimilar methodology now known or hereafter developed.
The use of general descriptive names, registered names, trademarks, service marks, etc. in this
publication does not imply, even in the absence of a specific statement, that such names are exempt
from the relevant protective laws and regulations and therefore free for general use.
The publisher, the authors and the editors are safe to assume that the advice and information in this
book are believed to be true and accurate at the date of publication. Neither the publisher nor the
authors or the editors give a warranty, express or implied, with respect to the material contained
herein or for any errors or omissions that may have been made.
Printed on acid-free paper
Springer International Publishing AG Switzerland is part of Springer Science+Business Media

(www.springer.com)
Preface and Acknowledgments
This book is a result of the Center for Research in Regulated Industries’ (CRRI)
23nd Conference on Postal and Delivery Economics, which was held June 3–6,
2015, at the Amarilia Hotel, Vouliagmeni, Greece. The first Conference was held in
1990 in the UK. During the 25 years since the first Conference, the industry has seen
considerable change. These include the opening of postal markets to competition
for most countries in the European Union on January 1, 2011. Even more important
is the increasing impact of multimodal competition. As a result of e-mail, social
networks, sophisticated handheld devices, and Internet advertising, important ques-
tions are being raised about the future of mail. The Conference and this book
attempt to address some of the resulting challenges. They follow earlier conferences
and workshops. This 23rd edited volume in CRRI’s program Postal and Delivery
Economics is the final volume in the series as this was the last CRRI Postal
Conference.
The Conference was made possible by the support of its generous sponsors. We
would like to thank sponsors not only for financial support and for supporting
service on the organizing committee but also for, along with others, their intellec-
tual contributions, advice, and encouragement: Mohammad Adra, Christos
Apostolou, Jody Berenblatt, Geoff Bickerton, Stephen Brogan, Jim Bruce, Larry
Buc, Carla Pace, Xenophon Chatzithanasis, Jo~ao Confraria, Margaret Cigno, Jeff
Colvin, Giota Christodoulou, Constantinos Delicostopoulos, Peter Dunn, Richard
Eccles, Colm Farrelly, Charles Fattore, Stephen Ferguson, Ruth Goldway, Stefano
Gori, Robert Hammond, John Hearn, Paul Hodgson, Adam Houck, Jim Holland,
George Houpis, Christian Jaag, Keith Kellison, George Kuehnbaum, Denis Joram,
Chantale La Casse, David M. Levy, François Lions, Martin Maegli, Francesco
Materia, Leonardo Mautino, Anna M€oller, Heikki Nikali, Veerle Nuyts, Chris
Paterson, Ted Pearsall, Alberto Pimenta, Michael Ravnitzky, Agustin Ros, Chris
Rowsell, Francisco SamaoJim Sauber, Michael Scanlon, Gennaro Scarfiglieri,
Soterios Soteri, Nancy Sparks, Robert Taub, Urs Trinkner, Mark van der Horst,
Tim Walsh, David Williams, and Ralf Wojtek.
v
vi Preface and Acknowledgments
This year’s Conference benefited greatly from the efforts of the Hellenic Post,
including Thalia Giannaki, ELTA’s IT staff and others, who were incredibly helpful
during the Conference, enabling it to operate very smoothly. They and colleagues
provided both advice and assistance on numerous occasions and contributed greatly
to the success of the event.
We would like to thank our distinguished dinner speakers: Rania Karra, Head of
Monitoring Community Policy Section, and Constantinos Delicostopoulos, Vice
President, Hellenic Post and Telecommunications Commission. These speeches
addressed strategy in addressing current issues of regulation and postal reform
against the background of increasing e-competition and declining letter volumes
in the postal sector.
The Amaralia Hotel was a great location for the Conference and provided
excellent conference facilities, as well as fine food and lodging. Andreas Paisios
and his staff did a wonderful job of taking care of the participants of the Conference.
In addition, we thank all authors and participants of the Conference. Absent their
contributions, the Conference and this book would not have been possible. The
usual disclaimers are applicable. In particular, the views expressed reflect the views
of the authors and are not necessarily those of the sponsors.
Timothy J. Brennan Michael A. Crew

Baltimore, MD Newark, NJ
List of Sponsors
Royal Mail
R.R. Donnelley & Sons Company
United Parcel Service
La Poste
Citizens Advice
CTT Correios de Portugal
Pitney Bowes
U.S. Postal Regulatory Commission
Poste Italiane
ARCEP
FedEx Express
NERA Economic Consulting
Swiss Post
Oxera
Canadian Union of Postal Workers
ANACOM
Frontier Economics
bpost
Commission for Communications Regulation
Copenhagen Economics
Cullen International
National Association of Letter Carriers
An Post
BIEK
Diversified Specifics
K&L Gates LLP
New Zealand Post Limited
GrayHair Advisors
vii
viii List of Sponsors
IBM Global Business Services

SLS Consulting, Inc.
Venable LLP
Swiss Economics
Contents
1 Price Cap Regulation and Declining Demand . . . . . . . . . . . . . . . . . 1

Timothy J. Brennan and Michael A. Crew
2 Scale Economies and Postal Price Caps in Europe:
Declining Volumes, Lower Productivity, Higher Postage? . . . . . . . 19
Antonia Niederprüm, Christian M. Bender, and Alex Kalevi Dieke
3 Origins of “Universal Service” . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29
Chris Rowsell
4 Implications of Recent Developments in e-Commerce
for Universal Service Providers and the USO . . . . . . . . . . . . . . . . . 43
John Hearn
5 Which Universal Service Obligation Attributes
Do Americans Value? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59
Jennifer Bradley, Michael D. Bradley, and Jeff Colvin
6 Impact of Market Dynamics on the Net Cost of the USO . . . . . . . . 75
Dariusz Nehrebecki, Leonardo Mautino, and Gavin Knott
7 Innovation, Disruption, and Partnering: Changing the DNA
of Posts from the Inside Out . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 91
Adam C. Houck
8 Postal Strategies in a Digital Age . . . . . . . . . . . . . . . . . . . . . . . . . . 103
Christian Jaag, Jose Parra Moyano, and Urs Trinkner
9 Case Studies in End-to-End Delivery Competition . . . . . . . . . . . . . 121
Philip Groves and Steven Cape
10 Promoting Competition at the Digital Age with an Application
to Belgium . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 137
Pierre Copée, Axel Gautier, and Mélanie Lefèvre
ix
x Contents
11 Compensation Fund Under EU Law: A Suitable

Solution for the Postal Market? . . . . . . . . . . . . . . . . . . . . . . . . . . . 153
Alessandra Fratini
12 The Shape of Postal Inefficiency . . . . . . . . . . . . . . . . . . . . . . . . . . . 169
Margaret M. Cigno, John P. Klingenberg, and Edward S. Pearsall
13 Differentiated Pricing of Delivery Services
in the e-Commerce Sector . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 191
Claire Borsenberger, Helmuth Cremer, Philippe De Donder,
and Denis Joram
14 Mode of Delivery and Customer Response to Advertising Mail . . . 213
Michael D. Bradley, Laraine Balk Hope, and John Pickett
15 Commercial and Regulatory Challenges for Postal
e-Services in Switzerland . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 231
Christian Jaag, Martin Maegli, and Denis Morel
16 Demand and Regulation for e-Commerce in Goods . . . . . . . . . . . . 247
Heikki Nikali, Juri Mattila, Ilona Rintanen, and Ville Huuhtanen
17 Personal Data and Privacy Issues and Postal Operators Stand . . . . 261
Claire Borsenberger, Denis Joram, Olaf Klargaard,
Philippe Regnard
18 Using Market Research to Analyze Consumer Preferences . . . . . . . 271
Stefano Gori, Bianca M. Martinelli, and Gennaro Scarfiglieri
19 An Economist’s Perspective on Terminal Dues . . . . . . . . . . . . . . . . 283
Henrik Ballebye Okholm, Anna M€oller Boivie,
Bruno Basalisco, and Julia Wahl
20 A Game-Theoretic Model of the Market for International Mail . . . 295
Edward S. Pearsall
21 Quantifying the Distortive Effects of UPU Terminal Dues . . . . . . . 313
James I. Campbell, Jr.
22 UPU Compensation Rates for Packages Under EU
Competition Law: Are There Lessons to Be Learned
from Other International Fee Arrangements? . . . . . . . . . . . . . . . . 331
Ralf Wojtek
Contributors
Bruno Basalisco Copenhagen Economics A/S, København K, Denmark

Christian M. Bender WIK, Bad Honnef, Germany
oller Boivie Copenhagen Economics A/S, København K, Denmark
Anna M€
Claire Borsenberger Department of Doctrine et Modélisation, Direction of Reg-
ulation and Institutional and European Affairs, Groupe La Poste, Paris, France;
Laboratoire d’Economie d’Orléans (LEO); and University François Rabelais,
Tours, France
Michael D. Bradley Department of Economics, George Washington University,
Washington, DC, USA
George Washington University, Washington, DC, USA
Jennifer Bradley USPS Office of Inspector General, Arlington, VA, USA
Timothy J. Brennan School of Public Policy, University of Maryland Baltimore
County (UMBC), Baltimore, MD, USA
Resources for the Future, Washington, DC, USA
James I. Campbell Jr. Independent Lawyer and Consultant, Potomac, MD, USA
Steven Cape Ofcom, London, UK
Margaret M. Cigno United States Postal Regulatory Commission, Washington,
DC, USA
Jeff Colvin USPS Office of Inspector General, Arlington, VA, USA
Pierre Copée University of Liege (ULg), Liege, Belgium
Helmuth Cremer Department of Economics, Toulouse School of Economics
(IDEI, GREMAQ and IuF), Toulouse, France
xi
xii Contributors
Michael A. Crew CRRI Distinguished Professor of Regulatory Economics,

Rutgers Business School, Newark, NJ, USA
Alex Kalevi Dieke WIK, Bad Honnef, Germany
Philippe De Donder Department of Economics, Toulouse School of Economics
(IDEI and GREMAQ-CNRS), Toulouse, France
Alessandra Fratini FratiniVergano - European Lawyers, Brussels, Belgium
Axel Gautier HEC Management School, Liege Competition and Innovation Insti-
tute (LCII) and CORE, University of Liege (ULg), Liege, Belgium
Stefano Gori CESPI, BEA and Poste Italiane, Rome, Italy
Philip Groves Ofcom, London, UK
John Hearn Formerly vice-chair CERP and Project Manager Postal Regulation,
ComReg, Malahide, Ireland
Laraine Balk Hope USPS Office of Inspector General, Arlington, VA, USA
Adam C. Houck IBM Global Business Services, Washington, DC, USA
Ville Huuhtanen Posti Ltd., Helsinki, Finland
Christian Jaag Swiss Economics, Zurich, Switzerland
Denis Joram Department of Regulation and Studies, Direction of Regulation and
Institutional and European Affairs, Groupe La Poste, Paris, France
Olaf Klargaard Direction of Digital Affairs, Groupe La Poste, Paris, France
John P. Klingenberg United States Postal Regulatory Commission, Washington,
DC, USA
Gavin Knott Competition and Markets Authority, London, UK
Mélanie Lefèvre HEC Management School, University of Liege (ULg), Liege,
Belgium
LICOS Centre for Institutions and Economic Performance, KU Leuven, Belgium
Martin Maegli Swiss Post, Bern, Switzerland
Bianca M. Martinelli Poste Italiane, Rome, Italy
Juri Mattila The Research Institute of the Finnish Economy (ETLA), Helsinki,
Finland
Leonardo Mautino Oxera Consulting LLP, London, UK
Denis Morel Swiss Post, Bern, Switzerland
Jose Parra Moyano Swiss Economics, Zurich, Switzerland
Contributors xiii
Dariusz Nehrebecki Oxera Consulting LLP, London, UK

Antonia Niederprüm WIK, Bad Honnef, Germany
Heikki Nikali Posti Ltd., Helsinki, Finland
Henrik Ballebye Okholm Copenhagen Economics A/S, København K, Denmark
Edward S. Pearsall Independent Consultant, Alexandria, VA, USA
John Pickett USPS Office of Inspector General, Arlington, VA, USA
Philippe Regnard Direction of Digital Affairs, Groupe La Poste, Paris, France
Ilona Rintanen Posti Ltd., Helsinki, Finland
Chris Rowsell Ofcom, London, UK
Gennaro Scarfiglieri Poste Italiane, Rome, Italy
Urs Trinkner Swiss Economics, Zurich, Switzerland
Julia Wahl Copenhagen Economics A/S, København K, Denmark
Ralf Wojtek Heuking Kühn Lüer Wojtek, Hamburg, Germany
Price Cap Regulation and Declining Demand
Timothy J. Brennan and Michael A. Crew
1 Introduction
PCR originated in the late 1970s in the USA with a new regulatory scheme for
Michigan Bell that replaced tradition cost-of-service regulation (COS). It gathered
considerable momentum in the 1980s, with its adaption for AT&T following the
Divestiture of the Regional Bell Operating Companies in 1984. PCR’s momentum
continued with the privatization of the utilities in the UK. All of the newly
privatized utilities were subject to PCR, not only telecommunications, but also
electricity, gas, and water.
The foundations for the adoption of PCR derived from two reports to the
British Government (Littlechild 1983, 1986). Stephen Littlechild held the view
that COS was inefficient and that PCR apparently provided stronger incentives for
efficiency. For example, COS, since revenue was a direct function of cost,
provided minimal incentives for cost minimization. In addition, COS involved
costs to regulated firms and regulators of making and evaluating requests to
increase rates.
PCR provides flexibility to regulated firms in that it allows the firm’s basket of
prices to rise by some index reflecting inflation. Frequently, the Consumer Price
Index (CPI) is employed. At its simplest, the allowed percentage change in the
firm’s weighted price basket is the change in the CPI less X, where X represents the
T.J. Brennan (*)

School of Public Policy, University of Maryland Baltimore County (UMBC),
Baltimore, MD, USA
Resources for the Future, Washington, DC, USA
e-mail: brennan@umbc.edu
M.A. Crew (*)
CRRI Distinguished Professor of Regulatory Economics, Rutgers Business School,
Newark, NJ, USA
e-mail: mcrew@business.rutgers.edu
© Springer International Publishing Switzerland 2016 1

M.A. Crew, T.J. Brennan (eds.), The Future of the Postal Sector in a Digital World,
Topics in Regulatory Economics and Policy 51, DOI 10.1007/978-3-319-24454-9_1
2 T.J. Brennan and M.A. Crew
decline in real prices provided to consumers under PCR.1 This is usually referred to
as “CPI X” regulation. The firm is compensated for inflation except for the preset
deduction of X. PCR is attractive to consumers in that the firm’s prices decrease in
real terms. It is beneficial to the firm in that it does not have the expense of a rate
hearing to obtain a price increase. In addition, it has some flexibility to an extent
that it can raise some prices greater than CPI X if it raises others by less than
CPI X.2 In addition, it provides incentives for internal efficiency.
This chapter examines the case where the regulated firm sells one product
regulated under PCR. The regulated firm will maximize profits by taking the regu-
lated price as given. In that sense, it becomes a price taker, just like a competitive
firm. Consequently, the regulated firm notionally has the same incentive to cut costs
as it would if it were operating in a competitive market. In contrast, under COS the
firm’s revenue is based directly on its costs. If the firm reduces its unit costs, its prices
are reduced to the lower average costs. Where demand is growing and technology is
advancing the regulated firm under COS can make greater than the allowed return.
The firm therefore has an incentive to delay filling a rate case until costs rise above
the previously allowed level. Despite the impact of regulatory lag, the COS firm has a
considerably weakened incentive to minimize cost than the PCR firm.
The stronger incentives to minimize costs for the profit-maximizing PCR firm were
a major benefit perceived by the proponents of PCR, e.g., Littlechild (1983). Another
benefit was that unlike a COS-regulated firm, a PCR firm did not gain by subsidizing
competitive products from regulated products (Brennan 1989; Braeutigam and Panzar
1989). However, when PCR was applied to United States Postal Service (USPS)
under PAEA, little or no attention was paid to the fact that USPS had no residual
claimants, which are required for the efficiency benefits of PCR to apply. Similarly,
the absence of residual claimants may mean that the organization’s incentive to avoid
cross subsidy of competitive products may be attenuated. Littlechild and others3 also
saw the limitations of PCR, including the likelihood that any efficiency gains would
likely be short run, raising the question of whether PCR would be an effective
regulatory scheme in the long run. The focus of this chapter is on the problem of
employing PCR when demand is decreasing, primarily in the postal sector.
PCR’s widespread adoption occurred later in the postal sector than in other
sectors. The last major postal operator to adopt PCR was the USPS. PCR was
incorporated into law as the mode of regulation with the passing of the Postal
Accountability and Enhancement Act of 2006 (PAEA) on December 20, 2006.
1
Nominally, X represents a sharing of benefits with consumers of the gains the regulated firm will
get created by its ability to profit from increasing efficiency. As noted below, X may also represent
sharing of profits the firm would achieve from economies of scale when demand increases.
However, in practice X is set through political negotiation rather than the application of economic
theory or evidence.
2
Brennan (1989) shows that with multiple products, a cap of an appropriately weighted sum of
prices of multiple regulated products converges to a set of prices that maximizes consumer welfare
given the profits the regulated firm is able to achieve.
3
See Armstrong and Sappington (2007) for a review of relevant research.
Price Cap Regulation and Declining Demand 3
The new regulatory process was administered by the Postal Regulatory Commis-
sion (PRC), the successor to the Postal Rate Commission. The Act was around a
decade in the making, and it has been widely criticized.
This chapter will not address the multitude of problems of PAEA but will concen-
trate on the impact of PCR on USPS when demand is falling. PCR was intended to
provide greater incentives for efficiency than COS and reduce the transactions costs of
regulation with the intention that USPS would prosper as a result. In fact, the reverse
has happened. Under PCR, it has experienced severe financial problems, which were
more serious than they would have been under the previous regime, and these
problems continue. Its problems arose because of the combined cyclical and secular
decline of mail. PCR is flawed when faced with declining demand, making it
unsuitable for industries with declining or sluggish growth in demand.
This chapter examines PCR in the face of declining demand. Section 2 describes
the effect of reductions in demand on USPS. Section 3 illustrates how falling
demand reduces the ability to cover cost under PCR and what might be done
about it. Section 4 presents the adjustment mechanism under PCR for declining
demand, including a numerical example. Section 5 examines briefly the relevance
of declining demand for regulated industries. Section 6 is by way of summary and
conclusions. The appendix provides formal derivations.
2 The Implications of Declining Demand for USPS
In the case of USPS, the effect of increased unit costs and the loss of revenue on the
reduction in output had serious financial implications. PAEA left it ill-equipped to
address the problems it was facing. Under the previous regulatory regime, it had a
mechanism for addressing this problem, namely, a rate hearing. So, under PAEA
cutting costs was effectively the only viable option perceived by USPS. Reducing
costs can be achieved in a number of ways including improved efficiency of
operations and lowering service quality. However, increasing efficiency faces an
uphill fight against the scale economies as demand decreases.
Given that USPS has scale economies, and therefore the impact of PCR is
severe, it needs to explore other avenues to avoid losses. However, the options
available to USPS under the price cap instituted by PAEA were highly limited.
Entering new lines of business was highly restricted. One option, which it
exercised, was to take advantage of the growing parcel business. Regrettably, this
growth has been insufficient to make up the shortfall resulting from the dramatic
fall in letter volume. This has been for a number of reasons. Parcel delivery is a
much smaller share of revenue than letters. In addition, USPS does not have the
market power in the parcel market that it does in the letter market. Furthermore,
scale economies may be lower in parcels than in letters. Expansion in parcels’
volume is likely to have less of an impact in reducing average costs than the loss in
letter volume is going to have in increasing average costs. In addition, since USPS
does not have the market power in parcels that it has in letters its margin is also
Table 1 USPS financial and operational information, fiscal years 2006 through 2013
Net income Career employees
Fiscal year ($ billions) Mail volume % decline (thousands)
2006 0.9 213 696
2007 (5.1) 212 0.47 685
2008 (2.8) 203 4.25 663
2009 (3.8) 177 12.81 623
2010 (8.5) 171 3.39 584
2011 (5.1) 168 1.75 557
2012 (15.9) 160 4.76 528
2013 (5.0) 158 1.25 491
Source: General Accounting Office (2013)
likely to be lower. The gaping hole left in the letter mail volume decline was not
going to be filled by increased parcel volume.
Given the limited options, USPS faced cutting costs by reducing service stan-
dards was management’s principal option. USPS proposed eliminating Saturday
delivery but without success. It was successful in reducing service quality. Mail
now takes at least a day longer to deliver, and overnight delivery locally has been
almost eliminated. The service reductions that have been introduced have had the
effect of making USPS products less valuable, which has contributed further to the
reduction in volume. Service reduction is another area where PCR can be problem-
atical. If the welfare-maximizing level service quality is higher than the profit-
maximizing level, the firm has an incentive to cut service standards.4
The measures taken by USPS were unsuccessful in becoming profitable or even
breaking even since 2006. So, in 2010, USPS filed an Exigent Rate Case before the
PRC resulting in a temporary price increase in excess of the CPI. Table 1 summa-
rizes the results for USPS since 2006. Every year since 2006, it has lost money. This
is despite the growth in parcel traffic and the cost savings from reductions in the
service quality.
A different approach is needed and this is the subject of what follows—the need
to modify the price cap formula to adjust for changes in demand.
3 The Declining Demand Problem Under PCR
PAEA was introduced at a time when the volume of mail peaked. It has declined
dramatically every year since then, albeit at a decreasing rate recently. Declining
demand has serious implications for the firm under PCR. Figure 1 illustrates the
baseline case in which a PCR firm is just covering cost, where Pcap is the regulated
4
Sappington (2005) provides an analysis of this issue. PAEA showed some recognition of this
issue in that it required USPS to seek an Advisory Opinion from the PRC when it is sought to
change service standards.
Pcap
AC
MC
D
Q
Q0
Fig. 1 PCR with price exactly covering cost
price under PCR, AC and MC are respectively average cost, marginal cost, and D0
is demand.5 The price is set to just equal average cost, as it would be under COS
regulation. This is a baseline to illustrate the solvency problem created by declining
demand.
Suppose the demand falls from D0 to D1, as illustrated in Fig. 2. Average cost
rises because of economies of scale, as output falls from Q0 to Q1. If price does not
rise, the PCR-regulated firm now is unable to cover its costs, purely as a result of the
decline in demand. The loss is shown by the shaded rectangle above Pcap, up to Q1.
The size of the loss depends on the height of the rectangle that, as explained in
the next section, depends on the elasticity of average cost. However, inspecting
Fig. 2 shows that the regulator simply to increase if the regulator simply increased
price to by the height of this rectangle, demand would fall below Q1, the actual
amount depending on the elasticity of demand. This, in turn, raises the required
price. To restore the PCR firm’s ability to cover costs, price has to rise to the point
where the new demand curve intersects the average cost curve, as indicated by
where the arrow is pointing in Fig. 3.
5
For clarity, the focus is on adjusting a capped price for a single product. This price adjustment is
intended to address only the question of solvency as presented by a regulator to the regulated firm,
and not to correct other potential problems, e.g., inefficient cost allocation.
Pcap
AC
MC
D1 D0
Q
Q1 Q0
Fig. 2 Losses to the PCR firm with declining demand
Pcap
AC
MC
D1 D0
Q
Q1 Q0
Fig. 3 The price adjustment to restore the ability to cover cost
4 Price Caps with Changing Demand
Revisiting the description in the introduction of PCR for a single product, the price
cap formula can be written as:
ΔP ΔCPI
¼ X;
P CPI
P is the price cap and ΔP is the change in the cap, so the left hand side is the
percentage change in the price cap. The fraction on the right hand side is the
inflation rate, and the last term is the X factor. An adjustment for changes in demand
needs to be added to the formula. Three variables representing the change in
demand, the resulting change in the average cost, and the elasticity of demand.
Change in demand. This is the percentage change in volume at the allowed price
cap, assumed to be the price the regulated firm is charging. This percentage change
in volume, Z, is assumed constant over the range of price between the current price
cap and the price cap adjusted to reflect changes in demand. If Q is the quantity
sold, the percentage change in volume defined as Z, as the percentage change in
price above, is
ΔQ
Z¼
Q
where ΔQ is the change in the change in volume. For declining demand, ΔQ and
thus Z are both negative.
It is important that this change in sales is independent of actions taken by the
regulated firm. If the firm were to be compensated by a price adjustment for every
reduction in demand, it would be able to charge a higher price by taking actions to
reduce demand, for example, reducing service quality. Similarly, a firm that
undertook efforts to improve quality and thus increase demand would be penalized
by a reduction in price. These would constitute perverse incentives to reduce
product quality. Accordingly, it should be clear that any price adjustment be in
response to only changes in demand outside the control of the regulated firm.
Average cost elasticity. While under PCR, prices are not set by regulators to
equal average cost—otherwise, the firm loses its incentive to minimize its costs—
prices need to be set with the expectation that the regulated firm will be able at least
to cover average cost. The intent is to adjust price by a percentage. So, how much
the percentage change in demand Z will change average cost, AC is required. The
term for this is the elasticity of average cost, eAC, defined as
ΔAC Q
eAC ¼ :
ΔQ AC
Multiplying this elasticity by the percentage change in demand will give the
percentage change in average costs. Since regulation generally applies to firms
with natural monopoly derived from significant economies of scale, average cost
falls as output increases, resulting in eAC < 0.
One problem is that this elasticity may not be easy to measure. The appendix
shows how to derive this elasticity from the elasticity of costs with respect to
quantity, which could be estimated statistically. However, a simple way to
understand and assess the magnitude of this elasticity is to calculate it from a

familiar example. Suppose that the cost C(Q) of providing Q units of a particular
service is given by
CðQÞ ¼ F þ MQ;
where F is fixed cost and M is a constant marginal cost for each unit of the service
delivered. Then,
ACðQÞ ¼ CðQÞ=Q ¼ F=Q þ M:
and
F
eAC ¼ ;
F þ MQ
The elasticity of average cost is the negative of the ratio of fixed cost to total
cost.6 If variable costs MQ are low relative to fixed costs, this elasticity will be close
to 1, its smallest value. In the extreme, if all of the costs are fixed (M ¼ 0), the
elasticity of average costs is 1; reducing quantity by any given percentage
increases average cost by the same percentage. In the other direction, if fixed
costs are low relative to variable cost, this elasticity will be close to zero—zero if
fixed cost F equals 0.7 In this case, price would not require adjustment for falling
demand because average cost would be M whether demand fell (or rose).8
Elasticity of demand. It might seem that this is all that is needed, since the
percentage change in demand times the elasticity of average cost with respect to
output gives the percentage change in average cost, which should determine how
much price would be adjusted. However, when price is adjusted, that too will have
an effect on quantity sold (unless demand is perfectly inelastic), which will in turn
affect how much average costs change. This effect means that the percentage
change in average cost will be the elasticity eAC times the sum of 2 % changes in
6
Another expression shown in the appendix for eAC is that for general cost functions, it equals
MC/AC 1, where MC is marginal cost. When MC is constant, the expression in the text follows.
Because the elasticity of total cost with respect to output is MC/AC, eAC equals the elasticity of
total cost minus 1.
7
To the extent that parcels have higher variable costs relative to letters, their scale economies are
lower than that of letters.
8
In theory, basing an adjustment formula on a parameter based on the ratio of fixed to total cost
could provide an incentive for the regulated firm to inefficiently substitute fixed for variable costs.
However, this appears unlikely. These adjustments would take place after the regulated firm
decides the production process that determines the ratio of fixed to variable costs, and before
declining demand would be expected. Moreover, as discussed below, this adjustment formula is
symmetric, in that it suggests price reductions if demand increases and the firm is better able to
cover its costs for reasons beyond its control. Increasing F/AC will lead to bigger automatic price
reductions in that case.
quantity sold: Z, from general shift in demand, and this second effect induced by the
adjustment itself.
The measure of this second effect on quantity sold is the percentage change in
price ΔP/P times the elasticity of demand for the service, eD. This means that the
adjustment to ΔP/P is determined by an expression that also has ΔP/P as part of the
adjustment itself. Specifically,

ΔP ΔP
¼ eAC Z þ eD
P P
Solving this for the percentage change in price—a derivation is provided in the
appendix—gives the expression to adjust for changes in demand Z:

ΔP eAC
¼Z :
P 1 eAC eD
The PCR equation incorporating the adjustment is then

ΔP ΔCPI eAC
¼ XþZ :
P CPI 1 eAC eD
The added adjustment term is based on three parameters, the change in demand, the
elasticity of average cost, and the elasticity of demand.9
Before getting to some illustrative examples, some general observations are in
order. It may be useful to look at some extreme cases. To reiterate, if demand is
perfectly inelastic (albeit at a lower level—a leftward shift of a vertical demand
curve), eD is zero. In that case, the denominator in the brackets is 1, and the
adjustment is simply Z[eAC], where price increases by the same percentage as
average cost resulting from the decline in demand. If eAC is zero, the expression
in the brackets is zero, and there would be no price adjustment, as average costs do
not change, and thus costs are covered without raising prices.
Typically, scale economies are present in situations where PCR is applied. Scale
economies, that is, declining average cost, will imply that eAC is negative. Because
the elasticity of demand is also negative, the product in the denominator above,
eACeD, is positive. If those elasticities are sufficiently small so their product is less
than 1, the expression in the brackets is negative. When demand falls, that is, when
Z is negative, the effect on price will be a negative number times the negative
expression in the brackets, implying a positive price adjustment. This effect is
expected. In most regulated sectors demand is inelastic, eD having absolute value of
9
If as envisioned this adjustment factor is applied annually, as are the other terms in this
expression, estimates of these elasticities are likely to be constant as the annual change in demand
is not likely to be large. However, these are likely to change over time, as does the inflation rate,
and the adjustment each year should reflect this.
Table 2 Price adjustments for a 10 % fall in demand

Elasticity of demand eD
Z ¼ .1 .2 .3 .4 .5 .6
Elasticity of average cost eAC .1 .010 .010 .010 .011 .011
.2 .021 .021 .022 .022 .023
.3 .032 .033 .034 .035 .037
.4 .043 .045 .048 .050 .053
.5 .056 .059 .063 .067 .071
less than one. The analysis above shows that eAC is also less than one in absolute
value when there are scale economies and positive variable costs. Consequently,
their product will generally be less than one.10 It is also worth observing that the
effect is symmetric. If demand grows for reasons outside the control of the
regulated firm, Z is positive, implying that prices could be adjusted downward,
increasing consumer welfare while not affecting the PCR firm’s ability to cover its
costs.
The numerical example in Table 2 provides a sense of the impact of the
adjustment. It shows the percentage adjustment to price from a fall in demand of
10 %, for demand elasticities eD between .2 and .6 and average cost elasticities
eAC between .1 and .5.
Absent the correction from using the elasticity of demand, the numbers in the
table would just equal the elasticity of average cost values on the left times the
change in demand, .1. For low values of both demand and average cost elasticity,
this is the case; when average cost elasticity is .1, the adjustment factors are
essentially just those. When both elasticities are large in absolute value, but less
than unity, the size of the adjustment can be substantial. For the largest elasticities
in the table, the price adjustment for a 10 % reduction in demand would be to
increase it by 7.1 %; with no correction for the demand effect, it would be 5 %.
For the current situation in the postal sector price elasticities in the range 0.3 to
0.4 seem reasonable.11 The range for cost elasticity is somewhat more difficult to
gauge, and its choice is based almost entirely on judgment. Combining an average
cost elasticity of 0.3 with a price elasticity of .3, Table 3 shows the price
adjustments for 2007–2013 implied from the changes in demand in Table 1.
10
If eACeD equals or exceeds one, the effect goes in reverse—falling demand that leads to higher
average cost would imply that the firm cuts price. The appropriate economic interpretation of this
perverse calculation is that if the elasticity of demand is sufficiently large, greater in absolute value
than 1/eAC, then there is no price adjustment with positive volume by which the firm could cover
its costs following a decline in demand. That is not a mathematical curiosity; it may be that any
attempt to increase price would reduce demand by so much that the revenue collected would be
less than the costs that remain. Under those conditions, the firm essentially lacks the market power
necessary to cover its costs. Therefore, the adjustment factor will be most likely useful in settings
where demand is inelastic at the regulated price before demand falls—as one would expect when
firms are regulated—and that demand has not fallen so far as to change that.
11
In the recent Exigency Case, USPS’ implicit demand elasticities were close to 0.3.
Table 3 Price adjustments Year Change in demand (%) Price adjustment (%)
for falling demand,
2007 0.47 0.15
2007–2013
2008 4.25 1.40
2009 12.81 4.22
2010 3.39 1.12
2011 1.75 0.58
2012 4.76 1.57
2013 1.25 0.41
The cumulative real price increase suggested by the data over this period is
9.8 %.12 These figures are only illustrative, in light of assumptions regarding the
relevant elasticities and the lack of adjustment for demand reductions driven by
prices rather than outside factors. However, this illustration indicates that the
potential adjustment under PCR due to falling demand could be quite substantial.
5 Relevance Across Regulated Sectors
While the motivation for this framework and its application is to the postal sector, it
could have applications in other sectors where the regulated firms have seen
declining demand, as summarized in Table 4.
Voice telephone service declined between 1992 and 2011, the percentage of
households with landline telephone service falling from 94.7 % in 1992 to 75 % in
2011.13 FCC data most recently for 2013 indicate that this trend has continued,
within connections provided by incumbent local telephone companies falling 16 %
from December 2011 to December 2013.14 This decline is coincident with rapid
growth in mobile telephony and telephone service provided by nonincumbent local
carriers providing telephone service over the Internet rather than through traditional
circuit-switched networks.
The decline in electricity service at this point is more prospective. Electricity
demand growth has fallen dramatically last 60 years, from around 11 % per year in
the late 1950s to barely positive, .02 %, between 2008 and 2013, with demand
falling in some of those years.15 Some of this was likely an anomaly due to the
12
Recall that under PCR prices should already be adjusted for inflation.
13
U.S. Census Bureau, “Extended Measures of Well-being: Living Conditions in the United
States, 2011,” Table 10, available at http://www.census.gov/hhes/well-being/publications/
extended-11.html.
14
Calculated from data in Federal Communications Commission, “Local Telephone Competition
as of 2013,” Table 1, p. 12, available at http://transition.fcc.gov/Daily_Releases/Daily_Business/
2015/db0219/DOC-329975A1.pdf.
15
Calculated from data in Energy Information Administration, Annual Energy Outlook 2014,
“Market Trends: Electricity Demand,” available at http://www.eia.gov/forecasts/aeo/MT_electric.
cfm#growth_elec.
Table 4 Factors leading to declining demand in regulated sectors

Industry Regulated provider Disruptive technology, factors
Mail State postal operator, Internet delivery, online bill payment, e-government
delivery e.g., USPS
Voice Landline telephone Mobile telephone service, digital voice-over-Internet
telephony company protocol (VoIP)
Electricity Local distribution utility Distributed generation, energy efficiency mandates,
emissions controls
Gas Local distribution utility Conservation, energy efficiency
Water Local distribution utility Conservation, drought
recession following the collapse of credit markets in the USA in 2007, but the long-
term trend remains striking. Moreover, the US utility industry is concerned that the
trend is likely to get worse, with declining demand threatening the financial
solvency of local electricity distributors (Kind 2013). Part of this long-term decline
in growth is the result of the decline in manufacturing and the more efficient use of
electricity.
More recently, specific factors eliciting concern include on-site solar power and
other technologies that “threaten the centralized utility model” and “energy effi-
ciency and DSM [demand side management] programs that also promote reduced
utility revenues while causing the utility to incur implementation costs”(Kind 2013
at 3). In the case of electricity, “decoupling” programs that attempt to guarantee
local distribution utility revenues may protect utilities, and reduce their incentive to
oppose public policies to promote efficiency programs that reduce electricity use
(Brennan 2010). The modification of PCR is one way to address the political or
economic need to institute programs to increase prices when usage falls.
Gas distribution and water utilities face similar problems. In the case of gas and
water per customer usage is being reduced by conservation programs. Water has
strong scale economies and is subject to considerable fluctuations in usage as a
result of restrictions imposed in time of drought (Crew and Kahlon 2014). A further
reform of gas, water, and electricity would be to recover more of the fixed costs of
distribution through monthly service charges. This would make distribution utilities
less vulnerable to the effect of falling demand (Brennan 2014 discusses rate reform
in the case of electricity).
6 Summary and Conclusion
Price cap regulation has significant advantages, in that it may induce regulated firms
to operate efficiently and reduce the transactions costs of regulation. However, PCR
can leave a regulated firm unable to cover costs when demand falls because with
scale economies unit costs increase as output declines. An adjustment to the price
cap formula, based on the elasticity of average cost with respect to output and the
elasticity of demand, can restore solvency to a price-cap-regulated firm. The

adjustment proposed to the price cap retains the incentives for internal efficiency
of PCR as, like the CPI adjustment it is exogenous to the company. All that is
required is to prove that volume has fallen, which is very different from cost of
service regulation, which requires proof that costs have increased.
The elasticity of average cost—the percent by which average cost rises or falls as
demand falls by a given percentage—can be expressed as a simple relationship
between marginal and average cost. When marginal cost is constant, the elasticity
of average cost is the negative of the ratio of fixed cost to total cost. Illustrative
examples suggest that increasing the price cap based on this adjustment could have
substantially mitigated the financial straits in which USPS found itself in recent
years. The effect is symmetric, in that if demand is increasing, the same formula
could be used to reduce prices, so consumers could share in the benefits of scale
economies created by expanding supply.
This remains a work in progress. To apply it appropriately, authoritative esti-
mates of the elasticity of demand and average cost. In practice there would be a
number of complexities. For example, a separate price adjustment for each class of
USPS products might be required as demand elasticities and perhaps scale econo-
mies differ across products. Applying the formula in this case may require relying
on regulatory rules that determine the size of the fixed cost contributions from any
particular product, as the basis for the cost recovery expected under PCR. In
addition, price adjustments to reduce the chance of insolvency reduce risk for the
regulated firm and, in principle, should reduce its cost of capital and, in the long run,
price. This shifts price risk to consumers, however, so the net effect taking risk into
account remains to be analyzed. Finally, it is important to remember that the
formula assumes that the quality of service is unchanged, in part to prevent an
inadvertent incentive to reduce demand by cutting quality in order to raise the cap.16
Acknowledgments The authors wish to thank Darryl Biggar, Khai Duong, Anthony Paul, Marc
Smith, Soterios Soteri, Menahem Spiegel, Ian Streule, Sheldon Switzer, Rodger Woock, Eric
Woychik, and participants in the CRRI March 27, 2015 workshop on the economics of postal
reform, 34th Eastern Conference, and 28th Western Conference, as well as participants in the 23rd
Postal Conference. They are not responsible for any errors.
Appendix
1. Equations for the change in average cost as output, the derivative of that change,
and the elasticity of average cost with respect to output
Let C(Q) be the total cost of producing Q, with average cost AC(Q) ¼ C(Q)/Q.
The derivative of average cost with respect to Q is
16
How PCR might be adjusted for service quality is a separate and complex question (Sappington
2005).
0
0 dAC QC C
AC ¼ ¼ :
dQ Q2
Rewriting C’ as MC, marginal cost, and dividing both the numerator and
denominator by Q gives
0 dAC MC AC
AC ¼ ¼ :
dQ Q
For natural monopolies with scale economies throughout, this expression is

negative, implying that MC < AC. The elasticity of average cost with respect
to Q, eAC, is then
dAC Q MC AC Q MC
eAC ¼ ¼ ¼ 1 < 0:
dQ AC Q AC AC
Because MC/AC is non-negative but less than 1 for firms with scale economies,
eAC is negative, greater than or equal to 1.
The second derivative of average costs shows whether the effect of changing
output on average cost increases or decreases with scale. If that second deriva-
tive is positive (negative), the derivative of average cost falls (rises) in absolute
value as scale increases. From the second expression above for AC0 ,
0 0
00 Q MC AC ½MC AC
AC ¼ :
Q2
Because Q[AC0 ] ¼ MC AC,

0 0
00 Q MC 2½MC AC MC 2½MC AC
AC ¼ ¼
Q2 Q Q2
The first term in the denominator on the right, MC0 , is non-negative if marginal
cost is constant or increasing, and because AC0 is negative when there are scale
economies, AC00 is positive in situations where prices are likely to be regulated.
Consequently, the larger the effect of a change in output on average cost, the
smaller the regulated firm; all else equal.
It may be useful to show these results for the setting, where costs C(Q) are
given by a fixed cost F and a constant marginal cost M, that is,
CðQÞ ¼ F þ MQ:
In that case,
F
AC ¼ þ M;
Q
QM
F
0 MC AC F
AC ¼ ¼M ¼ 2 < 0;
Q Q Q
MC M
eAC ¼ 1¼ F F
Q þ M 1 ¼ FþMQ 1 ¼ ¼ FþMQ < 0;
AC MQ MQFMQ
FþMQ
And, because MC0 ¼ 0.
00 2½MC AC 2F
AC ¼ ¼ 3>0
Q2 Q
2. A general derivation of the price adjustment formula with changes in demand

It is shown below that if demand changes by a given percentage Z, the price
p under a price cap should be adjusted in percentage terms (%Δ) by

eAC
%Δ p ¼ Z ;
1 ed eAC
where ed 0 is the elasticity of demand and eAC is the elasticity of average cost
with respect to output. If the firm has scale economies, eAC < 0.
To see why this holds, let demand be kD( p), where p is price and k is a
parameter reflecting a possible shift in the demand curve D. Z above will be the
percentage change in k, that is, dk/k. The percentage change in price will be dp/p.
To preserve the ability to cover cost regardless of the value of k, price p has to be
chosen to satisfy
pk½Dð pÞ ¼ CðkDð pÞÞ; ð1Þ
where C is the total cost function.

It is necessary to state dp/p as a function of dk/k for inclusion in the price cap
formula. To get this, first implicitly differentiate (1) to get
0 0 0 0
d p½kD þ dk½ pD þ pkD ½d p ¼ C D½dk þ C kD ½d p ð2Þ
Divide the left side of (2) by the left side of (1) and the right side of (2) by the
right side of (1) to get:
0 0 0 0
d p dk D d p C D dk C kD d p
þ þ ¼ þ :
p k D C C
Multiply the last term on the left by p/p, the first term on the right by k/k, and the
last term on the right by pD/pD, and get
0 0 0 0
d p dk D p d p C kD dk C kD D p d p
þ þ ¼ þ : ð3Þ
p k D p C k C D p
Because the elasticity of demand ed ¼ D0 p/D, (3) can be rewritten as

0 0
d p dk dp C kD dk C kD dp
þ þ ed ¼ þ ed : ð4Þ
p k p C k C p
Multiplying and dividing the first expression on the right side of (3) and (4) by
kD because kD is output, C0 kD/C, is MC/AC. However, it is more useful to
invoke the following relationship derived above:
MC=AC ¼ 1 þ eAC ; ð5Þ
Substituting (5) into (4) gives

d p dk dp dk dp
þ þ ed ¼ ½eAC þ 1 þ ½eAC þ 1ed : ð6Þ
p k p k p
Multiplying out the terms on the right gives

d p dk dp dk dk dp dp
þ þ ed ¼ þ eAC þ ed þ eAC ed : ð7Þ
p k p k k p p
Subtracting dk/k and ed[dp/p] from both sides of (7) gives

dp dk dp
¼ eAC þ eAC ed ; ð8Þ
p k p
Leading to

dp dk
½1 eAC ed ¼ eAC
p k
and

d p dk eAC
¼ : ð9Þ
p k 1 eAC ed
This is the expression at the beginning of this section of the appendix, where
dp/p is the percentage change in price and dk/k is the percentage change in
demand, Z, in that expression.
If a regulated firm has scale economies, eAC < 0. If demand falls, dk/k < 0.
Thus, if demand falls where there are scale economies, the price cap should go
up by the above expression to preserve the ability to cover costs under the
original price cap regime. This can happen if the denominator on the right
hand fraction in (9) is positive, that is
1 > eAC eD : ð10Þ
This is a reasonable assumption. If it is not true, when demand falls, if scale

economies are sufficiently great and demand is sufficiently elastic, (10) implies
that no price increase will be able to cover costs. The expression is symmetric.
So, an increase in demand under scale economies would imply a fall in the
price cap.
References
Armstrong M, Sappington D (2007) Recent developments in the theory of regulation. In:

Armstrong M, Porter R (eds) Handbook of industrial organization, vol 3. Elsevier, Amsterdam,
pp 1557–1700
Braeutigam RR, Panzar JC (1989) Diversification incentives under ‘price-based’ and ‘cost-based’
regulation. Rand J Econ 20:373–391
Brennan T (1989) Regulating by ‘capping’ prices. J Regul Econ 1:133–147
Brennan T (2010) Decoupling in electric utilities. J Regul Econ 38:49–69
Brennan T (2014) An expanded distribution utility business model: win-win, or win-maybe?”. In:
Sioshansi F (ed) Distributed generation and its implications for the utility industry. Academic,
Waltham, MA, pp 251–265
Crew M, Kahlon R (2014) Guaranteed return regulation: a case study of regulation of water in
California. J Regul Econ 46:112–121
Crew M, Kleindorfer P (1996) Incentive regulation in the United Kingdom and the United States:
some lessons. J Regul Econ 9:211–225
Kind P (2013) Disruptive challenges: financial implications and strategic responses to a changing
retail electric business. Edison Electric Institute, Washington, DC
Littlechild S (1983) Regulation of British telecommunications’ profitability. Department of Trade
and Industry, London
Littlechild S (1986) Economic regulation of privatised water authorities. Department of the
Environment, London
Sappington DEM (2005) Regulating service quality: a survey. J Regul Econ 27(2):123–154
U.S. Postal Service (2013) Urgent action needed to achieve financial stability. Government
Accounting Office, Washington, DC. Accessed 17 Apr 2013
Scale Economies and Postal Price Caps
in Europe: Declining Volumes, Lower
Productivity, Higher Postage?
Antonia Niederprüm, Christian M. Bender, and Alex Kalevi Dieke
1 Introduction
In many countries, postal regulators have implemented price cap regimes to provide
incentives for postal operators to improve efficiency and to allow more pricing
flexibility. In the price cap formula (RPI X), the X-factor reflects expected
productivity gains that usually reduce the scope of potential price increases such
that average prices increase less than inflation. In a world with growing mail
volume, efficiency gains were driven by economies of scale and scope and inno-
vations in postal operations. Today, most postal operators face declining letter
volumes so that productivity gains resulting from innovations are reduced or offset
by lower economies of scale. Consequently, regulators face the challenge that
average cost may be expected to increase, even where the regulated firm performs
well. As a result, regulators may have to set reduced or negative values for the X-
factor, allowing price levels to increase more than the rate of inflation.
Section 2 summarizes key trends in letter volume development and pricing in
selected European countries (Belgium, France, Germany, Ireland, the Netherlands,
Portugal, Sweden, and the UK). These countries account for more than 70 % of EU
letter post volume (WIK-Consult 2013a). Section 3 introduces price cap mecha-
nisms and compares price cap regimes, and X-factors implemented by regulators in
these countries. The chapter discusses specifically how regulators apply the price
cap formula to address both aspects—incentives for efficiency improvements and
This chapter presents the personal views of the authors only and should not be interpreted as the
views of Wissenschaftliches Institut für Infrastruktur und Kommunikationsdienste (WIK) or of
any clients of WIK.
A. Niederprüm (*) • C.M. Bender • A.K. Dieke
WIK, Bad Honnef, Germany
e-mail: a.niederpruem@wik.org

20 A. Niederprüm et al.
expected declines in mail volume. The chapter concludes that price cap regulation
can continue to provide incentives for efficiency in a declining market and presents
some recommendations for price cap regulation.
2 Volume and Price Trends
Letter post volume is declining in all countries for a number of reasons. National
postal operators (POs) face demand losses for structural reasons (e-substitution), a
weak economic environment and, in some countries, competition (e.g., in Germany,
the Netherlands, and Sweden).
Figure 1, on the left hand side, shows the letter post items per capita in relation to
the POs’ letter post volume and highlights average volume decline since 2010 on the
right hand side. It is clear that the POs are quite differently affected by volume decline.
While the Dutch universal service provider PostNL is strongly challenged with an
average volume decline of more than 9 % the German and the Belgian postal operators
are much less affected with an average decline of 1.5–2 % between 2010 and 2013. All
POs but An Post and CTT Correios deliver more than 150 letter post items per capita.
Four of the nine POs still deliver at least 200 letter post items per capita.
Figure 2 presents on the left hand side the current level of public tariffs related to
20 g letters delivered the next working day and, on the right hand side, the average
price change per year for single-piece letters in nominal and real terms for the period
2010–2015. Average price increases for priority single-piece items vary between
2.2 % (bpost, Belgium) and 9.4 % (PostNL, the Netherlands) in nominal, or 0.3 and
7.4 % in real terms. The POs with the lowest price increase (bpost and Deutsche
Post) are also the least affected by volume decline (see Fig. 1). The correlation
coefficient between the average volume decline (2010–2013) and average tariff rise
is 0.78 in nominal and 0.84 in real terms. This indicates that with increasing
volume decline the average increase in single-piece tariffs is also higher (keeping in
mind that, in most countries, these tariffs are subject to price cap regulation).
Letter post items per capita (2013) Average volume decline (2010-2013)
250
Deutsche Post
200 bpost
PostNord (SE)
150
An Post
100 La Poste
Royal Mail
50
CTT Correios
0 PostNL
bpost PostNord La Poste Royal PostNL Deutsche An Post CTT
Posten Mail Post Correios -10% -8% -6% -4% -2% 0%
Fig. 1 Per capita volume and volume developments of the national postal operators. Sources:
Based on regulators’ market reports and POs’ annual reports, UPU statistics, and Eurostat
(population data). Notes: Average volume growth is calculated as compound annual growth rate
(CAGR)
Scale Economies and Postal Price Caps in Europe: Declining Volumes, Lower. . . 21
20g Letter tariff as of 1 April 2015 (D+1) Average tariff increase (20g D+1 letters, 2010-2015)
0.90 €
PostNL
0.80 €
Royal Mail
0.70 €
Average tarif
0.60 € An Post growth in
nominal terms
0.50 € La Poste
0.40 € CTT Correios
Average tarif
growth in real
0.30 € terms
PostNord / Posten
0.20 €
Deutsche Post
0.10 €
0.00 € bpost
CTT Deutsche An Post PostNL La PostePostNord bpost Royal
Correios Post Posten Mail 0% 2% 4% 6% 8% 10%
Fig. 2 Single-piece rates (in EUR) and average price increases (2010–2015). Sources: Eurostat
(Exchange rate and consumer price index), postal operators’ price lists (year-end tariffs). Notes:
Foreign currency translation based on average annual exchange rate in 2014. Average tariff
increase in real terms calculated as difference between nominal average tariff change (CAGR
2010–2015) and average annual growth of the consumer price index (CAGR 2010–2014). In
Sweden (PostNord/Posten), the single-piece tariff includes VAT (25 %)
Generally, bulk mail tariffs are less transparent than public single-piece tariffs.
There are indications that bulk mail tariffs also increased in some countries but to a
lesser extent than single-piece tariffs. La Poste, for example, raised their postal
tariffs by 7 % on average in January 2015. While the single-piece tariff of 20 g
letters increased by 11–15 %, tariffs for bulk mail and direct mail were only
raised by 1.4–3 % (La Poste 2014). We observed similar trends at Deutsche
Post, PostNord/Posten, and PostNL (Thiele and Niederprüm 2015, p. 5).
POs increasingly use their pricing flexibility to de-average single-piece and bulk
mail tariffs.
3 Postal Price Caps in Europe
Many European regulators apply price cap mechanisms to regulate postal tariffs
(ERGP 2014). The price cap usually refers to a bundle of postal services (“basket”)
which allows the regulated PO to rebalance the price structure among the basket
services within the limits of the cap.
Xn
pi, t qi, t1
X ni¼1 1þIX
i¼1
p i, t1 q i, t1
The price cap formula is determined by the inflation rate I (usually approximated by
the percentage change of the consumer or retail price index) and the “X-factor”.1 In
theory, the X-factor represents the projected change in productivity in postal
1
For the theoretical underpinning and the properties of price cap regulation, see for example
Beesley and Littlechild (1989) and Brennan (1989).
operations.2 If a positive value is determined for the X-factor, weighted average

price will be allowed to increase less than inflation. If the projected productivity
rate declines, e.g., due to volume decline, average prices are allowed to increase
more inflation. In case that the company manages to raise productivity more than
projected, it can retain the realized “excess” profits. The effectiveness of efficiency
incentives depends on the price cap period and the credibility of the regulator’s
commitment to maintain the price cap design during the term. This is considered as
one of the major advantages of the price cap mechanism compared to the rate of
return regulation (Armstrong and Sappington 2007, p. 1606 et seqq).
The nature of postal service provision is characterized by substantial fixed costs.
Volume growth in combination with efficiency gains resulted in lower average
costs and thus improved profitability of postal operations. With more rapid decline
in mail demand, the situation has changed. Due to declining demand in letter
services, average costs now increase and therefore—under the given price cap
regime—the regulated company runs the risk to make losses.
Price cap regulation becomes more complex in a dynamic environment when the
cost and the demand conditions of the regulated company change during the price
cap period.3 We observe that price cap mechanisms were often kept rather simple
during times of growing mail volume. With declining mail volumes, regulators may
decide to include volume, revenue, and cost effects into the mechanism. In many EU
countries, the period of growing mail volume coincided with the efforts of former
postal administrations to modernize and become more commercial entities. How-
ever, price cap mechanisms that include the volume effect on average costs would
also have been an appropriate tool to skim average cost savings in times of growing
mail volume for the benefit of postal customers. From the viewpoint of the regulated
company, however, a simple price cap mechanism based mainly on the development
of the consumer price index has clear advantages when demand is growing.
Table 1 compares recent price cap decisions taken in Belgium, France,
Germany, Ireland, the Netherlands, Portugal, Sweden, and the UK.4
Tariff increases are always linked to the development of the consumer price
index. In three countries (Belgium, Sweden, and UK), it was decided not to include
an X-factor into the price cap model. While in Sweden the cap is limited by the
inflation rate (Swedish Postal Ordinance 2010), the Belgian postal operator bpost
can qualify for additional price increases by outperforming a weighted bundle of
quality of service targets (BIPT 2014). In both countries, the price cap formula was
2
In practice, the X-factor is often the result of a bargain between regulators and regulated firms on
profit sharing.
3
Crew and Kleindorfer (2008) already highlighted that more flexibility in price cap regulation
would allow postal operators sufficient commercial freedom to finance their operations under full
market opening, including the USO.
4
The presented X-factors (fourth column) refer to the value emerging in case that, given the
underlying formula, all assumptions would be met during the period of validity (fifth column).
Price adjustments are possible each year of the period (by request of the PO) and usually refer to
the PO’s financial year.
Table 1 Price cap models in Europe

Services included (one Period of Determined
Country basket) Formula X-factor validity by
Belgium Single-piece letters and ΔCPI Not Without a Decree
parcels + quality included fixed term
bonus
France Single-piece and bulk ΔCPI X 3.5 % 2015–2018 Regulator
letters, single-piece decision
parcel
Germany Single-piece letters (up to ΔCPI X 0.2 % 2014–2018 Regulator
1 kg) decision
Ireland Single-piece and bulk ΔCPI X 14.98 % 2014/ Regulator
letters (2014/15) 15–2018/ decision
Includes sub-cap for let- 1.35 % 19
ters (up to 2 kg)
The Single-piece standard ΔCPI X Pending Without a Decree
Netherlands and registered letters and fixed term
parcels
Portugal Single-piece and bulk ΔCPI X 1.6 % 2015–2017 Regulator
letters and single-piece decision
parcels
Sweden Single-piece letters (up to ΔCPI Not Without a Decree
500 g) included fixed term
UK Single-piece mail (only 53 % Not 2011/ Regulator
non-priority mail up to + ΔCPI included 12–2018/ decision
2 kg) 19
Sources: Regulators’ decisions (ARCEP, BNetzA, ComReg, ANACOM, and Ofcom) and Decrees
(Belgium, the Netherlands, and Sweden)
determined by decree without a fixed term and cannot be adapted by regulators’

decision.
PostNord/Posten is limited to the change in CPI and thus to has ensure that
average costs do not rise more than inflation in order not to jeopardize its financial
stability. Assuming that input costs rise at the same rate as the consumer price index,
PostNord/Posten fully bears the risk of average cost increases due to volume decline.
In the past, however, the company successfully managed to reduce costs in pace of
volume decline.5 Compared to the Swedish model, bpost has some additional scope
for price increases as long as the company outperforms the quality targets. Volume
decline is still low in Belgium, and bpost can fully retain its efficiency gains. This is
reflected in impressive profit margins of more than 20 % in the mail segment (bpost
2015, p. 51). In 2012, the British regulator Ofcom limited the basket of price-
regulated services to three single-piece non-priority services (letters, large letters,
5
The business segment PostNord/Mail Communications Sweden published profit margins of more
than 5 % until 2012; only in 2013 the margin reduced to less than 4 % (see PostNord, Annual
Reports). See also WIK-Consult (2013b) for more information on efficiency efforts made by
PostNord in Denmark and Sweden.
and packets up to 2 kg) and allowed a 53 % jump plus change of CPI (Ofcom 2012).
This prospect of getting higher prices if costs are higher basically weakens Royal
Mail’s incentives to become more cost-efficient. Consequently, Ofcom monitors
Royal Mail’s efforts to achieve efficiency gains, separately.
In the remaining five countries, price cap models address changes in volume,
revenue, and cost. In the Dutch case, the design of the price cap model has recently
been determined by decree. The Dutch regulator ACM is currently implementing
the model; the final decision is still pending (ACM 2014). In the other four
countries, the postal regulators are responsible for deciding on the design of the
mechanism and the calculation of the price cap.
The level of the X-factors (see Table 1) varies among the four countries. Only the
German regulator set a positive X-factor (0.2 %) which implies that the basket
tariffs can increase less than the inflation rate. The German regulator expects that
the increase in average costs by expected volume decline is overcompensated by
cost savings from efficiency gains (Bundesnetzagentur 2013, p. 34–35). Moreover,
volume decline at Deutsche Post was low compared to the other operators (see
Fig. 1). The other three regulators set negative X-factors ranging from 1.35 to
3.5 %. Additionally, the Irish regulator ComReg allowed An Post to substantially
increase the price level in the first price cap period by nearly 15 % to “enable a more
prompt return to an appropriate level of profitability for the price-controlled
universal postal services” (ComReg 2014, paragraph 86, p. 24). The French regu-
lator ARCEP allowed La Poste to consume half of the total scheduled 4-year cap
increase in the first year, i.e., in 2015 (ARCEP 2014, p. 7).
To incorporate the effects of demand and cost changes in the X-factor regulators
need assumptions for (1) future volume developments; (2) cost variations due to
volume changes; and (3) cost savings generated by efficiency gains. Additionally,
regulators take a reasonable rate of profit into account (either cost of capital,
WACC, or return on turnover).
All regulators rely on forecasts of basket volumes usually provided by the
regulated company and cross-checked by the regulator.6 The regulators apply
different approaches how to deal with deviations between the actual and projected
volume. The German regulator can opt to abrogate the decision or parts thereof in
case of substantial and cost-relevant deviations (Bundesnetzagentur 2013, p. 3). In
the Irish case, An Post can request a review of the price cap decision after 3 years
earliest. If ComReg accepts this request, the price cap model will be subject of a
new consultation. To limit the probability of an amendment after 3 years, ComReg
introduced a “buffer” mechanism by allowing for a 0.5 % points higher return on
turnover (ComReg 2014, p. 8).
6
The basket volume is usually different from the total letter volume of the PO. As indicated in
Table 1, the scope of the basket is limited to selected services, in some countries only to single-
piece services. For this reason, the basket volume can additionally be affected by migration from
regulated to non-regulated letter services. Depending on the allocation of (fixed) costs between
regulated and non-regulated services, this can further promote increasing average costs of basket
services.
In France, Portugal and the Netherlands, the price cap models have built-in
mechanisms in case of diverging projected and actual volume growth. If volume
declines more than projected, the X-factor changes in the French case by 0.7
percentage points per 1 % difference. Either the regulator has to take the initiative
or the regulated company, La Poste, can apply for adjustment (ARCEP 2014, p. 2 of
the appendix). In the Portuguese case, the X-factor declined by 0.375 percentage
points per 1 % difference.7 ANACOM, the Portuguese regulator, limited the
maximum size of the adjustment factor to [1.9 %, 1.9 %] which corresponds to
a maximum deviation of 5 % in volume growth. In the Dutch case, the formula
will be applied year by year, and it is foreseen to replace projected by actual volume
growth once available.8
The basket volume usually does not correspond to total mail volume processed
by the PO. We estimate that the proportions vary approximately between 20 and
90 % of total letter volume. Changes in the basket volume can be due to migration
to non-regulated letter services (e.g., from single-piece to bulk mail products)
and/or migration to other communication channels (e-substitution).9 While the
first option is partly under control of the regulated company, the second option is
not. Migration to non-regulated services becomes more important the lower the
proportion of the basket volume on total volume. This is generally the case where
the price cap is limited to single-piece mail services (i.e., in the Netherlands and
Germany). While this aspect is not considered in the Dutch approach (Staatsse-
cretaris van Economische Zaken 2015, Bijlage 3), the migration effect is included
in the decision of the German regulator (BNetzA 2013, p. 28).
All regulators took account of cost effects related to volume change. The cost
marginality is determined by the cost elasticity and varies between 0.25 in
ANACOM’s decision (ANACOM 2014, p. 28) and 0.45 in ARCEP’s decision
(Lions 2015, p. 9). ComReg assumes a cost elasticity of 0.36 (ComReg 2014). If
volume declines by 1 %, total cost falls less than proportional by 0.25–0.45 %.
Three regulators (Germany, Ireland, and Portugal) take potential cost savings
resulting from efficiency gains into account for calculating the X-factor. These cost
savings are considered to be independent from volume change. The Portuguese
regulator assumed cost savings of 1.27 % per year; but only half this rate is used to
reduce the X-factor. The Irish regulator assumes 2 % cost savings per year due to
efficiency gains.
7
This figure incorporates the change in costs due to volume change which is assumed with
0.25 volume growth rate (proportion of variable costs, see Table 2). Fifty percent of the change
in fixed costs is borne by CTT Correios. See ANACOM (2014, p. 32).
8
In this volume, Brennan and Crew (2015) discuss the introduction of an adjustment factor into the
X-factor to incorporate volume change (deviations from forecasted figures). The authors propose
how this adjustment factor could be designed.
9
A third option would be migration to competition. However, we consider this option as not being
relevant because only those services which are usually not offered under competition are subject to
ex ante price control.
Table 2 Price cap models with X-factors

Projected
volume
Inflation growth per Other
Country rate year Adjustment factor Cost change parameters
France 1.7 % (pro- 6.3 % Adjustment factors 0.45 (“cost
jection, if actual and elasticity”)
annual projected inflation
average) rate and/or volume
growth diverge
Germany Change of Confidential No adjustment Confidential
CPI as factor (cost base:
reported “efficient
costs”)
Ireland Change of Letters: No adjustment 0.36 (“cost Efficiency tar-
CPI as 4.2 % factor marginality”) get: 2 % per
reported year
Flats: Migration fac-
13.3 % tor: 5 % in the
next 2 years
Packets: Price elastic-
3.6 % ity: 0.22
Return on
turnover: 1 %
(first year) else
3.5 %
The Change of Pending Projections have to Pending Return on
Nether- CPI as be replaced by (proportion turnover: 10 %
lands reported actual numbers of fixed Correction
once available costs) factor if mar-
(refers to estimates gin
on volume growth rate > 10 %
and on the propor-
tion of fixed costs)
Portugal 0.7–1.1 % 2015: Adjustment factor 0.25 (propor- Efficiency tar-
(projection) 4.6 %, if actual and tion of vari- get: 0.63 %
2016: projected inflation able costs) (half of
4.1 %, rate and/or volume projected cost
2017: growth diverge savings of
3.7 % 1.27 % per
year)
Sources: Regulators’ decisions (ARCEP, BNetzA, ComReg, ANACOM, and Ofcom) and Decrees
(the Netherlands)
Overall, the design and the calculation of the price cap substantially differ
among the considered examples. The spectrum includes simple mechanisms
depending only on the inflation rate to highly sophisticated approaches with
substantial information requirements on current and future volume, revenues, and
costs. The majority of the price cap models take into account cost effects due to
volume changes. This reflects regulators’ ambitions to balance cost coverage and
efficiency incentives.
A striking feature of many recently implemented price cap models is the

possibility to adjust central parameters during the price cap period, in light of
volume developments. This automatism can substantially reduce the efficiency
incentives of price cap regulation. In the Dutch approach, the risk of changes in
average costs is fully shifted to postal customers, while in the French and the
Portuguese cases, customers have to bear 70 and 50 % of the cost risk. If the
cost-relevant parameters can be adjusted annually, the price cap mechanism con-
verges to a rate of return regulation, thus reducing efficiency incentives.
4 Conclusions
The comparison of price cap mechanisms shows that their complexity substantially
increases if volume, revenue, and cost changes are taken into account. This con-
firms in an impressive way that even in the postal sector “price caps had become
much more complicated than the simple price cap formulae envisaged in the
original Littlechild report” (Stern 2014). The development of more sophisticated
approaches to postal price regulation was mainly driven by falling mail volumes. In
periods of mail volume growth, however, the approaches were often simpler and
allowed the regulated companies to retain profits that resulted from declining
average costs.
When deciding on price caps in this challenging environment of declining
demand, regulators have to balance the objectives of cost coverage and efficiency
incentives. We identified three groups of decisions. The first group of regulators
still applies very simple price cap mechanisms without consideration of future cost
and demand developments (e.g., in Sweden and the UK). In the second group, the
regulators base the X-factor on projected revenue and cost development (including
the risk of increasing average costs) without the possibility to adjust the X-factor
during the price cap period (e.g., in Germany and Ireland). In the third group, the
price cap mechanisms have built-in mechanisms for volume and cost adjustments,
which can result in annual adjustments of the X-factor during the price cap period.
POs and regulators have recognized the importance of declining demand for
mail and its adverse impact on unit costs. Where this problem is addressed as part of
the price cap formula, where a decline in demand triggers a price increase in the
same manner as an increase in the economy-wide price index, then the efficiency
properties of price caps remain, as long as the formula for the price increase is
similarly unaffected by the firm’s efforts to control costs. However, where the price
increases are based upon the firm’s actual cost recovery, the process begins to look
much more like rate of return or cost of service regulation, which lack the incentives
for efficiency of price caps. For this reason, postal regulators should seek to retain
efficiency incentives when designing built-in mechanisms for X-factor adjustments
under conditions of serious volume decline.
References
ACM (2014) Final decision about 2015 stamp rates postponed, press release 29 August 2014
ANACOM (2014) Decision on the universal postal service pricing criteria, public version
ARCEP (2014) Décision no. 2014-0841 de l’Autorité de régulation des communications
électroniques et des postes en date du 22 juillet 2014 sur les caractéristiques d’encadrement
pluriannuel des tarifs des prestations du service universel postal
Armstrong M, Sappington DEM (2007) Recent developments in the theory of regulation. In:
Armstrong M, Porter RH (eds) Handbook of industrial organization, vol 3. Elsevier B.V.,
Amsterdam, pp 1557–1687
Beesley ME, Littlechild S (1989) The regulation of privatized monopolies in the United Kingdom.
Rand J Econ 20(3):454–472
BIPT (2014) Décision du conseil de l’IBTP du 8 Octobre 2014 concernant l’analyse de la
proposition tarifaire de bpost des tarifs pleins a la piece pour l’année
bpost (2015) Financial report 2014
Brennan TJ (1989) Regulating by capping prices. J Regul Econ 1(2):133–147
Brennan TJ, Crew MA (2015) Price cap regulation and declining demand. In: Crew M, Brennan TJ
(eds) The future of the postal sector in a digital world. Springer, New York
Bundesnetzagentur (2013) Beschluss in dem Verwaltungsverfahren BK5b-13/001,
Zusammenfassung von Dienstleistungen und Vorgabe von Maßgr€ oßen für die Price-Cap-
Regulierung für Briefsendungen bis 1000 Gramm ab 01.01.2014, Decision of 14 November
2013
ComReg (2014) Response to consultation and decision on price cap control for universal postal
services, 18 June 2014
Crew MA, Kleindorfer PR (2008) Regulation and the USO under entry. In: Crew MA, Kleindorfer
PR (eds) Competition and regulation in the postal and delivery sector. Edward Elgar, North-
ampton, pp 3–22
European Regulators Group for Postal Services (2014) ERGP report on tariff regulation in a
context of declining volumes. Draft report of November 2014
La Poste (2014) Evolution des tarifs du courrier et du colis au 1er janvier 2015, press release of
1 October 2014. http://legroupe.laposte.fr/Espace-Presse/Liste-des-communiques/Evolution-
des-tarifs-du-courrier-et-du-colis-au-1er-janvier-2015
Lions F (2015) Price regulation in the context of volume decline. Presentation held at the 15th
K€onigswinter postal seminar on postal economics: postal regulation and delivery markets in
transition on 11-12 February 2015 in K€ onigswinter
Littlechild S (1986) Economic regulation of privatised water authorities: a report submitted to the
Department of Environment, 26 January, London: HMSO, para 10.21 (recited from Stern J
(2014) The British utility regulation model: its recent history and future prospects. Util Policy
162–172)
Ofcom (2012) Securing the universal postal service—decision on the new regulatory framework.
Statement of 27 March 2012
Staatssecretaris van Economische Zaken (2015) Postregeling 2009, updated 13 February 2015
Stern J (2014) The British utility regulation model: its recent history and future prospects. Util
Policy 31:162–172
Swedish Postal Ordinance (2010), Postf€ orordning (2010: 1049)
Thiele S, Niederprüm A (2015) Pricing trends in the postal sector. Presentation held at the 15th
K€onigswinter postal seminar on postal economics: postal regulation and delivery markets in
transition on 11-12 February 2015 in K€ onigswinter
WIK-Consult (2013a) Main developments in postal sector (2010-2013). Study for the European
Commission, Directorate General for Internal Market and Services
WIK-Consult (2013b) Review of postal operator efficiency. Study for the British regulator Ofcom
Bundesnetzagentur (2013) Beschluss in dem Verwaltungsverfahren BKSb-13/001-
Zusammenfassung von Dienstleistungen und Vorgabe von Maßgr€ oßen für die Price-Cap-
Regulierung für Briefsendungen bis 1000 Gramm ab 01.01.2014
Origins of “Universal Service”
Chris Rowsell
1 Introduction and Motivation
“I think that the House is well aware what a universal service obligation means.”
This was Edward Leigh, a minister in the Department of Trade and Industry and
one of the leading proponents of the privatization of the Post Office1 in
John Major’s government, responding in 1993 to a Parliamentary question about
whether or not the second daily delivery was part of the universal service obliga-
tion. His answer—consistent with the position taken in the following year’s Green
Paper on the Future of the Post Office (1994, p. 21)—was that the second daily
delivery was not part of the universal service, as it was not provided throughout
the UK.
Following a further 8 years of continued second daily deliveries, Robert Camp-
bell could set out a view that would have been agreed by many:
The universal service [in the UK] has always been clear and well-articulated. . . . These
expectations are deeply ingrained in British society, where a system of twice-daily deliv-
ery—including early first-class delivery—is assumed to be a normal condition of life (2002,
p. 350).
The views expressed here are the author’s alone and do not necessarily represent those of Ofcom.
Traditions which appear or claim to be old are often quite recent in origin and sometimes invented.
Eric Hobsbawm (1983)
1
When referring to the UK’s incumbent postal operator, this chapter uses the contemporaneous
name of the group organization. Therefore, prior to 2001, it is the Post Office; from 2001 to 2002, it
is Consignia; after that Royal Mail.
C. Rowsell
Ofcom, London, UK
e-mail: chris.rowsell@ofcom.org.uk

30 C. Rowsell
With Consignia’s announcement of pretax losses of £1.1 billion in June 2002,

things changed quickly. In a statement to Parliament, Patricia Hewitt, Secretary of
State for Trade and Industry, supported Consignia’s chairman Allan Leighton’s plan
to save £350 million a year by moving to a single daily delivery, as the second daily
delivery accounted for 4 % of mail but 20 % of delivery costs. However, in her
statement she was clear that this was a commercial and operational matter for
management’s “stemming the losses and creating an efficient company” and “not
the result of any decision by the regulator [the recently established Postcomm]” or
by implication the Government (2002). Following trials in 2003, Royal Mail ceased
the second daily delivery during the financial year 2004–2005.
Recent debates regarding the definition of the universal service in 2012, Royal
Mail’s privatization in 2013 and the effects of end-to-end competition on the
provision of the universal service have raised similar concerns. In “Stopping first-
class post would prove we’ve become a second-class nation,” Dominic Sandbrook
referenced the establishment of the Royal Mail in the 1630s and experiences of
Edwardian tourists to argue in favor of the retention of the First Class post (2012).
Interestingly, it is worth noting that the introduction of the two-tier postal service
provoked many of the same arguments and pleas to antiquity in 1969. While
Hobsbawm (1983) was referring to the creation in the nineteenth and twentieth
centuries of supposedly ancient royal pageantry, his point holds for the discussion
of postal services, particularly in the UK.
These examples illustrate some of the key concerns of this chapter. Stakeholders
can agree that the universal service is clearly defined, but then hold diametrically
opposed views on what it comprises. There is often a conflation of the services
provided by a Universal Service Provider (USP) and the services that comprise the
universal service. It is also often unclear whether the USP’s services (both those
that are universal services and those that are not) are provided as a result of current
or historical social, commercial, or operational objectives. Indeed, at times it
appears that as increased codification instills the idea of universal service as a
consumer right, the potential for contention has increased.
This chapter is not concerned with the general economic definition of universal
service. Crew and Kleindorfer’s definition of a universal service obligation as the
“provision of a ubiquitous service at a uniform price and quality” (2000, p. 5) is clear
as a principle, but provides insufficient detail for its detailed codification. This chapter
also does not address the rationales for universal service, which are dealt with in detail
elsewhere (in particular, Crew and Kleindorfer 2000; Cremer et al. 2008). Rather this
chapter is interested in using the UK experience to study the development of the
concept of universal service (in particular among lawmakers); its increased codifica-
tion in legislation and regulation at the national and international level; and to consider
what the implications may arise from these moves.
Section 2 analyzes the development of the phrase “universal service” in UK
Parliamentary business up to 2004 (the latest year available in the searchable
database). Section 3 describes and discusses the relevant international postal legis-
lation, in particular that of the Universal Postal Union (UPU) and the European
Union’s Postal Services Directives (PSDs). Section 4 describes and discusses the
Origins of “Universal Service” 31
development of postal legislation in the UK. Section 5 sets out some concluding
thoughts on how universal service obligations have been codified and implications
for regulation and law-making in the context of structural letter volume decline and
increasing competition.
2 Developing the Language of Universal Service
Given the importance of the concept of the universal service to so many of the
current debates in postal issues, it is interesting how recent a development it
actually is. In Daunton’s Royal Mail: The Post Office since 1840, the phrase
“universal service” does not merit an index entry (1985). In Campbell-Smith’s
880-page Masters of the Post: The Authorised History of the Royal Mail, the
universal service is only mentioned 12 times, and in 10 of those cases it relates to
the period following 2000 (2011). A similar point was made in relation to USPS and
the universal service in the United States by John (2008).
To understand the development of the concept of “universal service,” I have
used the searchable archive of the British Parliament, which covers Parliamentary
debates, written answers, written statements, Lords reports, and Grand Committee
reports for the years 1803–2005. During this period, the phrase “universal service”
is used in three senses. First, up to 1944 “universal service” is used solely to refer to
a form of military conscription involving all eligible citizens, as opposed to
selective military conscription. This usage is not to be found in the Parliamentary
record later than the 1960s. Second, it is used in a literal sense to describe a service
available to all. The first parliamentary usage in this sense dates from 1944 and
refers to the planned creation of the National Health Service. This usage has
continued, for example, in relation to public libraries (1964), family planning
(1967), legal aid (1970), and the Connexions career service (2000). Third, it is
used in the sense relevant to this chapter to describe the “provision of a ubiquitous
public service at a uniform price and quality” (2000).
The first occurrence of “universal service” in the third sense is in 1952 during a
debate on transport in Scotland. Sir David Robertson MP stated “all the great public
services that have succeeded in the past have been built up on the same basis, from
the penny post onwards . . . it is the duty of a monopoly to give a universal service to
as many people as possible” (1952). Interestingly, with one exception, Sir David
Robertson is the only parliamentarian to use the phrase “universal service” between
1952 and 1960. He did so in debates on rail services, rural electricity supplies, and
BBC coverage but never directly with regard to postal services, although he often
used the Post Office as an exemplar.2
2
It is possibly only of coincidental interest that Sir David Robertson MP’s father was Chief
Inspector of the General Post Office in Glasgow. However, this perhaps explains his interest in
postal matters and could indicate an inherited terminology of “universal service.”
32 C. Rowsell
The chart below shows the usage of the phrase “universal service” in this third
sense since its first use in 1952–2004, the most recent full year available. The chart
counts the number of debates, reports, and written answers in which the phrase
“universal service” was used, not the number of times the phrase was used in total.
References to “universal service” in Parliamentary business, 1952-2004

60
50
40
30
20
10
0
1952
1953
1954
1955
1956
1957
1958
1959
1960
1961
1962
1963
1964
1965
1966
1967
1968
1969
1970
1971
1972
1973
1974
1975
1976
1977
1978
1979
1980
1981
1982
1983
1984
1985
1986
1987
1988
1989
1990
1991
1992
1993
1994
1995
1996
1997
1998
1999
2000
2001
2002
2003
2004
Postal services Communications Other
Source: Millbank Systems, Historic Hansard (2015)
As can be seen usage of “universal service” was infrequent before the early
1980s, where it is first used with regard to the debates leading to the Telecommu-
nications Act 1984. It only begins to be used consistently with reference to postal
services in the early 1990s. The spikes in 2000 and 2003 relate predominately to
debates and written questions on the passage of the Postal Services Act 2000 and
the Communications Act 2003, respectively. The spike in 2002 relates to significant
quality failures of Consignia and the ceasing of the second daily delivery. The
phrases “universal postal service,” “universal service obligation,” and “USO” all
follow a similar pattern with the first Parliamentary usage in 1983, 1981, and 1999,
respectively; and growing in usage throughout the 1990s and 2000s.
While the last decade’s Hansard reports are not yet in an easily accessible and
searchable format, it is clear that the phrase “universal service” has continued to
grow in popularity, both in relation to postal services and communications. For
example, the abortive Postal Services Bill of 2008/9, the passage of the Postal
Services Act, the privatization of Royal Mail, the entry of Whistl into end-to-end
competition with Royal Mail, and the issue of a broadband USO have all created a
welter of references to universal services in Parliament, government, and the media.
3 International Postal Legislation
3.1 European Postal Legislation: Postal Services Directives

(1997, 2002, and 2008)
As part of the European Union’s Single Market agenda, in June 1992 the European
Commission presented a Green Paper on the development of a single market for
postal services (1992). The Green Paper identified the areas of concern leading to
the proposal for intervention (which can be summarized as the effects of differing
universal services in Member States on consumers, citizens, businesses, and cross-
border trade); described the postal sector at that time; identified possible solutions
(complete liberalization, complete harmonization, status quo and “equilibrium,”
i.e., phased opening of the market while strengthening universal service); and
proposed—unsurprisingly—the “equilibrium” option.
This led in turn to the first PSD (97/67/EC) in 1997, which was subsequently
amended by the second PSD (2002/39/EC) in 2002 and the third PSD (2008/6/EC)
in 2008. In 1997, the first PSD set out the minimum characteristics of the universal
postal service to be provided in each member state; set limits on what area of the
postal market could be reserved for the USP; and required the creation of indepen-
dent national regulatory authorities for postal services. Under Article 3 of the first
PSD, all member states are required to ensure a universal service with deliveries
and collections at least every working day and not less than 5 days a week for items
up to 10 kg (with the discretion to increase to 20 kg and the obligation to deliver up
to 20 kg for inbound international items from other member states). It also required
services for registered items and insured items. Article 7 of this directive harmo-
nized the reserved area across member states as less than 350 g, where the price was
less than five times public tariff for first weight step of the fastest standard category.
Article 12 required tariffs for the universal service to be affordable, geared to costs
(while allowing for a uniform tariff), transparent, and nondiscriminatory. Articles
16 and 17 required the setting and monitoring of quality of service standards on
universal postal operators.
The second and third PSDs made no substantive change to the universal postal
service. Rather the second directive progressively reduced the reserved area: from
1 January 2003 the reserved area reduced to those items weighing less than 100 g
and costing less than three times the basic tariff; from 1 January 2006 weighing less
than 50 g and costing less than two-and-a-half times the basic tariff. The third
directive required full market opening by 31 December 2010 for most member
states (and 31 December 2012 in the remainder3). In addition, the third PSD set out
the apparatus for calculating and financing the cost of the universal service.
3
The member states with derogations to 2012 for full market opening were Czech Republic,
Cyprus, Greece, Hungary, Latvia, Lithuania, Luxembourg, Malta, Poland, Romania, and Slovakia.
34 C. Rowsell
It is clear that this European postal legislation was explicitly a harmonizing and
liberalizing project. However, the universal service specification mandated by the
first PSD in 1997 reflected the common provision of postal services in the 12 EU
member states4 at the time data was collected to inform the 1992 Green Paper
(1988) according to the EU study Main developments in the Postal Sector (2013).
This minimum specification of the universal service has not been reassessed since it
was introduced in 1997.
3.2 Universal Postal Union
At approximately the same time, the EU was legislating for universal postal service
in its member states, the UPU began to take a greater interest in the provision of
universal services within its members’ territories. In 1998, the UPU’s Council of
Administration incorporated the concept of a universal postal service into the
UPU’s mission statement. The 1999 Beijing Congress introduced the following
text as Article 3 of the Universal Postal Convention Letter Post Regulations:
Universal Postal Service

1. In order to support the concept of the single postal territory of the Union, member
countries shall ensure that all users/customers enjoy the right to a universal postal
service involving the permanent provision of quality basic postal services at all points
in their territory, at affordable prices.
2. With this aim in view, member countries shall set forth, within the framework of their
national postal legislation or by other customary means, the scope of the postal services
offered, and the requirement for quality and affordable prices, taking into account both
the needs of the population and their national conditions.
3. Member countries shall ensure that the offers of postal services and quality standards
will be achieved by the operators responsible for providing the universal postal service.
4. Member countries shall ensure that the universal postal service is provided on a viable
basis, thus guaranteeing its sustainability.
The 2004 UPU Congress supplemented the preamble to the Constitution of the
UPU by adding the text:
The mission of the Union is to stimulate the lasting development of efficient and accessible
universal postal services of quality in order to facilitate communication between the
inhabitants of the world.
Unlike the EU legislation, the UPU requirements leave significant discretion to

member states to determine what would constitute a universal postal service in their
own territories.
4
Belgium, Denmark, France, Germany, Greece, Ireland, Italy, Luxembourg, the Netherlands,
Portugal, Spain, and the United Kingdom.
4 UK Postal Legislation
There have been six major pieces of UK postal legislation since the Jubilee
concessions of 1897 guaranteed for the first time delivery to every address in the
UK at the same price. These are: the Post Office Act 1908 (“the 1908 Act”), the Post
Office Act 1953 (“the 1953 Act”), the Post Office Act 1969 (“the 1969 Act”), the
British Telecommunications Act 1981 (“the 1981 Act”), the Postal Services Act
2000 (“the 2000 Act”), and the Postal Services Act 2011 (“the 2011 Act”).
4.1 Post Office Acts 1908 and 1953
Neither the 1908 Act nor the 1953 Act significantly changed the governance or
operations of the Post Office. Neither defines a universal service obligation, each
using very similar language to define the Postmaster-General’s general powers and
duties to:
establish posts and post offices as he thinks expedient, and may collect, receive, forward,
convey, and deliver in such manner as he thinks expedient, all postal packets within or to or
from the British Islands or any British possession.
This is not to suggest that there was no expectation that the Post Office had a
social role. However, it is assumed that as a government department the service
provided by the Post Office implicitly took into account the need of the country.
However, while some Postmaster Generals clearly saw a social mission for the Post
Office (in particular, Fawcett and Blackwood), a key role of the Post Office was as a
revenue stream for the Treasury.
4.2 Post Office Act 1969
In contrast, the 1969 Act introduced significant changes. The Post Office ceased to
be a department of government and became a statutory corporation. The office of
the Postmaster General was abolished and replaced by a Minister of Posts and
Telecommunications.5 In Section 9 for the first time, UK primary legislation set out
social objectives for the postal services:
1. It shall be the duty of the Post Office (consistently with any directions given to it under
the following provisions of this Part of this Act) so to exercise its powers as to meet the
social, industrial, and commercial needs of the British Islands in regard to matters that
are subserved by those powers and, in particular, to provide throughout those Islands
5
The role of Minster of Posts and Telecommunications was itself abolished in 1974 by the
Ministry of Posts and Telecommunications (Dissolution) Order 1974, with all functions trans-
ferred to the Secretary of State.
36 C. Rowsell
(save in so far as the provision thereof is, in its opinion, impracticable or not reasonably
practicable) such services for the conveyance of letters and such telephone services as
satisfy all reasonable demands for them.
2. In discharging the duty imposed on it by the foregoing subsection, the Post Office shall
have regard—
(a) To the desirability of improving and developing its operating systems
(b) To developments in the field of communications
(c) To efficiency and economy
4.3 British Telecommunications Act 1981
The 1981 Act separated the postal and telecommunications operations of the
General Post Office, creating British Telecommunications. It also provided for
the suspension by secondary legislation of the Post Office’s monopoly for letters.
This was implemented by the Postal Privilege (Suspension) Order 1981, which
suspended for 25 years Royal Mail’s exclusive privilege for letters costing £1 or
more. The social duties of the Post Office were set out in Section 59 of the 1981 Act.
These were, with the exception of amendments to remove references to telecom-
munications matters and some re-ordering the clauses, identical to those set out in
the 1969 Act.
4.4 Postal Services Act 2000
The 2000 Act again changed the corporate status of the Post Office, making it a
public limited company albeit one that was 100 % owned by the Government,
represented from 2002 by the Shareholder Executive (ShEx). For the first time in
primary legislation, the 2000 Act used the phrase “universal service” and codified a
mandatory universal service.6 Section 3(1) of the 2000 Act set out that “The
Commission [Postcomm, the regulator] shall exercise its functions in the manner
which it considers is best calculated to ensure the provision of a universal postal
service.”
6
In order to meet the implementation timetable of the first PSD, these were implemented in The
Postal Services Regulations 1999 and then largely replicated in the following year’s Postal
Services Act 2000.
Section 4 of the 2000 Act set out the universal service obligation in some detail:
(a) Except in such geographical conditions or other circumstances as the Commission
considers to be exceptional—
• At least one delivery of relevant postal packets is made every working day to the
home or premises of every individual or other person in the UK or to such
identifiable points for the delivery of relevant postal packets as the Commission
may approve
• At least one collection of relevant postal packets is made every working day from
each access point
(b) A service of conveying relevant postal packets from one place to another by post and
the incidental services of receiving, collecting, sorting, and delivering such packets
are provided at affordable prices determined in accordance with a public tariff which
is uniform throughout the UK
(c) A registered post service is provided at such prices
The 2000 Act also established a licensing regime. Unlike the telecommunica-
tions licensing regime established by the Telecommunications Act 1984 licenses
were to be issued by the regulator rather than the government. In replacing a
statutory exclusivity (albeit one that could be suspended by secondary legislation)
with a licensing regime administered by the regulator, the 2000 Act paved the way
for removal of reserved area, and effectively delegated this decision to the
regulator.7
Therefore, it was at this point that the specification of the universal postal
service in the UK ceased to be the result of an ongoing negotiation between
different parts of Government (e.g., HM Treasury and DTI) and the incumbent
postal operator.
4.5 Postal Services Act 2011
The 2011 Act separated Post Office Limited from Royal Mail Group; allowed for
the privatization of Royal Mail Group and the mutualization of Post Office Limited
(the counters business); transferred regulatory responsibility to Ofcom, the com-
munications regulator; established a general authorization regime; implemented the
third PSD; and set out in further detail the obligations relating to the universal
postal service.
7
In 2006, Postcomm removed the last reserved area in the UK postal market and issued licenses
only to postal operators delivering postal items of 350 g or less, where the price was £1 or less.
38 C. Rowsell
In particular, Section 31 of the 2011 Act sets out the services that must as a
minimum be included in the universal postal service in the UK. These minimum
requirements are:
1. At least one delivery of letters every Monday to Saturday, and at least one
delivery of other postal packets every Monday to Friday
2. At least one collection of letters every Monday to Saturday, and at least one
collection of other postal packets every Monday to Friday
3. A service of conveying postal packets from one place to another by post at
affordable, geographically uniform prices throughout the UK
4. A registered items service at affordable, geographically uniform prices through-
out the UK
5. An insured items service at affordable, geographically uniform prices through-
out the UK
6. The provision of certain free services to blind/partially sighted people
7. The free conveyance of certain legislative petitions and addresses
During the passage of the Bill through Parliament, the Government committed to
not changing the minimum requirements during the life of the Parliament, i.e., up to
May 2015.
Further section 30(1) of the 2011 Act requires Ofcom to make a universal postal
service order setting out a description of the services that Ofcom considers should
be provided in the UK as a universal postal service, and the standards with which
those services are to comply. In making a universal postal service order, “OFCOM
must carry out an assessment of the extent to which the market for the provision of
postal services in the UK is meeting the reasonable needs of the users of those
services.”
Section 33 set out the exceptions to the universal postal service (items outside
the maximum size and weight; public holidays; exceptional geographic circum-
stances; in emergencies; and individual pricing agreements). Section 34 set out the
process by which the minimum requirements reflect the reasonable needs of users
of postal services and the legislative process for amending section 31. Section 35
describes the designation of USPs and Section 36 the regulatory conditions that can
be imposed on a designated USP (DUSP). Section 37 covers the publication of
quality of service information on the provision of the universal postal service;
Sections 38 and 39 the imposition on the USP of access and regulatory accounting
obligations, respectively. Sections 44–47 set out how Ofcom should review the
costs and fairness of the burden of the universal service obligations; the options for
addressing any unfair burden (including establishing a universal service fund and
procurement of the universal service to another provider); determining contribu-
tions for any universal service fund; and reporting on the sharing mechanism.
While the 2011 Act codified the universal service in more detail than ever
before, it also established an apparatus for the review and modification of the
universal service. There are two forms of review. A review of the minimum
requirements (under Section 34) can be carried out by Ofcom at any time and
results in a report to the Secretary of State. It would then be up to the Secretary of
State to make secondary legislation amending the minimum requirements, and this
would require the assent of both Houses of Parliament. The other review is under
Section 30 and relates to the requirements of the universal postal service order, the
secondary legislation implemented by Ofcom that sets out the further detail of the
universal service. In both cases the test is similar, i.e., whether the reasonable needs
of users are being met.
Therefore, unlike the European postal legislation, the UK legislation has a
mechanism for adapting the universal service in response to changing market
conditions, social needs and user demands. However, this flexibility is still less
than that enjoyed by Government and Post Office under the pre-1969 arrangements
and is also constrained by the minimum requirements of the European postal
legislation.
5 Concluding Thoughts
There is a clear consensus among postal operators, governments, regulators, users,

and other stakeholders that there are public and social elements to the provision of
postal services. However, the articulation of these social objectives using the term
of art “universal service” is considerably more recent and the explicit codification
of a universal postal service is as recent as the late 1990s, at least in the UK.
The UK example clearly illustrates the perceived need for greater codification of
the universal service obligation as a response to declining direct government
control and increasing liberalization. Prior to 1969, the services offered by the
Post Office were determined by the negotiation of sociopolitical, commercial, and
operational concerns by the Postmaster-General, HM Treasury, and Post Office
management. As described by Daunton (1985) and Campbell-Smith (2011), this
more flexible approach allowed for the introduction, for example, of uniform
postage rates in 1839, the Parcel Post in 1883, and the two-tier postal service in
1968. There was no need for the primary legislation to specify any universal service
obligation.
It was only with the corporatization of the Post Office in the 1969 Act and the
alleged diminution of direct government control that Parliament found it necessary
to place a duty on the Post Office “to exercise its powers as to meet the social,
industrial, and commercial needs of the British Islands [and] to provide throughout
those Islands (save in so far as the provision thereof is, in its opinion, impracticable,
or not reasonably practicable) such services for the conveyance of letters and such
telephone services as satisfy all reasonable demands for them.”
This pattern was repeated in the 1981, 2000, and 2011 Acts. As these acts further
liberalized postal markets, introduced independent regulation, and gave the incum-
bent greater commercial freedom (up to and including majority privatization), the
universal service obligation became increasingly specific. This raises the possibility
that the current universal service obligations are in reality best thought of as the
codification of the results of negotiations between Government and incumbent
40 C. Rowsell
postal operators and, therefore, to some extent the historic business decisions of
postal operators.
These developments must also be seen in the context of international develop-
ments. The first PSD had to be transposed and implemented during 1999. In
requiring certain elements of the universal service to be codified in UK law, the
first PSD made it politically inevitable that the Government would find it necessary
to codify the universal service at the prevailing UK standards, rather than rely on
the previous approach of more general and flexible duties.
Therefore, a number of factors have led to an increased codification of the
universal service obligation, at the same time the structural decline in letter
volumes, growth in parcel volumes and increase in competition is placing increas-
ing strain on the delivery of universal postal services.
As a result of declining volumes, several countries have already taken steps to
reduce the costs of their universal services. New Zealand has reduced to 3 days per
week the number of delivery days required in urban areas and Canada is ending
delivery to the home and using community mailboxes even in urban areas.
This has also been a matter of significant debate in Europe. An unintended
consequence of the EU’s liberalizing and harmonizing project is that the increased
codification has reduced the flexibility of incumbent postal operators to respond to
changing market and consumer needs compared to incumbent postal operators
outside the EU.
One of the conclusions of Main developments in the Postal Sector (2013)8 was
that “Requiring Member States to guarantee levels of universal postal services that
were considered essential in 1997 risks over-investment in postal services. . . . We
believe that the definition of universal service must move away from the one-size-
fits-all-and-always-will approach reflected in the current Directive. Member States
will need greater discretion in determining the scope of the USO.”
In September 2014, the European Regulators Group for Postal Services (ERGP)
published a discussion document Implementation of the Universal Service Obliga-
tion in the postal sector in view of the market developments (2014). This document
and a subsequent workshop in November 2014 (attended by regulators, postal
operators (incumbents and new entrants), user groups, and labor associations)
explored the consequences of declining mail volumes on the continued appropri-
ateness of current universal service obligations. A common theme was the need to
amend the minimum universal service obligation specified by the Directives to
ensure that it gave sufficient flexibility to adapt to the demands of users.
The next step would be to examine to what extent the UK experience (a move
from an implicit or agreed USO to greater codification driven by lessening govern-
ment control, greater liberalization, and European legislation) is typical or excep-
tional compared to other EU member states. In particular, this would involve
studying the formal status of the universal service in a number of countries before
and after the first PSD.
8
A study carried out by WIK-Consult for the European Commission.
For the case of the UK, it would be interesting to make greater use of Government
and Royal Mail archives to examine how the social elements of the Post Office’s
activities were described, considered, negotiated, and agreed before the existence of
a codified universal service obligation. One example for further examination would
be the development of Royal Mail’s Customer Charter and Code of Practice, a
product of the Major government’s Citizens’ Charter initiative. The key question is
the degree to which such a documents is merely a factual statement of what the USP
is currently providing, or whether it is a commitment of USP and government that
the USP will continue to supply the services even if not commercially viable.
References
Campbell R (2002) Politics of postal transformation: modernizing postal systems in the electronic
and global world. McGill-Queen’s University Press, Quebec
Campbell-Smith D (2011) Masters of the post: the authorized history of the royal mail. Penguin,
London
Carlisle M (1970) Criminal cases (legal aid). HC Deb, 4 December 1970, vol 807, cc1728-44.
Hansard, London
Cremer et al (2008) The cost of universal service in the postal sector. In: Crew MA, Kleindorfer PR
(eds) Competition and regulation in the postal and delivery sector. Elgar, Cheltenham
Crew MA, Kleindorfer PR (2000) Liberalization and the universal service obligation in postal
service. In: Crew MA, Kleindorfer PR (eds) Current directions in postal reform. Kluwer,
Boston
Daunton MJ (1985) Royal mail: the post office since 1840. Athlone, London
Department of Trade and Industry (1994) The future of postal services: a consultative document.
HMSO, London
Dieke A et al (2013) Main developments in the Postal Sector (2010-2013). WIK Consult
ERGP (2014) Discussion paper on the implementation of universal service in the postal sector and
the effects of recent changes in some countries on the scope of the USO
European Commission (1992) Green paper on the development of the single market for postal
services
Hewitt P (2002) Statement to the house on consignia. HC Deb, 13 June 2002, vol 386, cc1016-32.
Hansard, London
Hobsbawm E (1983) Introduction: inventing traditions. In: Hobsbawm E, Ranger T (eds) The
invention of tradition. Cambridge University Press, Cambridge
Irving S (1964) Public libraries and museums bill. HC Deb, 5 February 1964, vol 688, cc1170-287.
Hansard, London
John RR (2008) History of universal service and the postal monopoly. In Study of universal postal
service and the post monopoly, George Mason University
Millbank Systems (2015) Historic Hansard 1803-2005. http://hansard.millbanksystems.com.
Accessed 16 July 2015
Pike M (1967) National Health Service (Family Planning) Bill, 17 February 1967, vol 741, cc935-
1020. Hansard, London
Robertson D (1952) Transport, Scotland. HC Deb, 28 July 1952, vol 504, cc1109-217. Hansard,
London
Sandbrook D (2012) Stopping first-class post would prove we’ve become a second-class nation.
In: Daily Mail, 18 October 2012, London
Wicks M (2000) ConneXions Service. HC Deb, 10 April 2000, vol 348, cc43-4W. Hansard,
London
Implications of Recent Developments
in e-Commerce for Universal Service
Providers and the USO
John Hearn
1 Introduction
The purpose of this chapter is to identify how Universal Service Providers (USPs)
have been impacted by recent developments in e-commerce and to assess the
implications for the universal service they are currently obligated to provide
(the USO). Section 2 analyses how e-commerce has evolved over the last few
years. The delivery conundrum and the problems of returns are examined in Sect. 3.
Section 4 examines how selected USPs have responded to these issues and the
implications for the USO. Conclusions are set out in Sect. 5.
2 Evolution of e-Commerce
At the outset, it must be recognized that e-commerce can mean many different
things to different people. At its simplest, an email from a manufacturer asking a
supplier to send a specific quantity of raw materials might be considered to be part
of e-commerce; or booking a flight on airline’s website and receiving an email
confirming the booking and giving instructions as to how to check in. Indeed, it is
developments such as these that has led to the phenomenon labeled e-substitution
which, together with changes in the level of economic activity, has led to the
significant decline in letter post items in recent years—UPU (2013) reported
average annual volume declines of 3.8 % (domestic letters) and 5.3 % (international
letters) from 2010 to 2012; for 2013 the declines were 2.8 % and 5.6 %,
respectively.
J. Hearn (*)
Formerly vice-chair CERP and Project Manager Postal Regulation, ComReg,
Malahide, Ireland
e-mail: John@john-hearn.eu

44 J. Hearn
The particular focus of e-commerce that presents significant opportunities for

USPs is distance selling, or more specifically the retail sale of goods requiring
physical delivery to the end-purchaser. Hearn (2013) observed that few major
European retailers focused exclusively on a single distance selling business
model. Most major European retailers had adopted a multichannel approach with
online, mobile, and in-store ordering and a choice between delivery to the home and
in-store collection. Similarly, consumers were unlikely to rely exclusively on one
type of purchasing channel.
Two years on e-commerce has evolved, with further integration between the
in-store and online channels leading to what the leading edge retailers call the
omnichannel approach. Siloed channels—online, mobile or in-store—for the con-
sumer to make a decision to purchase are no longer good enough.
The John Lewis Partnership describes its approach in How We Shop, Live &
Look (2014):
The ability to mix and match purchase channels and delivery options, is now the norm as it
offers more convenience for customers. . . .
But how people buy in shops is changing, and more of the online world can be seen in
physical stores. . . .
A good example of the evolution of the shop is our York department store, which is our
most omni-channel offering to date. Its ‘See, Click, Wear’ initiative features an iPad bar,
allowing people to browse the full assortment of The Fashion Edit and order pieces for next
day delivery if an item is not on the shopfloor.
Pimkie, the French fashion retailer, also recently announced the launch of a new
store concept:
L’enseigne de vêtements Pimkie a lancé un nouveau concept de magasin a Paris, baptisé
Fashion Factory. Cette nouvelle formule mise sur une stratégie omnichannel, avec
notamment des tablettes en magasin permettant de faire son shopping online.1
According to PwC (2014)2, today’s consumers view multichannel shopping as a

given and demand greater integration.
“Convenient physical stores, a website capable of handling purchases, a mobile site or
app—these capabilities are simply the price of admission for a healthy relationship with a
consumer”. . . . [But] “it’s a high bar our survey participants have set for retailers:
compelling in-store technology, an “always-on” 24/7 service mentality, real-time insight
into product availability at individual stores, and consistent prices and offerings across a
retailer’s assets.
For example, the consumer may use the Internet at home to find out what is
available and who is selling it. Then, they may go to a retailer’s store to see and
1
Press Release April 21, 2015 Translation “Pimkie launched a new concept store in Paris, called
Fashion Factory. This new formula relies on a omnichannel strategy, especially instore tablets to
make shopping online.”
2
Each year PwC conduct a survey of online shoppers. The 2013 survey, the seventh in the series,
covered more than 15,000 online shoppers in 15 different territories on 5 continents. PwC (2014)
reports on the results of this survey.
Implications of Recent Developments in e-Commerce for USPs and the USO 45
Fig. 1 John Lewis Partnership Online Trends. Source: Internal John Lewis Partnership data
Table 1 Retailers used No. of retailers 2012 2013

in the last 12 months
Five or fewer retailers 43 % 58 %
Source: PwC (2014)
feel the actual item. While in the store they may use their smartphone to ask a
friend their opinion or to find out which retailer is offering the cheapest price.
They may then go home to order it online, but even if they decide to buy in store
the sales assistant may use a Tablet to place the order and check for stock, or if
the item or size is not available in-store, delivery to the customer’s home may be
offered.
The multidimensional nature of the purchasing transaction can best be observed
by examining data published by the John Lewis Partnership. In its annual report,
JLP (2014), it observes that “nearly two thirds of customers use both in-store and
online channels when shopping with John Lewis.” However in its annual retail
report How We Shop, Live & Look (2014), information is presented on the actual
choices made by the customers. This data shows that even though around 80 % of
customers use the online sales channel only 26 % of sales were made online—see
Fig. 1.
There are two other challenges that consumers are setting for successful
retailers. The first is that they will only use retailers they trust. The second is that
online shoppers are increasingly shopping with fewer retailers—see Table 1.
It’s not surprising that customers want to deal only with retailers they trust, given
the number of scams. Hearn (2012) reported the case of an Italian lady who ordered
electronic equipment over the Internet, paid significant sums of money to a ficti-
tious Irish “supplier,” expecting a fictitious British “postal undertaking” to deliver
the goods.
46 J. Hearn
What this means is that it will be difficult for SMEs based in one country to make
sales to customers in another that doesn’t know them, although the emergence of
“marketplaces”3 will help. Flash Eurobarometer 358 (2013) confirms the trend
noted in Hearn (2013) that EU consumers are significantly more likely to purchase
online from local sellers and providers (47 %) than from those in other EU countries
(15 %).4
JLP (2014) confirms that it is the consumer who is driving all these
developments:
Changes in consumer shopping habits, supported by rapid technological advances, have
had a significant impact on the UK retail market and altered the competitive landscape
forever. Consumers are now able to research products and reviews, compare prices and
order their goods 24/7 on-line from home or ‘on the move’ from mobile devices. There is
also demand for more flexible delivery and collection services which fit in with consumer
preferences.
Flash Eurobarometer 359 (2013) suggests that 71 % of retailers in the EU do not

sell products or services to customers in other EU countries. It identified ten main
barriers to cross-border sales development, including additional costs of compli-
ance with different consumer protection rules and contract law; potentially higher
costs of the risk of fraud and nonpayment; additional costs of compliance with
different national tax regulations; higher costs of cross-border delivery and poten-
tially higher costs in resolving cross-border complaints or disputes.
There are two ways of overcoming these barriers. The European Commission’s
approach appears to be to intervene in the markets to make it easier for cross-border
e-commerce. The Commission’s Digital Single Market Strategy for Europe
DSM-SWD (2015) makes the case as follows:
For EU businesses to benefit from the expanded market provided by cross-border e-com-
merce, the demand side must be fully unleashed. Consumer expenditure accounts for 57 %
of the EU’s GDP. However, there is consistent evidence to show that consumers have less
trust in cross-border e-commerce than in shopping online nationally and are more inclined
to buy from domestic sellers than cross-border. Whilst this “home bias” can be attributed
to some extent to cultural and proximity factors (language, brand recognition, etc.), a
strong and harmonised regulatory framework and consistent enforcement of consumer
rights across the EU will increase their willingness to engage with suppliers from other
Member States.
The alternative approach, which is used by many e-retailers trading internation-

ally, is to create a virtual presence in each targeted country so that the customer’s
perception is that he is purchasing locally rather than internationally. There are a
number of reasons why, all things being equal, retailers will opt to have country-
specific websites, even if they do not have a physical presence in the country.
The most obvious is that it is possible to address the potential customer in the local
3
Intermediate platforms between retailers and consumers, e.g., eBay.
4
The exceptions are in relatively small countries that share a common language and have other
associations with a larger and usually neighboring country: Luxembourg (41 % cross-border
vs. 14 % domestic), Malta (42 % vs. 11 %), Cyprus (31 % vs. 5 %).
language. Secondly, prices can be quoted in the national currency, notwithstanding

the use of the euro in many European countries. Thirdly, prices quoted can take
account of local VAT rates and other sales taxes. Within the EU retailers must
charge VAT at the rate applicable in the customers country of residence, which can
vary from 17 to 27 % (standard rate) or from 5 to 15 % (reduced rate).5 Also it is
possible to supply goods customized to local specifications, such as electrical
connections. The statistics of cross-border e-commerce might well reflect these
perceptions rather than the underlying reality.
The larger retailers the subject of the case studies in Hearn (2013) all have an
international presence, although there is a greater diversity in business models
compared with domestic markets. A significant majority have a bricks-and-mortar
presence in many countries. While some are owned by the parent company, in other
instances the international branches may be operated by a franchisee. In either case,
the bricks-and-mortar stores are complemented by country-specific websites. For
the retailer, the virtual presence approach demands a certain level of business in the
targeted country to justify the necessary investments. It should also be recognized
that relationships between retailers and customers are increasingly being personal-
ized based not only on location, determined for instance by the IP address used by
the consumer or by geo-position markers, but on records of previous transactions
and other records of that customers’ preferences.
A more recent strategy, reflecting the emergence of the omnichannel model, is
the establishment of a few flagship stores backed up by country-specific websites.
Marks and Spencer (2014) states its strategy as:
We will expand in Western Europe with a ‘bricks and clicks’ approach by opening flagship
stores, supported by Food stores and an online offer.
What is interesting is that a country-specific approach is adopted by the com-

pany. In France and the Netherlands, there are country-specific websites backed up
by flagship stores. In Germany, there is a virtual presence only. The “small print”
reveals that the potential customer is contracting not with the parent company or
with a German subsidiary but with Marks and Spencer Ireland Limited, although
the website also provides the address of its logistics agent in Kelsterbach for returns
and the address of its International Customer Services department in Bristol,
England. On the other hand in Greece, there is an extensive network of 29 stores,
a country-specific website but no online sales.
Even if there is complete transparency as to whom the purchaser is dealing with
there is always the possibility that the goods will be made up for dispatch directly
by the manufacturer in perhaps an Asian country and brought to Europe for
injection into domestic delivery systems. Which raises the question: Do statistics
about international e-commerce tell the whole story?
5
Sales to countries outside the EU are not subject to VAT but local taxes will be collected by the
government of the customer’s country of residence.
48 J. Hearn
3 The Delivery Conundrum and the Problems of Returns
According to an article in the UK’s Daily Telegraph (2013) newspaper, 60 % of

e-customers had problems with delivery and “Customers were bedevilled by the
dreaded ‘Sorry, you were out’ cards, missed deliveries and crashed websites.” Two
broad approaches to resolving these problems are observed.
The first approach is to make the home delivery model more effective. This
includes offering weekend and evening deliveries, nominated day deliveries or
allowing the recipient to specify time windows of 1 or 2 h. The use of mobile
phone technology to give recipients an indication of planned deliveries or to
nominate alternative delivery points is also an important development.
The alternative approach is to offer a facility to collect the parcel from a
convenient location, accessible either 24/7 or 7–11. There are three types of
collection points: (a) shops owned by the retailers (Click and Collect); (b) a network
of parcel lockers accessible 24/7; and (c) a network of independent CTN shops,6
petrol stations, etc. These are attractive options for the retailers as they tend to
generate impulse sales when the order is collected (or returned).
A conundrum is whether the decision about which options to offer is being made
by the retailer based on market research of the end consumers’ needs and a cost
appraisal or whether it is based on what the delivery companies’ offer. Organizing
deliveries to consumers’ homes is not currently a core competence of most retailers.
The evidence is contradictory. On the one hand, the European Commission7
reports that delivery to the home/work address is the norm:
those respondents who last purchased a tangible good online reported most often that it
was delivered to their home/work address (83 %). Other delivery options chosen included
picking it up in person from a shop (8 %), from a collection point/safe box in a public
location (5 %) or from a local post office (4 %).
UPS (2015) reports a lower figure for delivery to the home or work address
(74 %) but puts a different emphasis on the trends:
This year’s survey showed a decline in the preference of delivery parcels to the home, and
an increase in other collection locations. That’s one more indication that today’s flex
shopper is seeking convenience that may change based on personal circumstances and
location.
On the other hand, retailers see collection from store or other convenient
location as the preferred location. The Daily Telegraph (2013) observed that
It is becoming clear that “click and collect” holds the secret to the future of the high street.
The concept is very simple and very counter-intuitive. You, the customer, buy something
online. Then, rather than wait for the postman to ring the doorbell three days later, you go
to the shop and collect it yourself. Delivery charges are generally cheaper than by post, or
even waived.
6
Confectionery, Tobacco, Newsagents (UK, Ireland), Tabac (France).
7
DSM-SWD (2015).
...
“The reason click and collect took off over Christmas was because it solves the age-old
problem of the ‘final mile’ of delivery. People just don’t like waiting in for deliveries, but
most are happy to pick up from their local high street.”
Debenhams (2014) echoes this opinion:

Christmas 2013 was the first truly multi-channel Christmas. Convenience had a much
greater influence on customer behaviour and the winners were retailers who could offer
services such as next day click and collect. We were not able to do this and, whilst our
online sales continued to grow, in reality we fell further behind some of our competitors.
JLP (2014) gives some evidence of the importance of the click and collect offer
and innovation in this area:
Click & collect has grown strongly, up 57 %, . . .. Collect+8 was successfully launched in
September.
We are exploring more ways to integrate our channels and are trialling drive-through
collections in five shops. We began a pilot for [temperature controlled] collection lockers in
July 2013.
N Brown Group plc is a traditional mail order business that has embraced the
multichannel business model. Its Strategy (2015) acknowledges that “speed of
delivery is an area of focus and we continually improve our standard and next
day delivery times. In addition we are enhancing the range of delivery options
available to fit with the needs of our customers.”
Marks and Spencer (2014) also emphasizes the need for flexible delivery
options:
Our Shop Your Way service gives our customers flexibility, whether they are ordering
products at home for collection in-store or using our Browse and Order in-store screens for
home delivery.
According to the Daily Telegraph (2014):

Britain is already the biggest global user of the ‘click and collect’ service, which allows
shoppers to buy online and collect from their purchase from a local depot, with 35pc of
online buyers choosing the option.
It is estimated that 82m deliveries of that type will be made this year, a 17pc increase on
last year. The value of the market is also expected to soar by 112pc from £3bn last year to
£6.5bn in 2018.
It should be pointed out that even if the retailer does not offer an option to have
the items delivered to a collection point the customer can make arrangements with
the operator of a parcel locker network or parcel shops. In the case of the parcel
locker service provided by the Irish operator Nightline under the name Parcel
Motel, there is an interesting feature whereby the customer can provide a virtual
“UK delivery address,” which allows customers in Ireland to pay a domestic price
to British retailers overcoming the deterrent of comparatively high rates for
8
A network of over 5500 local neighborhood shops.
50 J. Hearn
shipping direct from the UK, and the controversial price differentials applicable to
Irish e-shoppers.
Another key issue concerning delivery is the demands retailers place on carrier
capacity. Carriers are used to managing the peaks and troughs of traffic due to the
business cycle. Postal operators are experienced in managing the Christmas pres-
sure from the quantities of Christmas and New Year cards. Retailers themselves
deal with the peaks by extending the opening hours of their shops, recruiting
seasonal staff, increasing waiting times at the caisse, etc. However, American
import Black Friday has given a new shape to the build-up of the Christmas sales
period since 2013. In 2014, the Black Friday surge in online sales caused unprec-
edented chaos to parcel deliveries in Britain. According to the Daily Mail (2014):
Couriers delivering goods for major retailers including Marks and Spencer, John Lewis
and Waterstones were last night said to be in chaos as they struggled to deal with huge
backlogs. They blamed an unprecedented surge in online shopping over the past fortnight.
...
Meanwhile, delivery firms were accused of ‘appalling’ customer service as shoppers
complained of two-week delays, lost goods and parcels left behind dustbins, under bushes
or even with strangers by over-worked couriers.
Closely interlinked with the problems of managing spikes in demand is the

problem of securing adequate compensation for delivering parcels. Major retailers
are renowned for demanding keen prices from their suppliers. On Christmas Day
2014, the BBC broadcast the following information9:
Parcel delivery company City Link, which employs 2,727 people, has gone into
administration.
The Coventry-based company, owned by investment firm Better Capital, called in
administrators on Christmas Eve after “substantial losses”.
. . ..
. . ., of Ernst and Young, said: “City Link Limited has incurred substantial losses over
several years.
”These losses reflect a combination of intense competition in the sector, changing
customer and parcel recipient preferences, and difficulties for the company in reducing
its cost base.
The interlinked issues of managing demand and sustainable pricing are key
issues that need to be tackled by postal companies and their commercial
competitors.
On the other hand, the European Commission10 reports that “for companies that
currently do not sell online but are trying to do so, 62 % say the fact that delivery
costs are too high is a problem.” The danger is that cut throat competition will lead
to unsustainable pricing, carrier failures and bring both the delivery industry and
e-commerce into disrepute. However, if the retailers have judged the views of their
customers correctly, rather than dealing with all the complexities of home delivery
there is a clear preference by many customers to specify the place and time of
9
See http://www.bbc.com/news/business-30602326. Accessed 7 May 2015.
10
COM (2015) 192 final.
collection at the outset and this may enable delivery costs and prices to be well
controlled. Innovation and customer focus stimulated by competition in the market
has brought significant benefits to users, retailers, and consumers.
A key issue for the cross-border retailer is to decide how to manage deliveries.
There are four options open to the retailer. The first is to contract with a single
operator capable of delivering to all the target countries. There are at least six such
operators11 capable of serving most addresses within the EU as well as several
regional players. The second is to use an operator to convey the parcels in bulk to
the country of destination and then to contract with a local operator for delivery of
the individual parcels. A third option is to use a broker or specialist logistics
company12 to make the necessary arrangements, subcontracting to the most appro-
priate delivery company in each target country. Finally, it is possible to use the
postal services. However, while the price can be relatively low, the involvement of
at least two companies in delivering each parcel renders this option inherently more
unreliable compared to the other options.
The flipside of the delivery conundrum is the problem of returning unwanted and
damaged goods. Features such as free returns and extended periods to return
unwanted goods in excess of the requirements of the EU Consumer Rights Direc-
tive (2011) were noted in Hearn (2013), and there has been no significant change in
the terms of sale since then. According to UPS (2015), however
. . . just over half of shoppers are satisfied with the ease of making returns and with the
clarity of the policies. . . .
The majority of shoppers say that free shipping on returns created a positive returns
experience and four in ten appreciated a hassle-free returns policy. They are also aware of
the length of time it takes to refund their goods, so “quick and easy” sums up post-purchase
expectations. . . .
Paying for return shipping and not receiving a refund quickly are the top negative
concerns when returning products.
The innovation in delivery options has been mirrored by new choices for
customers to enable them to return unwanted or damaged goods, and often free of
charge. It’s no longer the case that the only option is to queue up at the post office,
paying normal single-piece tariffs to return the damaged or unwanted goods. The
leading retailers now offer the choice of returning or exchanging damaged or
unwanted goods at a local store, leaving them with the parcel locker/parcel shop
from where they were collected or even arranging for the delivery company to
collect them. According to PwC (2015), the Russian fashion retailer Lamoda
delivers to customers’ homes, then allows them 15 minutes to try on their choices. Shoppers
pay only for what they want to keep, and the rest is promptly taken back to the store or
warehouse.
It is clear that there is also a financial imperative for retailers to minimize the
number of returns. Two of the main reasons for returns are that the goods were
11
In alphabetical order, DHL, DPD, FedEx, GLS, TNT, UPS.
12
For example, Aramex, MetaPack, Asendia (a joint venture between La Poste and Swiss Post).
52 J. Hearn
damaged during transit or in the case of clothing that they do not fit. N Brown
(2015) recognizes the need to tackle the issue of clothing sizes:
Over the past year we have made significant improvements to our product quality, further
improving our fit and size consistency and improving fabric quality. We are also developing
unique and innovative products, which build upon our considerable skills and expertise.
And there is also a need for retailer and carrier to work together to minimize
damage in transit.
4 The USP Response and the Implications for the USO
The postal market has changed forever. No longer are there state monopolies
offering a limited range of products and services on a non-negotiable basis but
even where the USPs remain in state ownership they must compete with other
operators, particularly for the delivery of goods, and with other media for transac-
tional and advertising communications.
The response of European USPs to the threats and opportunities of these market
developments needs to be considered from two perspectives. The first concerns the
structural and investment changes made. The second concerns the new products and
service options offered by the USPs.
An example of the structural changes made is the announcement by La Poste
(France) on 4 April 2014:
La Poste will be organised around a strong group and 5 business units. The Group will
ensure strategic and financial management of its business units based on the creation of
strategic committees for each one of them. These committees will be open to outside
persons.
Each business unit is responsible for a number of the main priorities in the strategic
plan:
• Mail-Parcels-Home Services: the development of the postman’s new services
• La Banque Postale: continued commercial development, particularly on the profes-
sional market
• GeoPost: expansion of a hybrid BtoB and BtoC express network in Europe
• The Network La Poste: development of services and new forms of postal presence
• Digital: the Group’s digital transformation, with the creation of a separate digital
business unit, bringing together all digital activities . . .
Three of the largest USPs, Deutsche Post, Royal Mail (UK), and La Poste
(France), all own European-wide parcel delivery networks, DHL, GLS, and
Geopost/DPD, respectively, and they continue to invest in the development of
these networks. Investments are also being made, often in joint ventures with
other USPs or with commercial companies, to take full advantage of the evolution
of e-commerce. Examples include the announcement on 28 January 2014 of a joint
venture to create and operate a network of secure automated lockers for the delivery
and return of parcels in France. The joint venture envisages an initial roll-out of
around 1500 lockers by 2016, and more than 3000 lockers in the long term to be
installed by Packcity France, a company jointly owned by Neopost and GeoPost

(part of La Poste). On 23 October 2013, Asendia13 announced a strategic invest-
ment in eShopWorld, an Irish company which has developed a software platform to
provide currency conversion, duty and taxes calculations, delivery solutions, cus-
toms clearance, returns management, and order payment processing.
As regards the new products and service options offered by the USPs, Table 2
provides a summary overview of some of the new delivery options offered by USPs
and/or their subsidiaries in support of e-commerce.
It is clear that many USPs have risen to the challenges, while the response of
others is still evolving. In terms of profitability, many of the USPs surveyed are
consistently reporting moderate profits—see Table 3.
In the meantime, the regulatory framework in which the USPs provide services
for e-commerce participants is under review. So far, the focus has been on the
relationship between USPs and senders. The EU Green Paper (2012) set out a view
of the main issues to be tackled to achieve “an integrated parcel delivery market for
the growth of e-commerce in the EU.” This was followed up with the EU’s
Roadmap (2013) to guide and organize the way forward with three main objectives:
increased transparency and information for all actors along the e-commerce value
chain; improved availability, quality, and affordability of delivery solutions; and
enhanced complaint handling and redress mechanisms for consumers. A review
was promised after 18 months.
In May 2015, the European Commission published its Digital Single Market
Strategy for Europe.14 With regard to parcel delivery, it includes a commitment that
The Commission will launch measures in the first half of 2016 to improve price transpar-
ency and enhance regulatory oversight of parcel delivery.
The justification is that

Affordable, high-quality cross-border delivery services can build consumer trust
in cross-border online sales. Stakeholders complain about a lack of transparency,
the excessive costs of small shipments and the lack of inter-operability between the
different operators typically involved in a cross-border shipment and the resulting
lack of convenience for the final consumer.
The Consumer Rights Directive (2011) that came into force on 13 June 2014
made a number of changes to consumer rights. In particular, the risk of loss or
damage passes to the consumer only when he (or his agent) is in physical possession
of the goods and the consumer’s right of withdrawal is extended to 14 days after the
receipt of goods (and to 12 months if the consumer has not been advised of his
rights). This means that the liability of the USP is exclusively to the retailer, and
inevitably prices and service specifications will be the subject of commercial
13
A joint venture between La Poste (France) and Swiss Post.
14
COM (2015) 192 final COM (2015) 192 final Communication from the Commission to the
European Parliament, the Council, the European Economic and Social Committee and the
Committee of the Regions, A Digital Single Market Strategy for Europe, and the accompanying
Commission Staff Working Document SWD (2015) 100 final.
54
Table 2 New delivery options offered by selected USPs

Weekend Evening Option to change Pre-advice of
USP Country Postal lockers Parcel shops deliveries deliveries delivery address delivery time
Deutsche Germany 2750 in Germany SAT YES YES
Post
La Poste France JV with neopost 1500 7000 pickup points SAT YES
by 2016
Royal UK Pickup at post SAT & SUN YES YES
Mail office only
Post NL The At main railway 2600 locations
Netherlands stations
bPost Belgium 125 in Belgium 1250 parcel points SAT & SUN
PostNord Denmark 500 in Denmark YES YES
Sweden YES YES YES
Itella Finland 459 in Finland Pick up at post YES YES YES
83 in Estonia offices only
An Post Ireland After failed delivery
only
J. Hearn
Table 3 Financial results of selected USPs
USP 2012 2014
Turnover €ma Group profit Profit margin (%) Turnover €ma Group profit Profit margin (%)
Deutsche Post Germany 2014 55,512 2665 4.8 56,630 2965 5.2
La Poste France 2014 21,658 816 3.8 22,163 719 3.2
Royal Mail UK Mar 2014 9,532 211 2.2 9456 430 4.5
Post NL The Netherlands 2014 4,330 758 17.5 4240 309 7.3
bPost Belgium 2014 2,416 323 13.4 2465 480 19.5
PostNord Denmark and Sweden 2014 39,173 380 1.0 39,950 351 0.9
Itella Finland 2014 1,947 39 2.0 1859 6 0.3
An Post Ireland 2014 807.3 (17.5) 2.2 821 6 0.7
a
Except Royal Mail in GB£m PostNord in SEKm
Implications of Recent Developments in e-Commerce for USPs and the USO
55
56 J. Hearn
negotiations between these parties. It must be questioned whether the European

Commission would be justified to extend economic regulation to such a commercial
and competitive market.
On the other hand, there are currently no concrete proposals to change the
universal service obligations (USO) designed to protect the needs of receivers
(end consumers). Within the EU, the concept of a USO for posts was first mooted
in the 1993 Green Paper and was legally defined in the 1997 Postal Directive. The
legal definition has not been changed since then. However in the meantime, the
demand for letter post services has declined dramatically and, as noted above, many
USPs have been focusing instead on the opportunities presented by e-commerce,
diversifying or refocusing their activities and making appropriate investments to
secure the future.
It should be noted that the EU Postal Directive provides that the USO “shall
evolve in response to the technical, economic and social environment and to the
needs of users.”15 It seems to the author that the changes in the technical, economic,
and social environment since the Postal Directive was enacted almost 20 years ago
are so significant that member states need to review how the needs of users can best
be met going forward. For many citizens, access 24/7 to a postal locker to receive
deliveries of goods ordered may be of greater necessity than daily deliveries of
letters (especially if advertising material is one of the principal contents). The key
question is whether USPs contracts with retailers to deliver and return e-commerce
parcels can create the profits that will make them better able to support a USO more
appropriate to the needs of users going forward.
5 Conclusion
In summary, e-commerce is evolving very rapidly and the boundaries between the
various sales channels are becoming blurred. Postal and other home delivery
options are no longer the automatic preference of consumers. For many, the ability
to collect the goods from a local store or from a parcel locker accessible 24/7 is
more desirable. The use of big data to personalize and localize the offers of retailers
blurs the distinction between domestic and international markets. The postal and
delivery sector is extremely competitive. The innovation and customer focus that
follows from this has allowed users, retailers, and consumers to benefit from
innovative services and efficient pricing. The best USPs are rising to these chal-
lenges and will survive, but those that fail to respond quickly enough face an
uncertain future.
15
EU Directive 97/67/EC (the Postal Directive) Article 5(1) fifth indent.
References
Consumer Rights Directive (2011) Directive 2011/83/EU of the European Parliament and of the
Council of 25 October 2011 on consumer rights, amending Council Directive 93/13/EEC and
Directive 1999/44/EC of the European Parliament and of the Council and repealing Council
Directive 85/577/EEC and Directive 97/7/EC of the European Parliament and of the Council
Daily Mail (2014) Online gifts may not arrive in time as couriers for major retailers admit they are
swamped. http://www.dailymail.co.uk/news/article-2872149/Online-gifts-not-arrive-time-cou
rier. 20 Dec 2014
Daily Telegraph (2013) Click and collect—the new way to go shopping. http://www.telegraph.co.
uk/finance/newsbysector/retailandconsumer/9785532/Click-and-collect-the-new-way-to-go-
shopping.html. 8 Jan 2013
Daily Telegraph (2014) ‘Click and collect’ services set to sweep the high street. http://www.
telegraph.co.uk/finance/newsbysector/retailandconsumer/11021825/Click-and-Collect-ser
vices-set-to-sweep-the-high-street.html. 9 Aug 2014
Debenhams (2014) Debenhams plc Annual Report & Accounts 2014. http://media.corporate-ir.
net/media_files/IROL/19/196805/agm2014/annual_report_and_accounts.pdf
DSM-SWD (2015) Commission staff working document a digital single market strategy for
Europe—analysis and evidence SWD (2015) 100 final
EU Green Paper (2012) COM (2012) 698 final. In: Green paper an integrated parcel delivery
market for the growth of e-commerce in the EU, Brussels, 15 January 2012
EU’s Roadmap (2013) COM (2013) 886 final. In: Communication from the commission a
roadmap for completing the single market for parcel delivery build trust in delivery services
and encourage online sales, Brussels, 16 December 2013
Flash Eurobarometer 359 (2013) Flash Eurobarometer 359 retailers’ attitudes towards cross-
border trade and consumer protection fieldwork: September–October 2012 Publication: June
2013 requested by the European Commission, Directorate-General for Health and Consumers
and coordinated by Directorate-General for Communication
Flash Eurobarometer 358 (2013) Flash Eurobarometer 358 consumers’ attitudes towards cross-
border trade and consumer protection fieldwork: September 2012 Publication: June 2013
requested by the European Commission, Directorate-General for Health and Consumers and
coordinated by Directorate-General for Communication
Hearn (2012) The confluence of the postal sector with the internet economy and regulation. In:
Crew MA, Kleindorfer PR (eds) Multi-modal competition and the future of mail. Edward
Elgar, Cheltenham
Hearn (2013) Delivering the goods to households—would further regulation help or hinder? In:
Crew MA, Brennan T (eds) The role of the postal and delivery sector in a digital age. Edward
Elgar, Cheltenham
How We Shop, Live & Look (2014) The John Lewis retail report 2014. How We Shop, Live &
Look. John Lewis Communications, London
JLP (2014) John Lewis Partnership plc annual report and accounts 2014. http://www.
johnlewispartnership.co.uk/content/dam/cws/pdfs/financials/annual%20reports/JLP-annual-
report-and-accounts-2014.pdf
Marks, Spencer (2014) Marks and Spencer Group plc annual report and financial statements 2014.
http://corporate.marksandspencer.com/investors/b73df1d3e4f54f429210f115ab11e2f6
N Brown (2015) N Brown Group plc full year results announcement for the 52 weeks ended
28 February 2015. http://www.nbrown.co.uk/news-item?item¼2054832613490688
N Brown Group plc Strategy (2015) http://www.nbrown.co.uk/strategy. Accessed 29 Apr 2015
PwC (2014) Achieving total retail consumer expectations driving the next retail business model.
PwC’s annual global total retail consumer survey, PricewaterhouseCoopers International
Limited (PwCIL), February 2014
58 J. Hearn
PwC (2015) Total retail 2015: retailers and the age of disruption. PwC’s annual global total retail
consumer survey, PricewaterhouseCoopers LLP, a Delaware limited liability partnership,
February 2015
UPS (2015) UPS Pulse of the Online Shopper™. United Parcel Service of America, Inc. www.ups.
com/media/en/2014-UPS-Pulse-of-the-Online-Shopper.pdf. Accessed Mar 2015
UPU (2013) Development of postal services in 2013. Universal Postal Union, Berne. http://www.
upu.int/en/resources/postal-statistics/2013-results.html
Which Universal Service Obligation
Attributes Do Americans Value?
Jennifer Bradley, Michael D. Bradley, and Jeff Colvin
1 Introduction
The U.S. Postal Service’s universal service obligation (USO) is a collection of

requirements that ensures all users receive a certain level of service at a reasonable
price. A USO can be fulfilled by one or more providers, and it is often a regulator’s
responsibility to ensure that it is being met. Today, the United States Postal Service
(USPS) has the sole responsibility for providing required postal services in the United
States; however, the current law does not contain a comprehensive and clear definition
of the USO. Instead, USPS’s USO is understood to be made up of various legal
requirements and regulations that, in most instances, provide only broad guidance.1 As
the changing communication market offers new digital alternatives to traditional hard
copy mail, the question arises as to the services desired by the American public.
To help answer that question, the authors worked closely with Gallup, an expert
in survey design and implementation, to develop a quantitative study to measure the
value of certain attributes of the USO. While similar surveys have been conducted in
other countries, this is the first of its kind for postal services in the United States.
While there have been several studies in the United States designed to elicit
postal customers’ needs, to date these surveys have been qualitative in nature. They
The views expressed are solely those of the authors and should not be construed to represent the
views of the USPS Office of the Inspector General.
1
For legislation that references various aspects of the USO, see U.S. Postal Service, Report on
Universal Postal Service and The Postal Monopoly, October 2008, http://about.usps.com/univer
sal-postal-service/usps-uso-report.pdf, pp. 9–11, and Postal Regulatory Commission, Report on
Universal Postal Service and the Postal Monopoly, December 2008, http://www.prc.gov/Docs/61/
61628/USO%20Report.pdf.
J. Bradley (*) • J. Colvin
USPS Office of Inspector General, Arlington, VA, USA
e-mail: Jcolvin@uspsoig.gov
M.D. Bradley
Department of Economics, George Washington University, Washington, DC, USA
e-mail: mdbrad@gwu.edu

60 J. Bradley et al.
ask the respondents for their opinions on what postal services they value. While
these studies can offer insight into what services postal customers want, they do not
provide quantitative estimates of the value customers place on specific services,
which are a feature of the study presented here. Moreover, it is important, in
designing such a study, to present respondents with enough information to make
their choices meaningful. For example, if the survey does not indicate that respon-
dents would have to pay a higher price in order to receive the higher level of service,
respondents will have little incentive to choose a lower level of service.
The study presented in this chapter has been designed to offer respondents a
trade-off between different levels of service and price. This forces respondents to
consider if a higher level of service is worth the extra cost. In this way, the study
provides quantitative measures of the value that consumers and businesses put on
various aspects of the Postal Service’s USO. These quantitative measures can
provide policy makers with insight into which services postal customers value the
most to help them design a USO.
Our survey used Stated Preference Discrete Choice Modeling (SPDCM) meth-
odology, which involves presenting respondents with a series of choice tasks. Their
responses were then used to estimate willingness to pay (WTP) for different USO
attributes. Section 2 provides a review of previous studies. The survey is described
in Sect. 3, and Sect. 4 lays out the model. Section 5 shows the WTP results, and
Sect. 6 presents an alternative way of analyzing the results. Section 7 breaks down
the results by subpopulation and Sect. 8 is offered by way of conclusions.
2 Review of Previous Studies

The survey was developed using lessons learned from previous surveys done in
other countries.2 Learning from previous studies,3 we strived to limit the
number of attributes4 and focus on attributes that respondents explicitly
2
These include Henrik Lindhjem and Simen Pedersen, “Should Publicly Funded Postal Services
be Reduced?; Copenhagen Economics, Main Developments in the Postal Sector; Rand Europe,
Study on Appropriate Methodologies to Better Measure Consumer Preferences, 2011, http://www.
rand.org/content/dam/rand/pubs/technical_reports/2011/RAND_TR1140.pdf.
Ofcom, Universal Service Obligation, Postal User Needs 2012, Quantitative research report,
2012, http://stakeholders.ofcom.org.uk/binaries/research/post/quantitative-oct2012/report.pdf.
Ofcom, Review of Postal Users’ Needs. An Assessment of the Reasonable Needs of Users in
Relation to the Market for the Provision of Postal Services in the United Kingdom, March 2013,
http://stakeholders.ofcom.org.uk/consultations/review-of-user-needs/; NERA, The Social Value of
the Post Office Network, Report for Postcomm, August 2009, http://www.nera.com/extImage/
PUB_Postcomm_Aug2009.pdf; and Accent, Postal Universal Service Obligation and Rob Sheldon
and Alison Lawrence, “The UK Service Obligation,” pp. 199–216.
3
A detailed summary of studies in other posts can be found in Appendix B of U.S. Postal Service
Office of Inspector General, Guiding Principles for a Universal Service Obligation, Report
No. RARC-WP-15-001, November 11, 2014, https://www.uspsoig.gov/sites/default/files/docu
ment-library-files/2014/rarc-wp-15-001.pdf.
4
Rand Europe, Study on Appropriate Methodologies to Better Measure Consumer Preferences,
2011, http://www.rand.org/content/dam/rand/pubs/technical_reports/2011/RAND_TR1140.pdf
and Ofcom, Universal Service Obligation, Postal User Needs 2012, Quantitative research report,
2012, http://stakeholders.ofcom.org.uk/binaries/research/post/quantitative-oct2012/report.pdf,
Which Universal Service Obligation Attributes Do Americans Value? 61
experience.5 For example, in the survey performed in Norway in 2011, Copenhagen

Economics (2010) found that the results related to excluding bulk mail from the
USO were not usable. The authors theorize that the type of questioning used was too
confusing for a survey, since it relies on the respondents to understand the impact
of including or excluding bulk mail from the USO.6 Another takeaway from
the same survey was that the payment mechanism contained price intervals that
were too small and needed to be increased to better reflect what respondents were
willing to pay.7
3 Survey
The survey was implemented in the summer of 2014 from Gallup, an internation-
ally recognized survey company that has established protocols to ensure the
statistical validity of polling results. The sample for consumers was drawn by the
Gallup Panel, a panel that represents the population of the United States. Gallup
selected members through the use of random-digit-dialing and address-based sam-
pling. The sample was stratified by age, education, race, and ethnicity. The survey
for consumers was administered by Internet and by mail, the latter to ensure that the
survey included respondents who were less computer literate. To ensure an ade-
quate data set, Gallup sent out a sample of 2486 web surveys and 625 mail panelist.
The Internet portion of the survey resulted in 759 completed surveys (30 %
response rate) and the mail resulted in 178 completes (28 % response rate) for a
total of 937 completed consumer surveys.
The sample for businesses was drawn from a database of commercial busi-
nesses licensed by Marketing Systems Group by Dun & Bradstreet. The sample
was stratified by size, as measured by number of employees, small (0–99),
medium (100–499), and large (500 or more). Gallup used phone calls to identify
the right point of contact and then e-mailed a link of the survey. Of the 810 indi-
viduals who were identified as best able to evaluate the USPS, there were
203 completed surveys including 28 small businesses, 99 medium businesses,
and 74 large businesses.
In order to ensure that the consumer sample was representative of US popula-
tion, Gallup applied base weights to account for the probability of selection of the
and Ofcom, Review of Postal Users’ Needs. An Assessment of the Reasonable Needs of Users in
Relation to the Market for the Provision of Postal Services in the United Kingdom, March 2013,
http://stakeholders.ofcom.org.uk/consultations/review-of-user-needs/.
5
Austria and other EU countries (2010) and Italy, Poland and Sweden (2011).
6
Copenhagen Economics, Main Developments in the Postal Sector (2008-2010) Final Report,
November 2010, http://ec.europa.eu/internal_market/post/doc/studies/2010-main-developments_
en.pdf, pp. 135–150 and Appendix A.
7
Ibid.
Table 1 List of attributes and levels of the attributes included in the survey
Attribute Levels of each attribute
Delivery • Monday through Saturday
frequency • Monday through Friday
• 3 days a week
Access to post • Post office open 8 h Monday through Friday and 4 h on Saturday
office • Post office open 4 h Monday through Saturday
• Postal counter at a commercial retail store, staffed by a non-postal employee,
where customers can purchase shipping and mailing services for both letters
and parcels, open 10 h a day
• Self-service kiosk, which is a secure automated device that allows customers
to purchase shipping and mailing services for both letters and parcels, open
24 h a day, 7 days a week
Mode of • Delivery to the door (of home or business)
delivery • Delivery to the mailbox at the curb
• Delivery to a locked cluster box located no more than ¼ miles from your home
• Parcel locker, which is a secure postal facility for retrieving parcels, acces-
sible 24 h a day, 7 days a week
Price* • For consumer-letters, 50, 60, 75, and 85 cents
• For business-letters, 38, 46, 54, and 62 cents
• For consumer and business-parcels, $11, $13, $16, and $18
Source: Gallup Report
*The lowest price reflects close proximities of current prices
individual into the Gallup consumer panel. All population parameters used in
weighting were obtained from the latest current population statistics available.
Weights were trimmed to ensure they did not have an undue effect on the variability
of the survey estimates and were normalized so the average weight equals one.
There were no weights for the business sample.8
The following attributes were included in the quantitative survey: frequency of
delivery, mode of delivery, access to postal services, and price. Price was necessary
in order to monetize the value consumers place on specific attribute levels. In other
words, respondents were asked, in effect, how great an increase in the stamp
(or other postal) price they would be willing to undergo in order to get a service
improvement of one kind or another.9 The list of attributes and their alternative
levels can be seen in the following table (Table 1).
8
For a more complete discussion of how the survey was developed and weighted, please see
U.S. Postal Service Office of Inspector General, What Postal Services Do People Value the Most?
A Quantitative Survey of the Postal Service Universal Service Obligation, Report No. RARC-WP-
15-007, February 23, 2015, https://www.uspsoig.gov/sites/default/files/document-library-files/
2015/rarc-wp-15-007.pdf.
9
Respondents were not given any direction on how to think about these questions—whether it
were how the USO change affected them personally or everyone.
4 Model
A stated preference discrete choice experiment produces a data set which reflects
participants’ choices from evaluating different USO alternatives. When faced with
these alternative USO combinations, participants selected the set of USO attributes
that provided them with the most utility. Because it can be derived from a random
utility model, a mixed logit model provides a mechanism for estimating the
marginal utilities associated with the different levels of each attribute.
Specifically, a mixed logit model is consistent with a random utility model that
specifies consumers get utility from the observed attributes (the xi) and an
unobserved random component (η). Consider a set of “j” alternatives (5-day
delivery, 4-day post office access, etc.), across “t” choice sets (the different USO
combinations in the survey) for n individuals (the survey participants). One can
write the utility associated with this as:
U n jt ¼ βn xn jt þ ηn jt
Note that the beta coefficients (which are ultimately estimated in the logit model)
measure how utility changes as the value of “x” changes. In the case of linear utility,
as expressed above, the mixed logit choice probability takes the following form:
ð
eβn xni
Pni ¼ X f ðβÞdβ;
e β n xn j
j
where f(β) is a density function.10

The choice experiment data set has four distinct subsets, one of consumers with
regard to letters, one for consumers with regard to parcels, one for businesses with
regard to letters and one for businesses with regard to parcels. This means that four
different mixed logit models were estimated.
To take advantage of the richness of the data set, individual logit models were
estimated for each participant using the hierarchical Bayes method.11 This was
implemented with effects coding so that the sum of the marginal utilities over the
different levels of each attribute is zero. That is why the lowest level for each
attribute has a negative part worth utility. For example, the next table presents the
average part worth utilities from the logit model for consumers with regard to
10
For a complete explanation of the relationship between random utility models and a mixed logit
model, see, Train, Kenneth, E., Discrete Choice Methods with Simulation, 2nd Ed., Cambridge
University Press, 2008.
11
For a discussion of the Hierarchical Bayes method see, Allenby, Greg M. and Rossi, Peter E. and
McCulloch, Robert E., Hierarchical Bayes Models: A Practitioners Guide (January 2005). Avail-
able at SSRN: http://ssrn.com/abstract¼655541.
Table 2 Average part worth utilities

Average part worth Std
Attribute Level utilities error T ratio
Frequency of Mon–Sat 0.210 0.0089 23.59
delivery Mon–Fri 0.169 0.0054 31.03
3 days a week 0.379 0.0124 30.62
Access to postal Post office open 8 h Mon–Fri 0.356 0.0128 27.86
services 4 h on Sat
Post office open 4 h Mon–Sat 0.094 0.0087 10.82
Postal counter at commercial 0.118 .0113 10.43
retail store
Self-service Kiosk 0.332 0.0139 23.93
Mode of delivery Delivery to the door of your 0.245 0.0139 17.56
home
Delivery to a mailbox at the 0.256 0.013 19.7
curb
Delivery to a locked cluster 0.500 0.0167 30.02
box
Price 50 cents 0.435 0.0128 33.9
60 cents 0.237 0.0058 40.61
75 cents 0.156 0.007 22.31
85 cents 0.516 0.0118 43.91
letters. These part worth utilities can then be used to calculate choice outcomes such
as WTP and price simulations (Table 2).12
5 Willingness to Pay Results
The WTP estimates are calculated separately for each attribute and provide an
estimate of how much value the respondents place on the base level relative to
alternative levels of service.13 For example, the WTP values for the access alter-
12
For other examples of the use of the logit model to estimate part worth utilities in postal and
non-postal contexts, see Rohr, Charlene, Trinkner, Urs, Lawrence, Alison, Kim, Chong Woo,
Potoglou, Dimitris and Sheldon, Rob, “Measuring Consumer Preferences for Postal Services,”
Swiss Economics Working Paper, July 2012, Kjær, Trine, Bech, Mickael, Kronborg, Christian and
Raun Mørkbak, Morten, “Public Preferences for Establishing Nephrology Facilities in Greenland:
Estimating Willingness-to-Pay using a Discrete Choice Experiment,” European Journal of Health
Economics, 2003, Vol. 14, 739-748, or NERA, “The Social Value of the Post Office Network, A
Report for Postcomm,” manuscript, 2009.
13
For a detailed description of the methodology used to calculate these values, see Gallup’s
Technical Report in Appendix C.
Table 3 WTP results for consumer/letter survey

Attribute WTP consumers/letters (cents)
Frequency of delivery (WTP to maintain Monday–Saturday)
Monday–Friday 0.23
3 days a week 2.06
Access to post office (WTP to maintain post office open 8 h Mon–Fri and 4 h on Sat)
Post office open 4 h Mon–Sat 0.77
Postal counter at commercial retail store 1.45
Self-service kiosk 2.07
Mode of delivery (WTP to maintain delivery to the curb)
Delivery to the door 0.72
Delivery to a locked cluster box 3.37
Table 4 WTP results for business/letter survey

Attribute WTP business/letters (cents)
Frequency of delivery (WTP to maintain Monday–Friday)
Monday–Saturday 0.49
3 days a week 3.40
Mode of delivery (WTP to maintain delivery to the door)
Delivery to the curb 1.16
natives reflect the value of accessing postal services via a Post Office open normal
business hours compared to the three alternatives. The higher the WTP estimate, the
more the respondents valued the base level of service over the alternative.14 In most
cases, the base level refers to the highest level of service. However, in a few
instances, the value of the second highest level of service was so close to the first
that respondents were essentially indifferent. As a result, for consistent comparison,
Gallup substituted the second highest level of service for the first to ensure positive
WTP measures. This happened in the case of mode of delivery for letters for
consumers and frequency of delivery for both letters and parcels for businesses.
14
We do not mean to imply that the WTP measure should be used as an attempt to identify the
maximum price the Postal Service should charge for its services.
The overall results for letters for both consumers and businesses are shown in
Tables 3 and 4. The WTP values for letters are given in cents and represent the
respondent’s WTP for each letter they send. When multiplied by the number of
letters sent each year, the number becomes quite substantial. For example, when the
estimated WTP value for consumers for 6-day delivery vs. 3-day delivery of letters
is multiplied by number of First-Class letters sent by households in FY 2013 (12.3
billion), the result is $258.3 million.
While it is useful to show a back-of-the-envelope calculation to demonstrate
how the WTP estimates can result in significant values, we caution against using the
WTP estimates to calculate a total value of each attribute for the purposes of
directly comparing to the Postal Service’s cost of providing the levels of service.
This is because the amounts can be influenced by factors such as which prices were
included in the survey. Further, while WTP values cast light upon the direct private
benefits of customers, they do not necessarily include the benefits to society as a
whole. For example, as indicated by a study commissioned by the Postal Regulatory
Commission, there is a value associated with having a government presence in the
community.15 It is not possible to know precisely what respondents considered
when answering the questions. Further, the total estimates vary depending on which
postal volumes are used for the calculation. Thus, while we recommend that the
WTP estimates not be used to estimate a measure of the total benefit of the USO
attributes, the survey’s results provide an important and useful measure of the
relative value across attributes and levels of attributes.
As can be seen by Table 3, when it comes to frequency of delivery for letters,
consumers are indifferent between Monday through Saturday and Monday through
Friday delivery. However, they place a high value on 5-day vs. 3-day delivery.
For access, the WTP estimates are derived by comparing alternative schedules to
post offices being opened normal business hours. The WTP estimate for post offices
being open for limited hours was fairly low, only 0.70 cents. However, the WTP
estimate to preserve the status quo vs. a counter in a non-postal store was fairly
high, and the WTP to avoid kiosks was even higher. This indicates that consumers
seem to value going to a Post Office manned by postal employees, even if the Post
Office is only open limited hours.
The survey also indicates that consumers place a high value on receiving their
letters at the door or curb. In fact, the highest WTP estimate for consumers in the
letter survey was for delivery to the curb as opposed to a locked cluster box.16 In
contrast, the WTP value for door over curb is low. This indicates that while
15
The Urban Institute, A Framework for Considering the Social Value of Postal Services, Final
Report. Prepared for the Postal Regulatory Commission, February 2010, http://www.prc.gov/prc-
docs/library/archived/Final_Report_Sent_to_PRC_Feb_3_943.pdf.
16
As mentioned earlier, Gallup used curb as the base for the consumer-letter survey in order to
avoid having a negative WTP value.
Table 5 WTP results for consumer/parcel survey

Attribute WTP consumer/parcels ($)
Frequency of delivery (WTP to maintain Monday–Saturday)
Monday–Friday 0.31
3 days a week 1.73
Postal Counter at commercial retail store 1.11
Delivery to a parcel locker 3.16
consumers are indifferent between curb and door delivery for letters, they greatly
prefer either of these two to delivery to a cluster box.
The WTP estimates for the business-letter survey demonstrated that overall
businesses have similar needs as consumers for their letter mail, with one exception.
While consumers were indifferent between the two, businesses seem to place some
value on maintaining delivery to the door over the curb. For the business-letter
survey, businesses placed the most value on maintaining 5-days-a-week delivery
over 3 days.
The WTP estimates for parcels for consumers and businesses are shown in the
Tables 5 and 6. It is important to note that the WTP estimates for parcels are in
dollars. As with letters, the higher the WTP value, the more value respondents place
on maintaining the higher level of service compared to the alternative.
For both consumers and businesses, the highest WTP estimates for the parcel
surveys were delivery to a parcel locker over door delivery, followed by delivery to
a locked cluster box over door delivery. Consumers had a fairly low but positive
WTP value delivery to the curb over the door, and businesses had a moderately high
WTP estimate for delivery to the curb over the door. This seems to indicate that
businesses place a high value on maintaining parcel delivery to the door, and
consumers value parcel delivery to the door or curb.
For frequency of delivery, both consumers and businesses had a positive esti-
mated WTP for 6-day delivery over Monday through Friday delivery, although it
was relatively small compared to the other WTP estimates. However, both con-
sumers and businesses place a high value on maintaining Saturday delivery when
compared to 3-day delivery.
As with the letter survey, the WTP estimates for access to postal services seem to
indicate that both consumers and businesses value post offices, even if only open
limited hours, over postal counters in a not-postal store or kiosks (Tables 5 and 6).
Table 6 WTP results for business/parcel survey

Attribute WTP business/parcels ($)
Frequency of delivery (WTP to maintain Monday–Friday)
Monday–Saturday 0.50
3 days a week 2.15
Delivery to a parcel locker 2.54
Table 7 Price sensitivity for retail service

Proportion of consumers who prefer full retail
Alternative service at a 50 cent stamp price (%)
M–F 4 h a day 72.1
Non-postal counter 85.3
6 Alternative Analysis of Results
To get a better understanding of the results, we explore an alternative analysis of the

results—identifying the sensitivity of consumers and businesses to change in the
price of USO attributes. We will explain what we did by walking through the
analysis for retail access.
Our baseline analysis indicated that consumers, as they relate to letters, had a
strong preference for retail facilities staffed by Postal Service employees over
non-postal counters (in grocery or department stores) or kiosks. There was also a
modest, but still substantial, preference for full retail hours of post office access
instead of just 4 hours a day, 5 days a week. These attribute results can be
summarized by looking how many people preferred full service (Monday through
Friday for 8 h a day and Saturday for 4 h a day) over lower service alternatives at a
constant stamp price of 50 cents (Table 7).
The next question is how strong these preferences remain when the cost of full
service is increased relative to the lower service alternative. To answer that
question, we investigate what proportion of consumers would continue to prefer
full service as the stamp price is increased. The part worth utilities are used to
calculate the choice proportions. Suppose we are considering n different choices.
Then, the part worth utility for each choice is given by θi. To find the proportion of
consumers that prefer a specific choice, ρi, we apply the following formula:
eθ i
ρi ¼ X n
i¼1
eθ i
The consumer choice is thus between full retail service at a higher stamp price or
a lower service alternative at a 50 cent stamp price. If the proportion of consumers
preferring full service falls sharply as the stamp price is increased, we can conclude
that they are very price sensitive with respect to retail access. If the proportion
preferring full service stays about the same as the stamp price is increased, we
would argue that consumers are not price sensitive with regard to retail access and
thus put a very high value on full service retail access.
The first, lower service alternative we examine maintains Postal Service retail
facilities but curtails access to just 4 h a day for Monday through Friday. The
following table investigates price sensitivity by showing how that preference falls
as the stamp price is increased (Table 8).
The table indicates that consumers are very price sensitive when comparing full
service and more limited service at Postal Service facilities. Only a quarter of
consumers would prefer full service if it required a 75 cent stamp price and less than
10 % of consumers would prefer full service if it required an 85 cent stamp price. In
everyday parlance, this could be described as “consumers prefer full retail access
hours, but they are not really willing to pay for them.”
We next compare full service retail access to a postal counter at a commercial
retail store, staffed by non-postal employees, open 10 h a day (Table 9).
This table reveals that consumers are more willing to pay the higher stamp price
to retain full service access if the alternative is a postal counter. Consumers are less
price sensitive to this alternative, indicating more of a dislike for postal counters.
Table 8 Price sensitivity for full retail service vs. limited hours
Proportion of people willing to pay the higher stamp price
Price to retain full service over Monday through Friday for 4 h (%)
$0.60 56.6
$0.75 24.5
$0.85 8.3
Table 9 Price sensitivity of full service vs. postal counter in non-postal store
Price to retain full service over a counter staffed by non-postal personnel (%)
$0.60 74.5
$0.75 42.1
$0.85 16.8
Table 10 Price sensitivity of full retail service vs. kiosk

Price to retain full service over a self-service kiosk (%)
$0.60 86.1
$0.75 60.5
$0.85 30.0
Table 11 Price sensitivity analysis for mode of delivery—consumers/letters

Proportion of people willing to pay the Proportion of people willing to pay the
higher price to retain door delivery over higher price to retain door delivery over
Price curb delivery (%) cluster box delivery (%)
$0.50 47.9 94.4
(same
for both)
$0.60 31.7 89.5
$0.75 10.3 68.0
$0.85 3.1 37.2
Nearly three-quarters of consumers are willing to pay a 10 cent higher stamp price
in return for full service access instead of having to use a postal counter.
Finally, consumers most dislike the kiosk alternative and many are willing to pay
higher stamp prices to maintain full service retail access. Eighty-six percent of
consumers are willing to pay a 10 cent higher stamp price and 60 % are willing to
pay 25 cent higher stamp price to retain full retail access over using a kiosk. In this
comparison, consumers are not very price sensitive (Table 10).
We did a similar analysis for consumers and parcels, and businesses with letter
and parcels and got fairly similar results. Although businesses appear to be a little
more price sensitive when it comes to mailing parcels.
The next analysis was for mode of delivery for consumers and letters (Table 11).
As you can see by the analysis, at the same price, consumers are indifferent
between door and curb delivery, but strongly favor door over cluster box delivery
for letters. Even as price increases to 75 cents, the majority of consumers would
prefer to pay a higher stamp price and maintain door delivery than to receive their
letters via cluster box.
The results favoring door delivery over cluster box are even more pronounced
for parcels. The table below shows the price sensitivity for consumers and parcels.
Even at the highest price in the survey, $18, 57 % of consumers would prefer to pay
the higher price to maintain door delivery over cluster box delivery and 67 % would
prefer to pay the higher price and maintain door delivery over delivery to a parcel
locker (Table 12).
The analysis for businesses was similar to consumers; however, businesses had a
stronger preference for door delivery compared to curb, than consumers. At the
current price, 38 cents, 80 % of businesses would prefer delivery to the door
vs. the curb.
Table 12 Price sensitivity analysis for mode of delivery—consumers/parcels

Proportion of people Proportion of people
Proportion of people willing to pay the higher willing to pay the higher
willing to pay the higher price to retain door price to retain door
price to retain door delivery delivery over cluster box delivery over parcel
Price over curb delivery (%) delivery (%) locker delivery (%)
$11.00 73.6 98.4 98.9
(same
for
both)
$13.00 53.2 96.1 97.4
$16.00 17.1 81.9 87.4
$18.00 5.7 57.2 67.1
Table 13 Price sensitivity analysis for frequency of delivery—consumers/letters

Proportion of people willing to pay the Proportion of people willing to pay the
higher price to retain 6-day delivery over higher price to retain 6-day delivery over
Price 5-day delivery (%) 3-day delivery (%)
$.50 53.6 89.5
$0.60 36.8 81.1
$0.75 12.6 51.6
$0.85 3.9 22.9
The third analysis was for frequency of delivery. Below is the analysis for
consumers and letters. As can be seen in the table, at a stamp price of 50 cents,
consumers are indifferent between 5-day and 6-day delivery. However, the results
for 3-day delivery are a lot different. At a 50 cent stamp price, 89.5 % of consumers
prefer 6-day delivery. Even at a price of 85 cents, almost a quarter of those surveyed
would prefer to pay the higher stamp price and maintain 6-day over 3-day delivery
(Table 13).
7 Willingness to Pay by Subpopulation
One of the advantages of estimating individual part worth utilities is that it allows
examining WTP results for important subpopulations. For consumers, it is of
interest to investigate breakouts of WTP values for four characteristics: income,
age, population density, and mode of delivery. These breakouts support examina-
tion of whether the WTP is consistent across age, income, or population density
and, if not, how it varies across these population characteristics.
To find appropriate subpopulations, we examined the individual category WTP
results to identify any natural breaks, above and below which there were observable
differences in WTP values. That examination produced the following binary cutoff
values: (a) if age was less than 40, (b) if income was less than or equal to $50,000,
and (c) if population density was in the bottom 40 % of responses. In addition, there
were certain comparisons that, for the full sample, never provided a WTP result as
large as one cent, for letters or parcels. Those comparisons are: (1) 6-day delivery
vs. 5-day delivery, (2) door delivery vs. curb delivery, and (3) full service Post
Office access vs. Monday through Friday for 4 h a day. These comparisons also
never yielded a WTP estimate as much as one cent for any of the subpopulations, so
they will not be included in the results presented below.
The first set of results are for consumers as they relate to letters. All figures are in
cents. The cents measure the WTP for the higher level of the attribute. Some
interesting patterns emerge. First, the WTP for higher levels of service is consis-
tently higher for those earning $50,000 a year or less. This is consistent with lower
income households having less access to alternative electronic means of commu-
nication. The differences in WTP are particularly large for retail access with lower
income households having stronger preferences for full retail access.
We also find that households that live in lower population density areas have a
higher WTP for postal services. The difference is very large for delivery mode, with
low density households expressing a much higher level of utility for door delivery
(Table 15).
Age does not seem to matter for WTP for 6-day delivery relative to 3-day
delivery, but consumers 40 and over have a higher WTP for full retail access and
door delivery. A final subpopulation breakout is by delivery mode. Households
were classified by the type of receptacle or location in which they received their
mail and separate WTP measures were calculated. Of particular note is the fact that
Table 14 WTP values for consumers as they relate to letters by age

Number 6-day Full service retail Full service retail Door delivery
of vs. 3-day access vs. postal access vs. self- vs. CBU
Age responses delivery counter service kiosk delivery
Under 197 2.04 1.21 1.80 2.43
40
40 and 739 1.91 1.63 2.36 2.59
over
Table 15 WTP values for consumers as they relate to letters by population density
Door
Number 6-day Full service retail Full service retail delivery
Population of vs. 3-day access vs. postal access vs. self- vs. CBU
density responses delivery counter service kiosk delivery
Least 339 2.03 1.68 2.40 3.19
dense
(40 %)
Most dense 597 1.89 1.48 2.16 2.24
(60 %)
Table 16 WTP values for consumers as they relate to letters by income

Door
of vs. 3-day access vs. postal access vs. self- vs. CBU
Income responses delivery counter service kiosk delivery
Up to 245 2.09 1.81 2.73 2.75
$50,000
Over 539 1.76 1.30 1.93 2.47
$50,000
Table 17 WTP values for consumers as they relate to letters by mode of delivery
Door
Delivery of vs. 3-day access vs. postal access vs. self- vs. CBU
mode responses delivery counter service kiosk delivery
Door 220 2.32 2.10 3.14 5.26
Curb 469 1.84 1.48 2.10 2.77
CBU 155 1.98 1.19 1.77 0.52
Post 54 1.59 1.44 2.06 0.62
office
Other 38 1.83 1.66 2.49 1.00
WTP for door delivery rather than cluster box delivery, among those households
currently receiving door delivery, is very strong. Curb recipients also expressed a
strong preference for door delivery over cluster box delivery, but cluster box
recipients seemed to be relatively satisfied with their current mode. It should be
noted that the same breakdown for parcels showed that cluster box recipients had
higher WTP values than other subpopulations for maintaining door delivery over
cluster box delivery of parcels.
8 Conclusion
The results of this survey indicate that both consumers and businesses still value
postal services, especially delivery to the door and curb and post offices. However,
the respondents do not seem to place as much value on Saturday delivery, especially
for letters. There is a limit, however, to how much people are willing to pay for
higher levels of service. These values, as well as the benefits to society as a whole,
need to be weighed against the costs to the Postal Service of providing these services.
References
Accent. Postal Universal Service Obligation and Rob Sheldon and Alison Lawrence, “The UK
Service Obligation”
Copenhagen Economics (2010) Main developments in the postal sector (2008–2010) final report.
http://ec.europa.eu/internal_market/post/doc/studies/2010-main-developments_en.pdf
Lindhjem H, Pedersen S. Should publicly funded postal services be reduced?
NERA (2009) The social value of the post office network, report for postcomm. http://www.nera.
com/extImage/PUB_Postcomm_Aug2009.pdf
Ofcom (2012) Universal service obligation, postal user needs 2012. Quantitative research report.
http://stakeholders.ofcom.org.uk/binaries/research/post/quantitative-oct2012/report.pdf
Ofcom (2013) Review of postal users’ needs. An assessment of the reasonable needs of users in
relation to the market for the provision of postal services in the United Kingdom. http://
stakeholders.ofcom.org.uk/consultations/review-of-user-needs/
Rand Europe (2011) Study on appropriate methodologies to better measure consumer preferences.
http://www.rand.org/content/dam/rand/pubs/technical_reports/2011/RAND_TR1140.pdf
The Urban Institute (2010) A framework for considering the social value of postal services, final
report. Prepared for the Postal Regulatory Commission. http://www.prc.gov/prcocs/library/
archived/Final_Report_Sent_to_PRC_Feb_3_943.pdf
U.S. Postal Service (2008) Report on universal postal service and the postal monopoly. http://
about.usps.com/universal-postal-service/usps-uso-report.pdf
U.S. Postal Service Office of Inspector General (2014) Guiding principles for a universal service
obligation, report no. RARC-WP-15-001. https://www.uspsoig.gov/sites/default/files/docu
ment-library-files/2014/rarc-wp-15-001.pdf
Impact of Market Dynamics on the Net Cost
of the USO
Dariusz Nehrebecki, Leonardo Mautino, and Gavin Knott
1 Introduction
The net cost of universal service has been the subject of extensive discussion in the
literature, in terms of the methodology to be followed, and also the practical
assessment of certain elements of the universal services. For example, previous
reviews of the potential savings from moving from 6-day to 5-day delivery indi-
cated it is highly likely that there is a positive net cost of this aspect of the USO.1
However, as further presented below, these papers have generally tended to be
relatively “static” in nature. The concept of the reference scenario approach gen-
erally taken to evaluate the net cost of the USO necessarily focuses on changes in
cash flow, which is therefore an annual measurement. The approach evaluates the
cost and revenue impact of removing the USO in that year.
This is useful for evaluating whether the USO has a net cost, and therefore
whether it is causing a market distortion in a particular year. However, it is less
useful for considering forward-looking public policy development. In assessing
whether, for example, the current 5-day USO specification within the Third Postal
Directive remains relevant, it is important to understand trends in the value of the
USO to users, and therefore trends in its net cost. Where markets are changing and
The chapter is the opinion of the authors only and should not be taken to represent the views of
their organizations.
1
See, for instance, Cuomo et al. (2013), Haller et al. (2014), Choain et al. (2015), and Robinson
et al. (2015).
D. Nehrebecki (*) • L. Mautino
Oxera Consulting LLP, 200 Aldersgate, London EC1A 4HD, UK
e-mail: dariusz.nehrebecki@oxera.com
G. Knott
Competition and Markets Authority, Victoria House, 37 Southampton Row,
London WC1B 4AD, UK

76 D. Nehrebecki et al.
uncertain, the use of analysis based on actual historical data cannot be reliably used
as a predictor for the future. Equally, the optimal specification of the USP’s
operation in the absence of the USO is likely to change as the total market shrinks.
In the case of post, while markets are changing and the trends cannot be
predicted perfectly, what is now well established is that there is an ongoing trend
of substitution of physical mail for electronic alternatives (e-substitution), along-
side an increase in delivery of parcels and other physical items (e-fulfilment).2 The
rate of e-substitution is uncertain, but there is both extensive practical evidence and
also evidence-based projections which suggest that there will be further
e-substitution into the future—and that the effects differ for different types of
mail. This will change the use of USO services, and therefore may have an impact
on the size of the net cost of the USO.
This chapter proceeds as follows. Section 2 reviews the relevant literature.
Section 3 considers the conceptual aspects of e-substitution. Section 4 outlines
our model. Section 5 presents the key results in the base case as well as in various
e-substitution scenarios. Section 6 discusses the impact of varying counterfactual
scenarios. Section 7 concludes and draws key policy recommendations. The
Appendix contains supporting evidence and further sensitivity results from our
model.
2 Review of Relevant Literature
In recent years, a number of papers have demonstrated that there are certain criteria
which assess whether a reduced service (e.g., 3 days/week delivery) would have a
net cost or benefit; under certain assumptions, reduced delivery could be more
efficient (e.g., Haller et al. 2014; Fustier et al. 2014; Choain et al. 2015; Robinson
et al. 2015). Most of these papers have primarily focused on the impact of a
respecification of the USO on the financial position of the USP. Some of them
considered whether, given current volume, the net cost represents a cost or benefit to
the USP, while others have tended to consider only limited aspects of the demand.
Other papers have focused on demand elements as part of their analysis of
changes in the scope of the USO. For example, Carslake et al. (2011) looked at
the customers’ potential responses to a reduction in the scope of the USO, in the
context of determining the net cost of the USO, while Houpis et al. (2015) consid-
ered estimates of the willingness to pay for features of postal services to analyze
whether certain characteristics of the USO are welfare-enhancing.
While these papers have considered changes in demand as a result of changes in
the specification of the USO, they have not evaluated the net cost of the USO if
demand fell as a result of e-substitution. One of the few exceptions, to our
2
Parcel deliveries feature prominently in European Commission’s Digital Single Market Strategy,
published in May 2015. See European Commission (2015).
Impact of Market Dynamics on the Net Cost of the USO 77
knowledge, is Cuomo et al. (2013), which explicitly accounted for the impact of
e-substitution in the analysis of the scope of the USO. In particular, they assessed
the impact of USO constraints on the USP cost structure, as well as their combined
effect in a market environment characterized by the shrinking of volumes due to
e-substitution. However, their paper did not differentiate the effect of electronic
substitution of different types of mail on the USO. Our paper explores further this
aspect.
3 Interaction Between e-Substitution and the USO
In this section, we consider the relationship between e-substitution and the net cost
of the USO. In understanding this relationship, it is useful to assess whether
e-substitution reduces the sending of mail by all users equally or affects different
types of mail differently. In planning a response to e-substitution for funding the
USO, it is crucial to see that e-substitution is more suited to certain types of mail
than others. In particular, e-substitution (including substitution by mobile technol-
ogy) is most suitable for replacing urgent communication, where mail is a relatively
slow form of communication. Table 1 considers some of the main applications of
mail and the suitability of e-substitution for those applications.
The exact pattern will vary by use of mail. However, we note that evidence from
the UK is consistent with these differential effects of e-substitution on different
priority of mail, with the use of USO and First Class mail as a share of total mail
falling sharply in the period where data was publicly available.3
Table 1 Comparison of main mail applications

Application Priority Suitability for e-substitution
Social mail Generally high High to very high
Advertising Low Low (electronic advertising is a different
product)
Transactional Low for regular communications, Potentially high, but has varied by cus-
higher for some messaging (e.g., tomer and industry
invoices) Greatest substitutability for urgent com-
munication, lower substitutability for
lower priority record-keeping (e.g., bank
statements)
Publications Variable Medium
Fulfillment Variable Low
3
See, for instance, Jarosik et al. (2013), which shows the declining trend of D + 1 mail (First Class)
as a proportion of total addressed inland mail volumes, and also the different impact of technology
on the demand for different types of mail. This is consistent with evidence on volumes in Royal
Mail (2010), page 4.
This chapter considers a model of the impact of differential effects on the net
cost of the USO. In particular, we consider the potential of this change in the use of
mail to result in a continuously increasing trend in the net cost of the USO. The
primary reason is that the USO, as specified in the Postal Services Directive,
requires companies to have the capacity to service priority mail. That capacity
represents an over-specification relative to what would be required to deliver mail
less affected by e-substitution (i.e., non-priority mail). As lower priority mail
increases as a share of total mail, the net cost of the USO will increase as a share
of total costs, as shown in Sect. 5.2 and the Appendix.
4 The Model
We have developed a simplified bottom-up model that is aimed at illustrating the

potential effect of changing various assumptions on the magnitude (and sign) of the
net cost of the USO. Figure 1 below illustrates the components of the model, which
are explained in the following subsections. For simplicity, we present the results
numerically, in absolute numbers, rather than conceptually.4
4.1 USP Revenues
On the revenue side, for simplicity we assume that the USP carries only two types
of traffic: priority mail (which requires the USO network) and regular mail, which
could be delivered via a lower cost non-USO counterfactual network. The former is
assumed to be a priority service, requiring delivery 6 days a week nationally; the
latter is delivered over the same network in practice, but to a lower specification that
would not require a 6-day delivery network. This is comparable to the service in
many EU countries, such as the UK, where Royal Mail offers a First Class (next
day) service, and a Second Class, D + 3 service. Table 2 below illustrates our base
case assumptions on the volumes and relative prices of the two services (prior to
any e-substitution or change in delivery specification, which are considered as
individual sensitivities).5
4
Our modelling is not based on any actual USP, but instead uses stylized assumptions which are
nonetheless consistent with the typical ranges observed among the European USPs. See Table 5 for
more information. The calculation method is in line with conventional approaches.
5
Note that, for simplicity, we do not explicitly model the relationship between the price of priority
(requiring the USO network) and regular mail (non-USO), the associated elasticities and changes
in demand. A full model would endogenize these relationships.
Factual (with USO) Counterfactual (USO absent)
Non-
Revenues Priority
priority
non-priority
volume marginality
priority to non-priority
Proce-
Costs Delivery Processing Overheads Delivery
ssing
Overheads
re-specified cost assumed

service quality marginality fixed
Counterfactual
Profits Current profits
profits
Net cost of the

USO
Fig. 1 Model outline
Table 2 Assumed revenues

Service line Volume (m) Unit price (€) Revenue (€ m)
Priority 25 1.2 30
Regular 75 1 75
Total 100 1.05 (average) 105
Note: These assumptions are consistent with observed evidence on volume distribution and price
differentials for major European postal operators. See Table 5 for more detail
4.2 USP Costs
On the cost side, we assume three broad categories of costs. Delivery costs
constitute the majority of the operating costs; we have assumed they are wholly
dependent on the number of days a week that the mail is delivered.6 The other two
are processing costs, which tend to have significant volume-related marginalities;
6
For small changes in volume, the delivery costs do not significantly change due to the relatively
small potential for realizing savings through changing the walk sequences. Large savings are
feasible with significant changes in volume, however. We have made the assumption that a class of
delivery costs can be defined which is sufficiently well correlated with the number of delivery days
that it can be assumed to be fully variable with delivery days.
Table 3 Assumed costs (€ million)

Activity Fixed Variable per 6 days Variable with volume Total
Delivery 0 50 0 50
Processing 0 0 25 25
Overheads 25 0 0 25
Total 25 50 25 100
and overheads, which tend to be fixed and not conditional on the volume covered
(at least for moderate changes in delivered volumes).
Table 3 above illustrates our starting assumptions as to the USP’s operating
costs, based on a 6 days/week operation. The cost distribution is broadly consistent
with Royal Mail’s cost distribution in 2010 (Royal Mail 2010). The delivery
network is assumed to be shared between the products.
We have assumed that the USP’s cost structure depends on both the volume it
handles and the specification of the USO as follows. For each day that is saved on
delivery, the USP is able to realize an approximate 16 % cost savings in delivery. In
other words, taking our numerical example, if a 6-day delivery network costs €50m
to operate, a 3-day delivery network would cost €25m (a saving of 50 %).7 In
addition, the USP is assumed to realize a 1 % reduction in processing costs for every
1 % reduction in volume.8
4.3 Hypothetical Absence of the USO
In the absence of the USO, we assume a reference scenario in line with the
conventional treatment in literature, where the USP would cease to operate a
6-day delivery service and instead switch to a service of lesser frequency—in the
base case, we have assumed that it would be a 3-day service.9 Furthermore, the USP
would cease to offer the priority product, and all its mail would be delivered on the
remaining network with a lower delivery specification. On the cost side, the larger
the difference in operating costs between the worlds with and without the USO, the
larger the resulting net cost. This has been extensively modelled in the past.
One key assumption that affects the changes to the revenue of the USP in the
absence of the USO is the volume diversion between priority and non-priority mail.
The higher the level of switching between the two types of mail, the larger the
7
This is a stylized linearity assumption, since the focus of this chapter is not on the optimization of
costs, but rather on understanding the impacts on revenues.
8
This assumption is backed by, for instance, Waterman et al. (2000), and represents a long-term
measure of cost-efficiency. A short-term cost saving would be significantly lower.
9
Albeit this is a simplification, one can think of transferring from a “full service” 6-day model to a
3-day-based XY model along similar lines. See further discussion in Sect. 6.1 below.
proportion of the priority mail that will remain within the USP network for delivery
on a non-priority basis after the USO service is removed.10
Consequently, the net cost would be positively related to this diversion ratio. As
the ratio approaches 1, a larger proportion of the mail handled would nonetheless
remain within the USP networks in the absence of the USO. Hence, removing the
USO would reduce costs without a significant volume downside (the impact on
revenues is dictated also by what happens to prices). In our base case, we have
considered an elasticity of 80 %.11
5 Results
This section of the chapter talks through the mechanics of the net cost calculation
and presents the base case results, as well as the specific sensitivities regarding
e-substitution.12
5.1 The Base Case
Based on the assumptions above, the total net cost of the USO is arrived at by
comparing the expected profits that the USP would achieve in the absence of the
USO (having taken account of all demand and supply side changes). In our base
case example, the USP achieves a €5m margin in the presence of the USO, and its
margin would increase to approximately €21m in the absence of the USO. This
gives an estimated net cost of approximately €16m. Table 4 below illustrates the
calculations.
We have, for simplicity, assumed that the USP would keep the price of its
remaining non-USO service unchanged, at the same level as in the presence of
the USO. Considering that, in the counterfactual, it no longer needs to introduce a
clear price differentiation between the two services, this is a conservative assump-
tion. Assuming any such increase did not dilute volumes, increases to unit price in
the absence of the USO would further increase the net cost of the USO estimation
(since the level of profits in the absence of the USO would increase further).
10
The remainder can transition to alternative delivery networks, digital communication
methods, etc.
11
Note that we have not explicitly modelled the relationship between the relative prices charged
by the USP for the priority and non-priority mail, the price of the single service in the counter-
factual and the degree of volume elasticity one would expect to see if the USO was removed. These
variables are, however, closely linked.
12
Further sensitivities are in the Appendix. It should be noted that all sensitivities have been
considered in isolation, and no feedback loop has been assumed into the other aspects of the model.
It is a fundamental simplifying assumption, and lends itself to further research in the future.
Table 4 Base case net cost results

In the absence
With the USO of the USO
Revenue side
Priority volumes m [A] 25 –
Priority unit price €/unit [B] 1.2 –
Non-priority volume m [C] 75 95
Non-priority unit price €/unit [D] 1.0 1.0
Total revenues €m [E] ¼ A B + C D 105 95
Cost side
Delivery €m [F] 50 25
Processing €m [G] 25 23.7
Overheads €m [H] 25 25
Total costs €m [I] ¼ F + G + H 100 73.8
Total profits €m [J] ¼ E I 5 21.3
Net cost €m Difference in [J] ~16
The model does not include any specific assumptions about the regulatory
framework and constraints that the USP faces in the counterfactual. We have also
assumed that the removal of the USO does not fundamentally change the level of
competition observed in the market (and in particular, there is no local or national
entry from direct delivery competitors). In this context, the assumption that the USP
does not increase prices in the absence of the USO is conservative.
The results we obtain depend largely on the assumptions selected. Additional
sensitivity results are presented in the Appendix.
5.2 Impact of e-Substitution
While the net cost calculation itself will vary according to the assumptions, our
sensitivity analysis demonstrates that the impact of e-substitution will strictly be to
increase the net cost over time, and the greater the rate of e-substitution, the
stronger this effect will be.
The speed of market decline varies noticeably between EU countries. Table 5
shows that for the five countries in the sample, e-substitution ranges from relatively
low levels in Germany to over 10 % decline in the Netherlands. This implies that,
going forward, the composition of the respective USPs’ portfolios is likely to
change rapidly, with significant effects on their operations, and financial
performance.
Table 5 Comparison of the selected postal operators

Germany
France (Deutsche Belgium Netherlands
Parameter UK (Royal Mail) (La Poste) Post AG) (bpost) (PostNL)
Price 1.17 (63p vs 54p) n/a n/a n/a n/a
differential
Addressed 4.0 % 5.8 % 0.7 % 4.4 % 10.7 %
mail vol-
ume
decline
Single ~25 %a n/a n/a n/a 23 %
piece mail
proportion
Delivery 6 days a week 6 days a 6 days a 5 days a 5 days a week
days week week week
Cash oper- 4.0 3.7 8.3 15.1 11.5
ating mar-
gin (%)
Notes 1c and 2c differ sig- n/a n/a n/a No clear quality
nificantly in terms of differentiation on
speed of service (D + 1 domestic mail
vs. D + 3) (all 24 h)
Note: When mail delivery forms part of a larger group of operations, only the segmental results
relating to the business division containing the USO business has been taken into account
Sources: UK—Royal Mail’s 2015 rate card, available online at http://www.royalmail.
com/business/system/files/Royal-Mail-UK-stamps-and-Franking-prices-wallchart-30-March-2015.
pdf, Royal Mail website, Royal Mail 2013–2014 Annual Report, Ofcom’s annual postal market
monitoring update, http://stakeholders.ofcom.org.uk/binaries/post/post/Annual_monitoring_
update_2012-13.pdf; DE—Deutsche Post 2014 Annual Report, and operator’s website; NL—
PostNL 2014 Annual Report, Post NL January 2015 rate card and website; BE—bpost 2014
Annual Report and operator’s website; FR—La Poste 2014 Annual Report and website
a
Refers to the proportion of USO mail
This can also be illustrated with reference to a net cost estimate for a notional
USP both today and in 5 years’ time. Figure 2 below shows the range of net cost
estimates for USP’s facing various rates of USO mail e-substitution.
The overall net cost, not surprisingly, increases over time, as the USO mail is
eroded. There are two effects which are driving the increase in the net cost; first,
that there is lower volume but some fixed costs; and second, that there is a lower
proportion of volume which requires the USO network rather than the counterfac-
tual network. These sensitivities are discussed in the Appendix. The two effects are
broadly similar in magnitude.
This carries an important implication for policy design. Using backward-looking
data and assumptions, in particular on volume trends, presents a challenge to a
successful operation of USPs. This is particularly true in countries characterized by
high levels of USO mail erosion.
For example, consider a policy choice on whether to maintain the current USO
or reduce the USO specification. This would generally take a number of years to
22
20
18
16
14
12
10
2015 (reference, 2020, after 2020, after 2020, after 2020, after
base case) annual 1% USO annual 4% USO annual 6% USO annual 10% USO
mail e-sub mail e-sub mail e-sub mail e-sub
Fig. 2 Net cost of the USO today and given a future e-substitution pattern (€ million). Note: Each
scenario assumes e-substitution for non-USO mail at 2 % per annum, and no price response by the
USP to a fall in volumes; no efficiency improvements are factored in. If such a response was
allowed, such that the starting level of margin in all scenarios remained at €5 m, the expected level
of net cost is reduced
implement. Examples from the Netherlands and New Zealand have demonstrated
that there could be a long lead time to implement changes. For example, consider a
5-year process to amend the USO. If a net cost, calculated on the basis of historical
data, is used to understand the impact of policy changes, this could understate the
net cost in 5 years time by up to a third (in our model, and assuming high levels of
e-substitution).
Similarly, consider the impact of competition on the financing of the USO, and
the policy options for implementing liberalization. A static analysis may consider
that there is no adverse impact of competition and that any costs from competition
will be materially outweighed by the benefits. However, if the net cost of the USO is
strictly increasing over time, then the impact of competitive distortion caused by the
USO will also be increasing, and this will affect the need to intervene and the
benefits of policies which have the effect of improving the competitive position of
other postal operators (such as rules on access obligations).
6 Extensions and Discussion
The specific results on net cost, and the associated policy implications, rely
significantly on the assumption about the relevant counterfactual. We already
explored some of the main cost-side sensitivity, namely the service quality in the
absence of the USO (proxied here by the number of days of the non-priority service
per week in the counterfactual). In this section, in order to explore the wider policy
implications, we explore the following adjustments to our base model. First,
varying service quality either on a zonal basis, or introducing quality differentiation
in the non-priority services. Second, introducing some volume diversion of priority
mail into the USP’s premium network (which typically carries time-sensitive letters
and parcels).13
6.1 Zonal Service Variations and the XY Delivery Model
In the absence of the USO, a plausible scenario would be to assume that the USP
would change the delivery specification on a zonal basis, such that areas that are
characterized with higher cost to serve are subjected to a lower delivery specifica-
tion. Typically, this would affect rural areas, as well as some less densely populated
suburban areas, with urban centers retaining frequent deliveries. An example of a
country that has recently implemented a USO respecification along these lines is
New Zealand.14
Any change in the specification of the USO that lowers the service requirement
allows to remove some of the incremental costs linked with the priority service, and
hence is more likely to reduce the overall net cost. This assumes that the cost
savings linked with the respecification outweigh the revenue losses associated with
lower service quality in certain areas. The same logic applies to analyzing coun-
terfactual operations of the USP in the absence of the USO, and hence can have a
significant impact on the level of the overall net cost in the system.
One variant of the change of delivery specification would be a shift from a “full
service” model, whereby both priority and other mail is delivered within each zone
every day, to a “XY” model. Post Danmark introduced this setup in 2008.15 In their
case, the USP splits the delivery routes in sections, and while priority mail and daily
newspapers are delivered in all sections on a daily basis, non-priority, business,
advertising, and similar mail are delivered only every 2 days to a particular delivery
13
For example, in the UK, Royal Mail’s selected “Special Delivery” products are delivered via
such a network, which maintains large degrees of separation from the remainder of the USP’s
operation.
14
From 1 July 2015, the minimum delivery frequency is: not less than 3 day per week deliveries to
99.88 % of delivery points; and 1 day per week deliveries to the remainder of delivery points.
Notwithstanding the above, New Zealand Post shall continue to provide: 5 day per week delivery
to all delivery points which received the rural delivery service as at 30 June 2013, except for those
points which had a lower delivery frequency; and 2 or 3 day per week deliveries to each delivery
point that received 2 or 3 day per week deliveries as at 30 June 2013; and 5 day per week deliveries
to PO Box and Private Bags, except where they received less than 5 day per week deliveries as at
30 June 2013. Source: Extract from the Postal Deed of Understanding, agreed in 1998, amended in
2010 and 2013, available online at http://www.med.govt.nz/sectors-industries/technology-commu
nication/postal-policy/postal-deed-of-understanding.
15
See, for instance, Sondrup (2011).
point (i.e., some sections of the route get service today and others tomorrow). This
allows for a portion of the overall delivery costs to be saved, albeit the overall
impact on the net cost calculation would still depend on the balance of cost savings
and revenue impacts.
6.2 Premium Network
The analysis in this chapter shows that in the context of a shared network which
services both priority and regular mail, the net cost will increase as the proportion of
regular mail increases. We noted in Sect. 3 above that the primary application of
mail which has been resilient to the ongoing decline in mail has been fulfillment.
The fulfillment of physical items has increased as a share of total mail carried on the
USO network. Therefore, the counterfactual assumption in respect of fulfillment
will become increasingly important in understanding the net cost of the USO.
In this context, we note that while for standard items, the alternative to the USO
is a slower (low priority) service, there are generally two alternatives for senders of
physical items—the low priority network or a premium network. The premium
network will be higher price and (assuming a share of shared and common costs)
will have a higher incremental value to the operator.
As a result, there is a simple relationship between the existence of a premium
network and the net cost of the USO. For every item which would be switched to a
premium network (which would continue to be offered under a commercial coun-
terfactual), this will strictly increase the net cost of the USO.
In other words, the USO requires the postal operator to offer a priority fulfillment
service which can be used as an alternative to the commercial premium service. In
the absence of the USO, senders of mail using the priority fulfillment service have
three options: downtrade to lower priority commercial service, uptrade to higher
priority commercial premium service, or stop sending items (including switching to
a non-USO provider).
The “best” outcome for the USP is that senders uptrade to a higher priority
service—although this outcome is a valid counterfactual only for fulfillment items
and where there is a premium network (we have not modelled this explicitly).
Therefore, if (1) there are more fulfillment items, (2) there is a premium network,
and (3) some customers choose to uptrade, then this would improve the commercial
position in the absence of the USO, which further increases the net cost of the USO.
7 Key Conclusions for Policy Design
Our modelling highlights a number of findings that would need to be considered

when undertaking a net cost of the USO assessment, in particular in making
forward-looking policy decisions. In terms of modelling the net cost itself, there
are two key findings for practitioners. Specification of the demand scenario in the
counterfactual (i.e., absence of the USO) is as important, if not more important, than
the corresponding cost savings. How the USP would respond in terms of pricing its
non-USO offering, and correspondingly what volumes it would carry across its
redesigned network constitutes a major element of the net cost estimate.
e-Substitution erodes the volumes carried by the USP, and in particular the
volumes of priority mail. This could significantly affect the structure of the
USP’s portfolio, which has been changing in recent years, and is likely to continue
to change in the future. We have shown that the proportion of priority mail in the
USP’s network is a major factor in net cost calculation. Policy design should factor
in not only the contemporaneous, but also the forward-looking expectations on the
USP’s portfolio. Deriving a backward-looking net cost for the purpose of setting up
a compensation mechanism or another USP subsidy may not fully reflect the rapidly
changing realities of today’s mail markets, in particular in certain European coun-
tries characterized by very high rates of e-substitution.
Appendix: Robustness of Our Assumptions
The assumptions used in our model are stylized, but have been based broadly on the
typical values observed within the European postal sector. As a benchmark, Table 5
below shows the relevant metrics for a range of operators.
Selected Model Sensitivity Results
Table 6 below shows a cross-tabulation of net cost results for a range of input
assumptions on quality of service in the absence of the USO and the level of volume
cross-elasticity once the priority service is removed.
Positive values represent a net cost. Negative values indicate that the commer-
cial offering associated with the USO confers a form of a “benefit” on the USP, in
other words, in the absence of the USO, the profitability of the USO would be lower
if the USP were to move to the notional reference scenario. If that is the case, the
USP could choose in practice to continue to provide a “USO-like” service, and
continue to capture the associated profits, and therefore the true net cost of the USO
(at least relating to delivery specification) is zero.
It should be noted that the extreme results are unlikely to occur in practice, due to a
number of secondary effects (for instance, competitive entry or regulation). Any such
feedback loops have not been included in our estimates of counterfactual profitability.
Nonetheless, the approach is primarily aimed at comparing first-order effects.
The USO network requires a certain level of volume leverage in order to be
viable. If the priority mail constitutes a significant proportion of total mail carried
by the USP, one would expect that this mail would give the USP a significant
Table 6 Dependence of the net cost on volume estaticity and counterfactual service specification
(€ million)
Number of service days per week in the absence of the
USO
1 2 3 4 5 6
Volume elasticity priority to 0% 18 10 1 7 15 24
regular mail 10 % 20 11 3 5 14 22
20 % 22 13 5 3 12 20
30 % 24 15 7 1 10 18
40 % 25 17 9 0 8 16
50 % 27 19 11 2 6 14
60 % 29 21 13 4 4 13
70 % 31 23 14 6 2 11
80 % 33 25 16 8 0 9
90 % 35 26 18 10 1 7
100 % 37 28 20 12 3 5
revenue boost, and hence offset the incremental operating costs. This intuition is
confirmed by our findings in Table 7 below.
If the world without the USO is characterized by a significant reduction in
service quality (approximated here with the days of service), the net cost increases.
Furthermore, the lower the proportion of priority mail in the USP’s network, the net
cost increases as well. For instance, starting from the base case of 25 % of priority
mail and 3-day service in the absence of the USO (which gives a net cost of €16m),
the net cost increases to ~€23–25m if either there is only 5 % of priority mail in the
system, or the optimal service level in the absence of the USO is 2 days.
Summary on the Importance of Various Parameters

for the Sign and Magnitude of the Net Cost
Out of all the various assumptions considered, the number of days of service in the
absence of the USO has the largest implications for the magnitude of the net cost of
the USO (in our numerical example, up to approximately €8m).16 Among the other
factors, as Fig. 3 below indicates, the blend of priority and regular mail is crucial in
determining whether the USO would impose a net cost or confer a net “benefit,”
with the remaining assumptions having broadly the same impact.
16
This assumes that the delivery costs are fully scalable and linearly linked with the number of
days of service—and corresponds to an approximate 16 % reduction in delivery costs (and a 8 %
reduction in total costs). When this assumption is relaxed, as shown in Sect. 5.5, the magnitude of
this effect diminishes.
Table 7 Dependence of the net cost on distribution of the mail carried and counterfactual service
specification (€ million)
Days per week without USO
1 2 3 4 5 6
Proportion of priority mail (%) 0 42 33 25 17 8 0
5 40 32 23 15 7 2
10 38 30 22 13 5 4
15 36 28 20 11 3 5
20 35 26 18 10 1 7
25 33 25 16 8 0 9
30 31 23 15 6 2 11
35 29 21 13 4 4 12
40 28 19 11 3 6 14
8
7
6
5
4
3
2
1
0
One fewer Proportion of Priority price Delivery Delivery cost Volume
delivery day in priority mail differential efficiency marginality elasticity
the absence of increased by increased by increased by higher by 10% higher by 10%
USO 10% 10% 10%
Fig. 3 Comparison of modelled net cost responses to changes in various input assumptions
(€ million)
This confirms our original hypothesis that the volume handled by the USP, in
particular the blend of priority and non-priority mail in the network, has important
implications for the net cost. In particular, in the presence of rapidly changing
markets and high rates of e-substitution in some mail segments, this presents a clear
challenge for the net cost calculation, and correspondingly affects whether the
imposition of the USO confers a net cost on the USP.
References
Carslake I, Houpis G, Strobel C (2011) Costing the universal service: evaluating the demand
response’. In: Crew M, Kleindorfer P (eds) Reinventing the postal sector in an electronic age.
Edward Elgar, Cheltenham, pp 142–151
Choain L, Fustier F, Janin L (2015) Potential gains and losses of (partial) lifting of home delivery
obligations. In: Crew M, Brennan T (eds) Postal and delivery innovation in the digital
economy. Springer, New York, pp 143–154
Cuomo M, Nardone T, Rovero A, Scarfiglieri G (2013) Electronic substitution and USO scope
definition’. In: Crew M, Kleindorfer P (eds) Reforming the postal sector in the face of
electronic competition. Elgar, Cheltenham, Chapter 13
European Commission (2015) Communication from the Commission to the European Parliament,
the Council, the European Economic and Social Committee and the Committee of the Regions:
a digital single market strategy for Europe. COM (2015) 192, 6 May
Fustier F, Janin L, Sahly R (2014) Net cost calculation: a practical example concerning La Poste
and its territorial presence obligation. In: Crew M, Brennan T (eds) The role of the postal and
delivery sector in a digital age. Edward Elgar, Cheltenham, pp 214–226
Haller A, Jaag C, Trinkner U (2014) Calculating the net cost of home delivery obligations. In:
Crew M, Brennan T (eds) The role of the postal and delivery sector in a digital age. Edward
Elgar, Cheltenham, pp 227–239
Houpis G, Rodriguez J, Strobel C (2015) Welfare impact of changes in the scope of the USO. In:
Crew M, Brennan T (eds) Postal and delivery innovation in the digital economy. Springer,
New York, pp 131–142
Jarosik M, Nankervis J, Pope J, Soteri S, Veruete-McKay L (2013) Letter traffic demand in the
UK: some new evidence and review of econometric analysis of the past decade. In: Crew M,
Kleindorfer P (eds) Reforming the postal sector in the face of electronic competition. Edward
Elgar, Cheltenham, pp 194–210
Robinson M, Klingenberg J, Haller A, Trinkner U (2015) Estimating the financial impact of
discontinuing Saturday delivery of letters and flats in the U.S. In: Crew M, Brennan T (eds)
Postal and delivery innovation in the digital economy. Springer, New York, pp 155–168
Royal Mail (2010) Regulatory financial statements 2009-2010. http://www.royalmailgroup.com/
sites/default/files/regulatory_financial_statements_2009-10_audited.pdf
Sondrup J (2011) Operations and delivery systems as a short term response to universal service
demands. http://www.wik.org/fileadmin/Konferenzbeitraege/2011/13th_Koenigswinter_semi
nar/S3_1_Sondrup.pdf
Waterman R, Rubin D, Thomas N, Gelman A (2000) Simulation modelling for cost estimation’.
In: Crew M, Kleindorfer P (eds) Current directions in postal reform. Kluwer, Boston, pp
171–193
Innovation, Disruption, and Partnering:
Changing the DNA of Posts from
the Inside Out
Adam C. Houck
1 Introduction
The global postal industry finds itself in the midst of a disruptive transformation:
one that was born of globalization, liberalization, and corporatization, was fueled
by technological substitution, and is now being accelerated by the growth and
proliferation of other new technologies such as the Internet of Things (IoT), social
media, and cognitive computing. Whereas technology-based substitution or more
specifically, the pace of such substitution, struck fear in the hearts of Postal
Operators (POs), and dominated the transformation agendas of the recent past, it
is the premise of this chapter that technology adoption, and the pace of such
adoption, present an equally significant yet far more agreeable opportunity for the
future. It is an opportunity to transform last mile delivery, product development,
and customer and employee experiences. Indeed, it can even transform the very role
of the PO within society.
POs have a long history of applying technologies to improve performance and
reduce the costs of core processes, but the payback on such efforts has become
harder to achieve. Automation has already been applied to so much of the value
chain and declining volumes make supporting additional investments in research
and development (R&D) more challenging. Finding and adopting new technologies
for other areas of the business or to expand the business presents challenges,
perhaps the most significant being cultural.
The views expressed are those of the author and do not necessarily represent those of IBM Global
Business Services. The author is indebted to Lorraine Galloway and Jason Carlin for their
contributions to this chapter.
A.C. Houck
IBM Global Business Services, Washington, DC, USA
e-mail: AHouck@us.iBM.com

92 A.C. Houck
Successful private sector firms such as Amazon, Google, and Yamato have
created cultures of innovation and experimentation. They encourage calculated
risk taking to explore new technologies and their implications for product devel-
opment, customer demand, and customer acquisition. In contrast, flourishing POs of
the past 10 years have achieved most of their success by embracing diversification,
acquisition, and territorial expansion strategies. To succeed in the next phase of
transformation, POs will need to adopt new cultural attributes and strategies that
mirror those of their successful private sector counterparts. This is especially true
with regard to exploring and applying technology. Identifying, testing, and adopting
cutting-edge technologies will likely require POs to partner with technology firms
to acquire new solutions and skills. Indeed, for POs to capitalize on the new
opportunities, they may very well need to change elements of their DNA, from
their approach to partnering to their appetite for calculated risk taking and exper-
imentation in order to become more essential to not only senders and recipients, but
citizens as well.
Section 2 explores the nuances between innovation and strategic intent. Section 3
examines some of the disruptive technology trends that are most directly applicable
to POs. Section 4 explores how POs can partner with technology firms to acquire
the competencies needed to identify and react to new trends; it also highlights the
need for POs to adopt cultures of innovation that mirror the organizational models
of successful private sector counterparts. Conclusions are presented in Sect. 5.
2 Innovation and Strategic Intent
Innovation is a concept that is often elusively defined yet wholly recognizable when
manifested in its purest form, and it differs from strategic intent. Consider some of
the most well-known inventions, including X-rays, radar, Velcro, and the telescope.
All are clearly seen as innovative approaches to solving existing challenges. All are
acknowledged as having changed the status quo. They meet the Merriam Webster
definition of innovation as “do[ing] something in a new way”; therefore, true
innovation is defined by the firm offering a service, pricing a service, or performing
a service in a new way. Although the recent actions of POs to diversify into
financial, insurance, banking, and e-government services are significant and trans-
formative, they are not necessarily innovative, because the services were not
characterized by new elements and efficiencies.
Myriad forces shape the need for firms to pursue innovation. Perhaps no force is
greater than that of an industry undergoing extreme disruption and transformation.
As noted in an IBM Institute for Business Value (IBM) study, “During periods of
extensive industry change . . ., companies must choose to either shake up their
industries—harness disruptive technologies, go after new customer segments, dis-
lodge competitors—or face their own demise” Giesen et al. (2009). This topic was
also explored by Houck (2014). The global postal industry, perhaps more than any
other industry including telephony, radio, and music, has experienced the most
Innovation, Disruption, and Partnering: Changing the DNA of Posts. . . 93
profound disruption and transformation over the last decade. Fundamental

e-substitution effects, new entrants into the value chain, and liberalization have
reshaped not only the sender–receiver relationship, but also the very role of the PO
in society. Perhaps once solely viewed as a communication mediator among
citizens, governments, and businesses, POs can now occupy a position with other
firms such as Amazon, UPS, and FedEx at the center of local and national
ecosystems that enable the growth of global commerce. As POs seek to become
more essential within ecosystems, we must recognize that innovation can look
dramatically different in the new sharing economy than in the past.
Indeed, “innovation has become more open and increasingly occurs within
economic ecosystems. . .it is about building organizational and ecosystem-wide
capabilities to execute and deliver” (Burstein and Black 2014, p. 1). Classical
examples of innovations in core postal operations such as the first United States
Postal Service (USPS) semiautomatic parcel sorting machine in 1956 and the first
computer-driven optical character reader (OCR) in 1982 can today manifest quite
differently. Instead of the next best parcel sorter, today’s POs are investigating and
pursuing IoT ecosystems of sensor and communication assets that leverage existing
physical infrastructures. These sensor networks—both stationary and mobile—can
exist in and travel through neighborhoods as many as 7 days per week and can
monitor and collect environmental air quality data and video analytics data to help
cities manage traffic flows, improve the movement of vehicles and goods, and aid
emergency organizations in more granular disaster planning. These networks become
only more valuable in developing countries, where POs can aid governments in
delivering services far beyond simple mail delivery in improving citizen lives.
Additionally, the proliferation of global e-commerce has sparked a commensu-
rate shift in global PO supply chains, from mail collection and processing to
delivery. POs now find themselves needing to adjust their thinking from simply
trying to fill letter density in the last mile to understanding how to become more
essential to every citizen. The growth of e-commerce drives the focus to deliver an
outstanding customer experience within the entire value chain and forces POs to
think more broadly about innovation in order to become more essential within the
ecosystem.
Unlike previous advances in automation technology which occurred within the
core of postal operations, innovation is now occurring at the edge. The innovations
within existing core operations have allowed POs to reduce cost and maintain price
and service quality, but now the energy that drove these advances must be
channeled to motivating all areas of the business to innovate and drive a culture
as seen in fringe startups and those firms who are cream-skimming profitable
delivery volumes in the last mile. Evidence suggests past product and service
offerings in these markets by POs were nationally focused and based on a USO
coverage mentality. This mentality, however, runs contrary to what drives innova-
tion today as evidenced by successful startups both inside and outside the postal
sector. Companies such as Capital Bikeshare and ZipCar have created innovative
product offerings that addressed niche geographical needs, then allowed the
94 A.C. Houck
business models to proliferate instead of seeking a single national product offering

that could address every need in every local community.
This brings us to strategic potential and intent. POs must break this USO mindset
when it comes to innovation, especially in markets where products are not protected
by monopoly. Except for the introduction of scanning technologies and services
such as Delivery Confirmation, conceptually, First Class Mail (FCM) letter delivery
has remained relatively unchanged in the USA since the early twentieth century
with the advent of Rural Free Delivery. The physical process of delivering mail
today still mirrors processes centuries old. Indeed, citizens could very well have the
same physical mailboxes from the nineteenth century and be acceptable delivery
receptacles for POs today. POs must clearly not ignore core products like FCM and
operations at the expense of seeking new business opportunities, but must acknowl-
edge the importance of the larger ecosystem and the competitive advantages they
possess to deliver greater value to citizens.
No other entities possess the physical presence of POs. In every neighborhood
6 and 7 days per week, with over 600,000 employees, 211,000 vehicles, 172,000
mail collection boxes, 48,000 facilities, and 1.2 billion miles driven annually, the
physical network of the USPS cannot be matched (USPS 2014). As similar infra-
structures exist for other POs globally, this omnipresence creates an advantage for
the PO to play a much larger role in local ecosystems than simple mail delivery.
However, in order to do so, the PO must expand the boundaries when seeking
innovative products, services, and roles it can offer to citizens.
POs can capitalize on this omnipresence by adopting new technologies, analyt-
ical techniques, and data platform strategies to transform their roles from the
“spoke” to the “hub” of local ecosystems. Some POs are already seizing this
opportunity and offering innovative last mile services. An example is bpost’s
City Logistics Project in Antwerp, which “cuts the number of individual deliveries
made to the city center by consolidating items within a facility on the outskirts of
the city first” (Post and Parcel 2014, online1). The reduction in overall traffic has
maintained a high delivery quality (99.5 %) and is estimated to deliver a 30 %
reduction in the number of kilometers driven and pollutants reduced (Ibid.)
In an age defined by the creation, accumulation, and analysis of data, POs are in
a unique position to create new products and improve existing products and support
and facilitate change for others in the ecosystem as well. Not only can POs provide
additional data to help advertisers better micro-target customer segments for direct
mail pieces, but POs can “find ways to exploit the linkages between parcels and
advertising mail themselves or assist retailers and e-tailers in this endeavor as a
trusted third party information broker” (Houck 2014, p. 296). Indeed, POs can play
a critical role throughout the letter and parcel value chains as they both seek
innovative new solutions to existing problems and assist others within the ecosys-
tem to increase the benefits to customers and citizens.
1
http://postandparcel.info/63493/news/innovation/bpost-set-to-expand-urban-logistics-project-
to-brussels/
3 Disruptive Technology Forces
As the mail mix continues to shift and parcels garner a larger share of total volumes,
regionalized innovation has proliferated into cream-skimming activities created by
smaller technology and logistics providers. The continued forecasted growth of
both domestic and global e-commerce suggests that the list of innovative offerings
will continue to grow. Innovation and disruption in the last mile has seen fringe
companies such as Zipments and Shutl utilize data, analytics, and their brand to
cream-skim profitable delivery volumes in high density metropolitan areas.
Several key forces are fueling this growth. Customers are demanding more
flexible, customized delivery experiences; but at the same time want low shipping
costs or no charge for the delivery of e-commerce parcels. Models abound that
deliver on both of these expectations, demonstrating the ability of market leaders to
solve the chicken-or-the-egg problem of whether price or quality matters most by
appearing to offer both, simultaneously; Amazon Prime is probably the most well-
known example. Interestingly, whereas airlines are going the way of displaying the
costs of individual service components to consumers, e-commerce firms are choos-
ing to bundle all service components into a single product price and offer volume
discounts consumers may or may not choose to use such as with Amazon Prime.
Global governments are concurrently pursuing ways to offer more value to citizens
and digitize traditional government services both autonomously and by require-
ment, such as the Digital Agenda in the European Union.
Concerning delivery, the last mile is about filling density, and not only for letter
volumes. Online grocery delivery services such as Amazon Fresh and Instacart,
along with other innovative offerings such as bposts’ City Logistics project,
continue to strengthen the density in the last mile that high volume logistics
networks require in order to function optimally. Evidence would suggest programs
like bpost’s are the tip of the spear and that regional collaborative logistics is just in
its infancy. The data, analytics, and strategic intent to collaborate and partner in the
new sharing economy are reaching a maturity where providers, both large and
small, understand the benefits and possess the technologies and capabilities to
partner within the ecosystem. They also realize that regional market needs are
heterogeneous and that developing innovative products means solving these local
challenges and allowing those products to grow and evolve in order to meet more
national needs as they arise.
Innovation, entry, and disruption seen in the last mile are now appearing in the
first mile. For parcel products, innovative new companies like Shyp are offering
pickup and logistics services that are disrupting the traditional parcel shipping
model. For a flat USD $5 fee, “one of Shyp’s couriers will pick it (the item) up,
package it, and ship it off using the best carrier option: FedEx, DHL, UPS, or the
U.S. Postal Service” (O’Brien 2015, online2). The growth of these models, just as in
2
http://money.cnn.com/2015/04/21/technology/shyp-series-b/
96 A.C. Houck
the last mile, demonstrates that customers are willing to pay a premium for
convenience and quality of service.
A significant challenge facing POs is that these new providers are delaying
profits while they first seek deep market penetration, not dissimilar from the larger
Amazon penetration strategy. Amazon “recently announced free shipping for
anyone on products that weigh less than eight ounces and cost less than $10. . .
(Amazon) is betting that by investing in shipping and perfecting the fulfillment
process, it will create a huge barrier to competition” (Kline 2015, online3). Not
being able to adopt similar strategies due to existing regulations and reductions in
overall R&D funding, POs must carefully evaluate the strategic moves they must
make and seek ways to not only improve delivery, but leverage their omnipresent
physical networks to offer new innovative services that combat this new entry.
To its credit, USPS has demonstrated a recent willingness to test nontraditional
service offerings such as its experiment with providing fulfillment services for
small to mid-size e-commerce businesses. From a warehouse outside urban Chi-
cago, USPS is examining “the feasibility of providing pre-staged warehousing
delivery services for third-party companies” (Solomon 2015, online4). A recent
USPS Office of Inspector General (OIG) report indicated that USPS could indeed
offer basic fulfillment services, which could position the PO for additional service
offerings such as warehousing, inventory management, and logistics. Importantly,
the move “could expand USPS’ offerings beyond the handling and delivery of mail
and allow it to join the ranks of global posts that have long since expanded into
nontraditional segments” (Solomon 2015, online5). The key question remains
whether these new services would be truly innovative or simply demonstrate a
willingness to diversify service offerings into nontraditional spaces for USPS, for as
Hamel and Prahalad (1989) noted, “Imitation may be the most sincere form of
flattery, but it will not lead to competitive revitalization” (p. 2).
Evidence suggests that private investors see the opportunity for these alternative
industry models to succeed, as Shyp has attracted USD $50 million in recent
venture capital funding, bringing its total funding to USD $62m (O’Brien 2015,
online6). Although this approach is likely not a long-term strategy for POs, it is
important to understand how the competitive landscape is evolving. Indeed, the
approach of establishing innovative service offerings in the first and last mile, then
using that foothold to begin addressing other market needs such as solving the ever-
present issue of returns and reverse logistics, is a strategy that fringe firms can likely
adopt more easily than POs. This short-term profit tradeoff that erects these barriers
3
http://www.fool.com/investing/general/2015/06/08/is-this-the-biggest-threat-to-amazons-busi
ness-mod.aspx
4
http://www.dcvelocity.com/print/article/20150423-usps-testing-e-fulfillment-services-for-small-
mid-size-e-commerce-businesses/
5
http://www.dcvelocity.com/print/article/20150423-usps-testing-e-fulfillment-services-for-small-
mid-size-e-commerce-businesses/
6
http://money.cnn.com/2015/04/21/technology/shyp-series-b/
to entry suggests significant, long-term challenges for POs, due to regulatory

constraints and other factors such as declining capital investment and R&D bud-
gets. However, that is not to mean that POs should not continue to seek innovative
approaches to deliver value within local ecosystems and throughout the value chain.
PO entry in these markets can provide a unique opportunity when compared to
other firms, who possess a greater ability to discriminate and cross-subsidize
product costs; the ubiquity as well as existing regulation can benefit consumers.
In the USA, USPS is required by statute to take the fairness, equity, and simplicity
of its rate structures as well as the relationships among prices, production costs, and
the value of the service provided into consideration. Such constraints can help
ensure that products and services exist in these markets that reflect the actual cost of
providing the service and reduce potential distortions.
The vast physical infrastructure of POs in neighborhoods is also an area that
cannot be ignored when considering how POs can provide innovative services
within local ecosystems. The Internet of Things provides the ability to draw a
richer picture of the physical systems that POs possess across the globe. While not
an overall new concept, the size and cost of sensor technology, the enhancements in
connectivity and cloud technologies, and culturally, the very features and conve-
nience people demand in the products they purchase is seeing IoT technologies
introduced to industries that are not historically Information Technology intensive.
The postal industry might very well be the next largest area of opportunity for
the IoT.
The omnipresent physical postal infrastructures that exist in every neighborhood
position POs at the center of local ecosystems that can be used to improve citizen
lives. Instrumenting this ecosystem could not only help reduce the cost of failed
parcel delivery attempts and reduce mail theft; new services can be offered such as
the recently announced venture by Japan Post to use letter carriers to offer a range
of citizen services for seniors to help address the significant challenges of an aging
population in Japan. Enhancing existing postal products with IoT features such as
temperature monitoring and real-time GPS monitoring, not just for high value
shipments, could provide the ability to improve on even the leading delivery
provider models. Further, IoT technologies could help deliver a customer experi-
ence that goes beyond connecting senders and recipients, it connects the greater
needs and wants in specific geographies, with local postmen filling potential roles
for crowd-sourced ecosystems of local delivery like bpost.
Outfitting vehicles and collection boxes with IoT sensor technologies could also
allow POs to create information platforms others could use to increase citizen safety
and value in society through mobile air quality data collection, video analytics that
aid cities in increasing traffic and parking quality, aiding transportation agencies in
assessing the changing flows of vehicles and goods in neighborhoods to better plan
the needed infrastructure for evolving population patterns, and aiding emergency
organizations on more granular disaster planning for basic services in the event of
natural disasters. These concepts were most notably explored by Ravnitzky (2011).
It is important to note that these opportunities are not simply theoretical thought
experiments about the art of the possible; these ideas are now in production.
98 A.C. Houck
Chicago’s “Array of Things” initiative “is deploying sensor boxes on light posts
around the City of Chicago, with the goal of providing the government, citizens,
and researchers with information like air quality, noise, and temperature” (Johnson
2014, online7). The question is whether POs will sit idly by and allow nonmobile
IoT deployments to occur in cities such as Chicago or offer up their omnipresent
mobile physical infrastructures to deliver a more innovative approach.
This omnipresence provides POs with the opportunity to layer alternative,
innovative business models to capture new revenues and transform existing eco-
system roles, all seamlessly upon existing operations. However, as IoT ecosystem
innovation is happening right now, most of the activity is occurring at the margins
in specific geographies. As much of the most valuable data can be captured through
technologies that have no impact on operations, POs possess the unique advantage
of collecting this data without affecting the collection, processing, transportation, or
delivery of mail.
4 Changing the DNA of Postal Operators
The promise and opportunities that exist in adopting disruptive, cutting-edge

technologies to the postal industry is not devoid of significant challenges. Cultur-
ally, determining the best business models will require experimentation, adopting a
culture that encourages calculated risk taking, a larger appetite for absorbable
failure, and an extensive evaluation of PO brand. Further, POs must acknowledge
that many of the core capabilities in these new technologies do not exist in house
and as a result will need to partner with technology firms to acquire these capabil-
ities. The very process by which POs partner with firms to drive both innovation and
strategic direction requires a careful examination, as the problems likely lie not in
identifying the potential, but in adopting an innovation culture across the entire
organization that encourages experimentation in the first place.
Layering a culture of innovation over a clear path of strategic direction is an
approach that the most essential, forward-thinking firms utilize to great success.
Compared to partnering for innovation, firms and POs have traditionally shown
strong ability to seek out leading partners and strategic consulting firms to set this
clear direction. Indeed, the global postal industry is not lacking in strategic
roadmaps and 10-year plans to increase revenues, decrease costs, and increase
customer satisfaction. What many POs lack, however, is a clear-cut approach to
pursuing innovation across the enterprise and the postal value chain itself.
Beyond the classical examples which illuminate the need to directly pursue
innovation, recent forces have accelerated the need for all organizations, both
private sector and state-owned enterprises (SOEs), to quicken their pursuit. “The
poor economy of the last 6 years resulted in government budget cuts, spurring
7
https://blog.tempoiq.com/the-array-of-things-and-chicago-as-a-data-platform
efforts to find ways to do more with less. More accessible technologies like apps
and social media, and a user community more comfortable with a wide variety of
technologies. . .have also hastened the establishment of government innovation
programs” (Burstein and Black 2014, p. 13). Given the immense pressures facing
the industry, the In-Q-Tel (IQT) model in the US intelligence sector might provide
some examples for consideration to be applied in the postal industry.
Formed in 1999, IQT is an independent, nonprofit organization created to seek
out cutting-edge technologies that have direct application to the intelligence com-
munity and drive innovation within the industry. IQT “identify and invest in
venture-backed startups developing technologies that will provide ‘ready-soon
innovation’ (within 36 months) vital to the intelligence community mission” (IQT
2015, online8). Elements of this approach could prove beneficial for POs as
innovation typically lags behind the pace of technology development in private
sector counterparts, POs can find it quite difficult to attract and retain employees
with the core technology capabilities, and regulatory limitations can prevent POs
from investing significant capital into areas of opportunity such as IoT. Tapping
into a larger ecosystem of innovators would allow POs to have their fingers on the
pulse of the most innovative developments germane to their challenges, more
quickly develop innovative products that meet more demanding customer needs,
and accelerate the speed to market.
Innovations and breakthroughs are clearly not always planned, but that does not
mean that organizations cannot plan for innovation; indeed, some opportunities
should be purposefully pursued even if they possess a relatively small chance of
success. Industry leaders place an emphasis on innovation across the organization
then create specific approaches for each element of business activities to excel in
innovation. A 2014 study from the IBM Institute for Business Value in collabora-
tion with the Economist Intelligence Unit interviewed more than 1000 C-suite
executives and found that “the most successful organizations do indeed approach
innovation differently. The top 9 % of organizations in both operating efficiency
and revenue growth pursue distinct strategies in innovation organization, culture,
and process” (Burstein and Black 2014, p. 1). POs must strategically plan for this
innovation by identifying the areas of greatest promise then directing resources and
energy to capturing those opportunities.
For POs to capitalize on new opportunities, they may very well need to change
elements of their DNA, from their approach to identifying and exploiting many
promising opportunities within the postal value chain to partnering to their appetite
for calculated risk taking and experimentation. Adopting new technologies and
developing new service offerings presents significant challenges for POs, especially
for those with traditionally mail-driven businesses that have not diversified into
other service lines such as banking, insurance, mobile telecommunications, and
e-government services. Innovation requires experimentation and an understanding
that experimentation does not or should not always result in immediate success.
8
https://www.iqt.org/about-iqt/
100 A.C. Houck
Interestingly in the USA, USPS’s risk averse culture has resulted in so few
innovations that those it has are almost always successful. Conversely, when
initiatives of well-known innovative firms, e.g., Amazon fail, they do so frequently,
quickly, and with no regulator closely observing their actions.
Fortunately for POs, they do possess certain key advantages over private sector
firms that position them to capture valuable opportunities. The scale of POs allow
them to take more calculated risks and pursue projects that might not even break
even, as they can more easily absorb the failures compared to smaller fringe start-up
companies. This is not to imply that absorption results in cost recovery. Without
cross-subsidizing products, USPS can simply absorb more failure because of its
size. Traditional SOEs have typically needed to only adopt “sure win” strategies, as
monopoly protections carve out spaces for agencies to operate, just like FCM letters
in the USA. There is typically not an impact on market share when an agency such
as the U.S. Social Security Administration changes the pricing structure on its core
offering; it still owns the entire market.
While POs are indeed SOEs, they must be treated quite differently when it
comes to experimentation and failure as they comprise both market dominant and
competitive product offerings that compete with private sector firms. Interestingly,
Sappington and Sidak (2003) found “the fact that public enterprises may have both
greater incentive and ability than private enterprises to pursue anticompetitive
actions suggests that the costs of public enterprises need to be weighed carefully
against any benefits that such firms may provide” (p. 200). Their research suggests
that public enterprises, such as the USPS, can attempt to broadly define the limits of
monopoly protections to preclude competition and that difficulties arise when
“public enterprises may have even stronger incentives than their profit-maximizing
counterparts to engage in (anticompetitive) activities. . .that serve to exclude rivals
from the marketplace altogether” (p. 196).
Changes to product attributes, service levels, and pricing methods indeed have
an impact on the brand and the financial bottom line of the PO. As such, POs must
look to both the private and public sectors when evaluating approaches to innova-
tion, experimentation, partnering, and failure when introducing new products and
services. For example, a city government might want to take advantage of the
ubiquitous presence of USPS in all neighborhoods to provide monitoring of infra-
structure to promote public safety.
Learning how to fail quickly and well is a defining characteristic of successful
businesses that endure. IBM’s 2014 survey revealed that 31 % of the C-suite
executives interviewed believed that tolerance of failure is critical and that
outperforming firms “are 25 % more likely to accept that some innovation projects
will not succeed” (Burstein and Black 2014, p. 8). In essence, these firms have an
overall greater tolerance of failure and understand that calculated risk taking and
absorbable failure is an essential ingredient to innovation and progress. This is good
news for POs, and confirms that absorbable risks should be strongly encouraged to
test and identify what services current and future customers will demand. USPS
recently demonstrated an increasing appetite for this calculated risk taking with the
market test for the MetroPost same-day delivery product in San Francisco.
By many accounts the product was a failure, as the test failed to generate
impressive volumes. However, this kind of experimentation should be celebrated
as the USPS not only introduced an innovative new product offering tailored to the
needs of a specific geography; it did so quickly, failed quickly, then modified the
attributes of the product and reintroduced the market test in the Washington, DC
metropolitan area in January 2015 with improved success. Learning how to fail
quickly and properly is a skill and one that POs must acquire in order to become
better at risk-reward sharing within the postal ecosystem of potential suppliers and
partners, whether it is in core PO operations from collection to delivery, or new
potential roles POs can offer to local ecosystems.
5 Conclusion
The future holds tremendous promise and opportunity for POs. Compared to
previous technology disruptions that threatened the core value proposition of
POs, current technologies possess the opportunity to benefit POs in equally signif-
icant ways. The technology building blocks already exist, and the challenge is not
necessarily creating new technologies, but adopting cultures that explore and
experiment with how the most applicable technologies can drive value within the
entire postal value chain. In a world of declining volumes and declining R&D
funding, POs must continue to seek ways to mimic the innovative cultures of fringe
last mile delivery firms and larger innovative firms such as Amazon and Google in
order to continue exploring new innovative offerings. Not all ventures will succeed;
and nor should they.
Beyond innovative opportunities in the first and last mile, POs now have the
prospect of redefining their roles in society using IoT technologies, but simply
focusing here is not enough. To identify the most promising opportunities, POs
must seek new engagement models to grow the number of partners within the
ecosystem to cultivate these ideas. The postal industry must learn from successful
ventures and engagement models such as those in the US intelligence community or
risk not only simple stagnation, but irrelevance itself in valuable markets. Failure to
develop non-USO-based innovative products and services will mean letter delivery
will continue to resemble an activity born in the nineteenth century, render letter
delivery as a simple advertising broadcast medium, and keep PO package delivery
as the low cost, last mile option in a marketplace that demands innovative new
service offerings, flexibility, and quality.
102 A.C. Houck
References
Burstein R, Black A (2014) A guide for making innovation offices work. IBM Center for The
Business of Government
Giesen E, Riddleberger E, Christner R, Bell R (2009) Seizing the advantage: when and how to
innovate your business model. IBM Institute for Business Value
Hamel G, Prahalad CK (1989) Strategic intent. Harv Bus Rev 67(3):2–14. Print
Houck AC (2014) The competitive fringe: informing new customer experiences and driving
additional value for recipients. In: Crew MA, Brennan TA (eds) Postal and delivery innovation
in the digital economy. Springer, New York, pp 291–300
In-Q-Tel (2015) About IQT. Accessed 2 Apr 2015
Johnson S (2014) The array of things and Chicago as a data platform. Accessed 1 Apr 2015
Kline, D (2015) Is this the biggest threat to Amazon’s Business Model? Accessed 13 June 2015
O’Brien, SA (2015) Thousands are bypassing the post office with this app. Accessed 22 Apr 2015
Post and Parcel (2014) bpost set to expand urban logistics project to Brussels. Accessed 20 Jan
2015
Ravnitzky M (2011) Offering sensor network services using the postal delivery fleet. In: Crew
MA, Kleindorfer PR (eds) Reinventing the postal sector in the delivery age. Edward Elgar,
Cheltenham, pp 366–379
Sappington DEM, Sidak JG (2003) Incentives for anticompetitive behavior by public enterprises.
Rev Ind Organ 22:183–206
Solomon, MB (2015) USPS testing e-fulfillment services for small, mid-size e-commerce busi-
nesses. Accessed 25 Apr 2015
United States Postal Service (2014) 2014 annual report to Congress. Accessed 29 Mar 2015
Postal Strategies in a Digital Age
Christian Jaag, Jose Parra Moyano, and Urs Trinkner
1 Introduction
Electronic communications impact all of postal operators’ businesses. Traditional

letter mail can be substituted by various electronic means, whereas quality parcels
services are of essence for online shops bypassing the traditional retail channels.
Table 1 illustrates the development of letter mail and parcels volumes for selected
postal incumbent operators. The table reveals important decreases of letter mail of
up to 30 % within 4 years. At the same time, most incumbent operators benefit from
growing e-commerce and see their parcels volumes rising despite the economic
crisis in 2009. According to IPC (2012), parcels may globally increase by up to
69 % until 2020.
These developments represent major challenges and opportunities for postal
service providers. Incumbent postal operators (POs) have reacted and are reacting
differently to those challenges.
In their letter mail business, PO actions have included cutting costs, e.g., by
reorganizing processes throughout the value chain, applying increased automation,
outsourcing of carriers, reducing real wages, increasing prices above inflation in
particular for single-piece mail, reducing service standards and/or universal service
obligations (USO), improving services, for example, intelligent mail or hybrid
services, and leveraging the delivery network, for example, by providing commu-
nity and doorstep services or cross-selling third-party products.
To participate in growing e-commerce markets, incumbent POs have at the same
time steadily developed their parcels business. Many have improved end-to-end
transit times, e.g., by better delivery times (Saturday delivery, Sunday delivery,
evening delivery, selectable time slots) and faster delivery speed. Similarly, many
have improved sender and recipient services (track and trace, collection at home,
C. Jaag (*) • J.P. Moyano • U. Trinkner

Swiss Economics, Abeggweg 15, Zurich 8057, Switzerland
e-mail: christian.jaag@swiss-economics.ch

104
Table 1 Development of letter mail and parcels volumes of selected incumbent operators
AT AU CA CH DE ES FI FR IR IT LU NL NO NZ PT UK US
Mail CAGR 08-12 4.7 3.9 4.8 3.8 3.1 8.2 2.9 4.0 5.1 8.3 2.9 7.5 4.5 4.6 3.7 2.3 5.9
Total % 17.4 14.6 17.7 14.3 11.9 28.9 11.1 15.0 19.0 29.4 11.1 26.8 17.0 17.3 14.1 8.7 21.7
Parcel CAGR 08-12 6.6 NA 0.8 1.3 5.4 NA NA NA 4.1 2.0 NA 10.7 1.6 0.6 1.1 NA 1.1
Total % 29.0 NA 3.0 5.2 23.5 NA NA NA 15.2 7.6 NA 50.0 6.6 2.6 4.2 NA 4.6
Source: IPC (2012)
C. Jaag et al.
Postal Strategies in a Digital Age 105
flexible delivery points, parcel lockers, etc.). Many operators are also expanding
toward logistics (all-in-one solutions, warehousing, etc.).
Similarly, POs have developed their post office networks into different direc-
tions. Many have replaced their traditional post offices by agencies where basic
postal services are provided in third-party retail outlets, e.g., grocery stores. Other
measures include outsourcing or franchising of post offices to third parties, and
leveraging their post office infrastructure to enter new markets, e.g., financial
services, insurance services, or high value retailing.
Incumbent POs have chosen different combinations of these measures. Conse-
quently, they are developing into different directions. For example, Poste Italiane
has become a much diversified group with postal services accounting for less than
15 % of revenues, whereas USPS is constrained by statute to focus on its traditional
mail and parcels business. As a result, postal incumbents strongly differ in the
composition of their revenue distributions, as shown in Table 2. Based on their
revenue mix, the operators are classified as follows: Important role of letter mail
(M), parcels (P) logistics (L), retail services (R), financial services (F), information
services (I), and telecommunications services (T). Note that a sharp delineation of
the revenue components is not possible (e.g., between parcels and logistics). The
values in the table are therefore only indicative.
This chapter aims to derive generic long-term strategies for incumbent operators.
Special attention is given to the legal and regulatory environment which critically
affects a post’s ability to adopt its business strategy. Doing so, the chapter extends
the analysis of Buser et al. (2008) focusing on post offices to a broader context. In a
first step, the various approaches observed in practice are explored by means of
selected case studies. In a second step, a broader overview of performance indica-
tors of postal incumbent operators is provided. The indicators reveal particularly
successful strategies. In a third step, generic strategies are derived.
The chapter is structured as follows. Section 2 provides a brief literature review
on incumbent strategies. Section 3 summarizes selected country studies and
describes the specific strategies pursued by postal incumbents in light of their
regulatory background. In Sect. 4, performance indicators are established and the
incumbent operators are assessed against these indicators. Section 5 develops three
generic strategies. Section 6 concludes.
2 Literature Review
Several authors have studied the strategies of incumbent POs before. Patrovic
(2005) recognizes the role of the state in influencing the strategies of incumbent
POs by setting the regulatory framework in which they must operate. Buser
et al. (2008) discuss the role of the post office in the marketplace and aims to
contribute to a better understanding of post office network optimization and to
identify key strategic issues in light of the regulatory framework. Bailly and
Meidinger (2013) conclude that in view of the decline of the traditional core
106
Table 2 Revenue composition in % and classification of selected incumbent operators

Revenue
composition AT AU BE CA CH DE ES FI FR IN IR IT JP LU NL NO NZ PT RU S-D UK US
Mail 58 36 76 56 33 18 91 49 38 36 75 12 11 25 85 41 48 74 40 56 49 81
Parcels and express 36 50 11 38 15 31 8 12 23 2 15 59 16 18 13 51 19
Logistics and freight 2 51 28 38
Postal retail 6 14 4 14 16 9 25 7 12
Financial services 9 31 22 55 85 82 3 36 8 35
Information services 4 7 1 11 2 6
Telecommunications 0 1 72
Classification P R M P F L M I F F R F F T M P F M F L P M
Source: Swiss Economics based on IPC (2012)
C. Jaag et al.
business of incumbent POs, diversification is the best way forward to find new
sources of revenue and guarantee their sustainable economic development. Green-
ing et al. (2013) analyze the options of POs to enter the area of telecommunications
as a natural complement to the traditional postal business, and briefly compare the
strategies of different POs in this field.
In 2013, the International Post Corporation (IPC) and Boston Consulting Group
published a report on the future perspectives of the postal sector, indicating that
building a new compelling position for incumbent POs requires many fundamental
changes, and that POs need to accelerate from evolutionary to revolutionary
transformation to accommodate revenue decline from increased substitution, and
to seize the window of opportunity in e-commerce. More recently, the Universal
Postal Union (2014) published a book on development strategies for the postal
sector in which it considers specific features of postal markets in developing
countries and traces the emergence of new legislative and regulatory frameworks
in Sub-Saharan Africa. The book also presents the role and strategies of post offices
in the delivery of basic financial services to all citizens. The authors also state that
exploring and making growing use of big postal data, particularly at the interna-
tional level, will empower postal stakeholders, enabling them to take control of the
future of the postal sector, unleash all its untapped potential, and reinvent postal
services for the twenty-first century.
This chapter contributes to the literature by linking the specific regulatory
context in which postal strategies are derived to the behavior and performance of
incumbent operators. Specifically, it outlines the USO in each country and
describes the strategic direction derived by the POs in each of the frameworks.
Additionally, the chapter compares the performance of these POs and derives
generic strategies of proven success.
3 Selected Country Studies
Business models of incumbent POs are bounded to the regulatory framework in

which they are conceived (see also Patrovic 2005). With digital substitution threat-
ening the viability of traditional postal services, the strategies that POs choose to
create new value for customers and maintain or increase their revenues—strategies
not only restricted to enhancing their digital services, but also exploring other
business areas than traditional postal services—are quite diverse. In this section,
illustrative examples from the following six countries are presented: Australia,
New Zealand, the UK, Switzerland, the United States, and Italy. Each example
starts with an overview of the regulatory framework that describes the basic
requirements for the designated Universal Service Provider (USP), followed by
the implemented and planned strategies of the respective incumbent operator. It is
found that the chosen strategies differ significantly. Their performance is analyzed
in Sect. 4.
108 C. Jaag et al.
3.1 Australia Post
As prescribed by the Australian Postal Corporation Act, the postal USP must
provide a universal letter service at a single uniform rate of postage (uniform tariff),
on an equitable basis. Moreover, the USP is required to maintain a minimum of
4000 outlets, of which at least 50 % but not fewer than 2500 must be in rural or
remote areas. These outlets can adopt the form of corporate post offices, licensed
post offices, community postal agencies, or franchises (Howard 2015). Addition-
ally, the USP is to maintain a network of at least 10,000 letter boxes and provide a
mail collection point at each retail outlet. The frequency of delivery is 5-days-a-
week for 98 % of all the addresses and a 2-days-a-week for the remaining 2 %. The
cost of the Community Service Obligations (CSO) is funded by an internal cross-
subsidy within the letter service. The USP has a reserved area to deliver letters
weighting less than 250 g, which are directed to an Australian address.
Australia Post is the incumbent POs in Australia and is a Government Business
Enterprise, fully owned by the Commonwealth of Australia. The corporate vision of
Australia Post, communicated through its Annual Report 2014, is keeping Australia
connected, by delivering mail, helping businesses and consumers with parcels, and
providing a national retail network that supports local communities and organiza-
tions. As indicated in the Annual Report, 2560 of the 4417 postal offices are located
in rural or remote areas, where citizens have reduced access to other than postal
services or products. The primary segments, in which the group operates, are
“Postal Services” that comprise the Mail and Retail segments, and “Parcel Ser-
vices” that currently deliver the strongest performance.1 The strategy for the “Postal
Services” segment is to implement changes that secure the sustainability of the
regulated mail services business, such as increasing the basic postage rate, adjusting
the USO for letters, or continue increasing the number of providers that sign to use
the Digital Mailbox.2
Australia Post offers a very high variety of retail products to its customers in
every post outlier and aims to be the one-stop shop to suit a wide range of needs of
its customers. The product range includes “agency services” for already more than
750 businesses and government bodies in 2011, including banking services for
70 financial institutions (Australia Post Annual Report 2009–10), insurance and
medical services as well as diverse retail supplies such as furniture and food.
The strategy for the “Parcel Service” segment is to create a world-class
multichannel parcels and freight business to harness the growth in online shopping
and ecommerce by, e.g., integrating Australia Post and StarTrack3 to create a
1
Operating EBIT of $337.5 million in 2014 (up 20.8 %) and revenue growth of 16.4 %.
2
The Digital Mailbox is an online platform that gives businesses and government agencies a
secure way to connect to their customers, who can use it as a mobile payment and storage solution
to receive and pay bills, as well as to archive important documents.
3
In October 2012, Australia Post acquired Qantas’ 50 % interest in StarTrack, making Australia
Post the sole shareholder in StarTrack.
logistics provider that serves both, business and consumer markets with an broader
suite of delivery services, or by stimulating e-commerce growth by offering its
customers the choice to direct their deliveries to locations other than their homes.
Moreover, and in order to build capability in its delivery network, scanning devices
have been rolled out across its postmen to enable postmen to deliver small parcels
that require tracking and signature-on-delivery (Australia Post Annual Report
2015).
3.2 New Zealand Post
The USO in New Zealand is defined in the Deed of Understanding between

New Zealand Post and the Minister of Communications and Information Technol-
ogy.4 The USP has to operate a minimum of 880 collection points from which at
least 240 must provide personal assistance. Until June 30, 2015 there must be a
6-days-per-week delivery to more than 95 % of the delivery points, a 5- or 6-days-
per-week delivery to more than 99.88 % of the delivery points and 1- to 4-days-per-
week delivery to the remaining delivery points. This will change from July 1, 2015,
due to an amendment to the Deed of Understanding. From this date on, 99.88 % of
the delivery points must have a not fewer than 3-days-per-week delivery frequency,
and the rest of the delivery points must have at least a 1-day-per-week delivery
frequency. Since there is no compensation fund, all costs associated with the USO
are borne by the designated operator.
New Zealand Post is a state-owned enterprise held by the Minister of Finance
and the Minister for State-owned Enterprises that acts as USP. The corporate vision
of the New Zealand Post, stated in its Annual Report 2014, is being invaluable to
New Zealanders’ lives and businesses. The company consists primarily of two
“clusters of businesses,” namely “Logistics & Communications” services that
include, among others, the provision of the USO, the operation of a nationwide
store network, or the management of time-sensitive delivery warehouse services,
and “Financial Services” that with Kiwibank has a strong growing presence in
personal and small business banking. Kiwibank acts as a very significant contrib-
utor to the Posts’s success.
The strategy of New Zealand Post for the two business clusters encompasses the
following elements: First, build a sustainable physical network that offers the
flexibility to align to changes in demand (e.g., progressively change the delivery
mode from walking and cycling to a walking and vehicle support that will allow its
postmen to accommodate the continued growth in parcels). Second, create a
platform upon which a range of valuable digital services can be delivered to
New Zealanders. Third, develop and deliver, in an economic and sustainable way,
an exceptional customer experience by providing effective products and services
4
https://www.nzpost.co.nz/sites/default/files/uploads/shared/deedofunderstanding.pdf
110 C. Jaag et al.
(e.g., increasingly using agency type models where the New Zealand Post service
offering is co-located within a host business, allowing greater convenience and
longer opening hours). Fourth, ensure long-term value creation for New Zealand
Post through the further development and growth of Kiwibank5 as well as ensure
the internal structure, processes, and frameworks in operation at New Zealand Post,
to offer the most efficient and profitable outcome for the group. Fifth, ensure that all
employees have the skills and capabilities they require to succeed, and create a
culture in the organization that enables and encourages them to strive for great
customer outcomes (Statement of Corporate Intent 2011–2014, New Zealand Post).
3.3 Poste Italiane
In Italy, the USO includes letter mail up to 2 kg and universal parcels up to 20 kg

(both non-express), as well as registered and insured mail, press and editorial items,
electoral items, and all mail related to administrative and judicial procedures.6 The
designated USP must ensure that 75 % of the population has an access point at a
distance of less than 3 km from its residence. Moreover, it has to make sure that
92.5 % of the population has an access point at a distance of less than 5 km from its
residence. Similarly, 97.5 % of the population must be provided by the USP with an
access point at a distance of less than 6 km from its residence. The USP is obliged to
deliver with a 5-days-a-week frequency (with some exemptions for barely inhabited
areas). The USP enjoys no reserved area, except for judicial acts and traffic fine
notices. The USO is partly financed by public compensations from the state budget,
using a “subsidy cap” formula and additionally, POs finance a compensation fund.
However, the yearly contributions to this fund made by the POs have not yet been
significant.
Gruppo Poste Italiane is a 100 % state-owned public limited company, divided
into four operating segments: Postal and Business Services, Financial Services,
Insurance Services, and Other Services. The Postal and Business Service is carried
out by Poste Italiane SpA and certain subsidiaries which is, by Legislative Decree
58/2011 the designated PO for Italy. The mission of Gruppo Poste Italiane is to
5
The support to Kiwibank reveals as logic, given the significant financial contributions of
Kiwibank to the New Zealand Post. Kiwibank has been focusing on the provision of high levels
of service through a wide branch network and high innovation levels for internet and mobile
banking. By June 30th 2014, the segment earnings before income tax reached the first position
within the group, contributing NZD 139M. The success of Kiwibank is based on the accessibility it
offers to its clients by leveraging its existing branch network, the long opening hours and a high
focus on innovation. In its aim to provide innovative solutions to its clients, Kiwibank was the first
bank in introducing real-time mobile phone banking, beating the world’s biggest banks to win the
international 2007 Financial Innovation Award. The heritage of this innovation is appreciated in its
service today with initiatives like “Fetch” to fasten the withdraw speed.
6
Ministerial Decree of 07 October 2008 and AGCOM Decision 342/14/CONS.
consolidate its role as a global operator, drive the development of communications,

payment and logistics solutions, improve services, and create opportunities for
Italy.
The Gruppo Poste Italiane is based on two commercial pillars and three business
areas, supported by corporate units, which address, manage, control, and provide
services to support business processes. The two commercial pillars are the “Post
Office Network” and the “Business Sales and Public Administration.” “Post Office
Network” manages the retail and SME segments and coordinates the post office and
Poste Impresa network, thereby representing the main access channel to Poste
Italiane services and products. “Business Sales and Public Administration” serves
large companies, utilities, banks, and public administrations. “Mail, Logistic, and
Communication Services” provides letters, parcel services, and advertising mail.
“BancoPosta” is the business unit for financial services, whereas “PosteVita” is the
business unit committed to insurance services.
Gruppo Poste Italiane reports its yearly financial activity according to its
four operating segments: “Postal and Business Services,” “Financial Services,”
“Insurance Services,” and “Other Services” (Poste Italiane 2015). One of the
strategic objectives of Poste Italiane in the “Postal and Business Services” market
is to position itself as a provider of highly qualitative business mail services,
providing new shipping and delivery solutions. For letters, however, Poste Italiane
tries to achieve relaxations regarding the numbers of delivery days. In the field of
the “Financial Services,” Poste Italiane aims at further diversify the range of
products offered, strengthen the current position in the account and payments
systems market and consolidate its leadership in the prepaid cards market, by
introducing innovative services in the areas of bank transfers, remittances, and
standing orders. For the “Insurance Service” market, Poste Italiane aims to keep
focusing on the PosteVita insurance group by developing tailored products for its
target segments in combination with multichannel operations. Since its beginning,
PosteVita has developed very successfully and accounted for over 60 % of group
revenues in 2013. Additionally, Poste Italiane has found a growth opportunity in the
non-life segment of insurances, in which it is active through PosteAssicura (Poste
Italiane 2014).7
7
The case of Poste Vita is of special interest, since in only 10 years since its creation in 1999, it
became the first insurance company of the country in terms of premiums. The focus of Poste Vita is
to manage the “new welfare” model that has emerged after the economic crisis. The “new welfare”
model is driven by a growing demand to cut public expenditures. Since often, these cuts let parts of
the society «unprotected» in terms of health-, self-sufficiency-, savings-, and asset-protection,
Postevita focuses on these segments to fulfill their needs and explode this market niche.
112 C. Jaag et al.
3.4 Royal Mail Group
The UK Postal Services Act of 2011 defines the USO requirements for the USP.
Letters have to be collected at every access point and must be delivered on
6 weekdays. Parcels have to be delivered on 5 weekdays. Public prices for a
package service must be affordable and uniform. Moreover, the USP must guaran-
tee that at least 95 % of users are located within 5 km of an access point capable of
receiving the largest relevant postal packages and registered mail. Furthermore, the
premises of not less than 95 % of the users in each postcode area must be within
10 km of an access point capable of receiving the largest relevant postal packages
and registered mail. The USP has no reserved area, since the postal market is fully
open since January 1st 2006. As a result of the provisions of the Postal Services Act,
Ofcom has implemented a new regulatory framework for financing the USO, by
which the universal service provider should be allowed to make a “reasonable
commercial rate of return” while fulfilling the USO. Ofcom recognizes that EBIT
margins between 5 and 10 % for the reported business activities should enable a
financing of the universal postal service on a sustainable basis.8
Royal Mail Group Ltd. is a limited company, owned by the Government of the
UK (30 %), the company employees (minimum required by law of 10 %) and
private investors (60 % free float). It is designated as the USP.9 The vision of Royal
Mail Group is to be recognized as the best delivery company in the UK and across
Europe, by sustaining the continued provision of the Universal Service in the UK
and generating sustainable shareholder value. The two divisions of Royal Mail
Group Ltd. are the UK Parcels, International and Letter (UKPIL) that carries out the
Universal Service Obligation in the UK, and the General Logistic Service (GLS)
that provides business-to-business parcel and express services, as well as logistics
solutions through its network in the domestic markets in which it operates. Both
divisions reported positive operating profits in the years 2013 and 2014.
The strategy for UKIPL is to be a successful parcel business by continuing to
grow parcels and maintain its strong position in the consumer, micro SME and SME
market segment. Moreover, it aims to deliver significant IT upgrades (such as the
launch of Local Collect, a click and collect solution with the Post Office), introduce
new products and maximize the proportion of traffic it can handle in the UK in a
profitable way. Additionally, UKIPL aims to become more customer focused and to
give customers more choices by offering longer opening hours. Since 2002, Royal
Mail Group has increased the prices of first- and second-class stamps for regular
8
Securing the Universal Postal Service, decision on the new regulatory framework, Ofcom,
27 March 2012, § 5.41, p. 50.
9
The facilities to fulfill the USO are provided and owned by Post Office Ltd., as stated in a 10-year
inter-business commercial agreement that allows Post Office Ltd. to continue issuing stamps and
handling letters and parcels for Royal Mail. Post Office Ltd. and Royal Mail Group Ltd. are sister
companies, since both are part of Royal Mail Holdings Ltd.
letters (up to 100 g) considerably.10 The strategy for GLS focuses on the continu-
ation of its expansion and growth, and on the improvement of its offering by, e.g.,
increasing the number of B2C delivery options.
3.5 Swiss Post
In Switzerland, the USO includes the daily conveyance of addressed letters and
parcels, newspapers and magazines, and outbound international letters and the
provision of basic financial services such as bank accounts for residents and
payments. Moreover, Swiss Post has to ensure that its retail network can be reached
by 90 % of the population within 20 min by foot or public transport (for financial
services 30 min). Prices have to be affordable and uniform for single-piece letters,
single-piece parcels, and newspapers. Standard letters and direct mail up to 50 g
currently remain reserved for Swiss Post. Around 70 % of the letters market is
already deregulated. The residual monopoly on domestic letters up to 50 g is an
important pillar for Swiss Post’s financing of the universal service.
Swiss Post is a public limited company under special law that acts as the
designated USP and is fully owned by the Swiss State. The vision of Swiss Post
is to be “simple yet systematic” and connect the physical and digital worlds, setting
new standards with its products and integrated solutions, as well as make it easier
for its customers to operate in today’s complex environment, giving them greater
scope to succeed.
Swiss Post is an incorporated holding comprising three main operating sub-
sidiaries: Post CH Ltd. (with the divisions PostMail, PostLogistics, Swiss Post
Solutions, and Post Office & Sales), PostFinance Ltd. (with the division
PostFinance), and PostBus Switzerland Ltd. (with the division PostBus). Since
2013, Swiss Post is focusing on maintaining and increasing the value of the
company and on achieving industry-standard returns in all business areas. For
PostLogistics, Swiss Post aims to maintain its leading position in Swiss
e-commerce and continue to foster its cooperation with partners to add more than
660 postal agencies to its network by the first half of 2015. Moreover, and in order
to satisfy the market needs, Swiss Post has enlarged its palette of services and
focuses on online vendors. An example of new services is “YellowCube.”11 The
division Swiss Post Solutions aims to further develop the internationalization of
10
The price of a first-class stamp for a standard letter in 2002 was of 27 pennies, whereas by the
year 2015 the same stamp would cost 63 pennies. A similar pattern is to be observed in the prices
of the second-class stamp for standard letters that cost 19 pennies in 2002 and 53 pennies as of
March 2015.
11
“YellowCube” is an all-in logistics solution for distance selling made for online vendors. With
this service, Swiss Post takes care of storage, picking and packing, fast shipping and return
management, so that vendors can fully outsource their warehouse, shipping, and packaging duties.
114 C. Jaag et al.
Swiss Post’s business and to broaden the range of postal-related services in the
context of the digital transformation, focusing on its three core markets, Germany,
the UK, and the United States. PostMail experienced in 2014 a moderate decrease
in income due to lower mail volumes. Post Office and Sales operates the retail
network and postal services for private customers and aims to further tailor its
services to the needs of its retail customers.
3.6 US Postal Service
According to US legislation, the US Postal Service (USPS) must provide prompt,

reliable, and efficient services in all the areas and render postal services to all
communities and serve the entire population of the United States as nearly as
practicable. USPS must receive, transmit, and deliver throughout the United States,
its territories and possessions, written and printed material, parcels, and other
materials, and maintain at least one class of mail for letters with a uniform price
throughout the United States. Uniform rates for First-Class Mail must be provided
for articles weighing up to 13 oz (369 g). The USO does not define a specific
number of post offices or letter boxes. USPS currently fulfills a 6-days-a-week
delivery policy. There is no compensation fund to finance the USO. Profits from
competitive products can be used to help defray costs for meeting the Universal
Service Obligation.
The USPS is an independent establishment of the Executive Branch of the US
Government and has a defined business model imposed by Federal Laws that does
not allow it to utilize its existing asset base to generate additional revenue to offset
the recent large mail declines (Crew and Richard Geddes 2014). Hence, USPS is
forced to restrict its activities to the traditional postal services, such as delivery of
first- and second-class mail, which accounts for over two thirds of its annual
operating revenue, and shipping and parcels, which accounts for the remaining
operating revenue (Annual Report to Congress 2014).
The corporate vision of the USPS, outlined in its Annual Report 2014, is to
improve its services, products, and capabilities to adapt to the changing needs of
customers in the digital age. Given the restrictions imposed by the regulatory
framework, USPS only operates within the traditional mail service with one
division: Mail/Shipping & Packages.
The Office of Inspector General defined in 2014 the five strategic focus areas for
the USPS for the period 2014–2018 that align with the Postal Service’s strategic
goals—preserving reliable and affordable universal service and implementing
comprehensive transformation for a long-term sustainable future—articulated in
the Postal Service’s April 2013 Five-Year Business Plan. The five strategic focus
areas are: “Meeting the Emerging Needs of the Digital/Global Era” by informing
and connecting the postal enterprise, stakeholders, and the OIG through the use of
innovative technology12; “Achieving Twenty-first Century Modernization” by

making recommendations that enhance service to customers, improving operational
economy and effectiveness13; “Ensuring Mailer and Recipient Value” by exploring
emerging opportunities in integrated delivery, micro logistics, and managed deliv-
ery14; “Enhance the Brand and Integrity” by conducting investigations and audits
that enhance the integrity and security of Postal Service products, services, and
people15; and “Recovering and Preventing Improper Payments” by proactively
addressing vulnerabilities and accelerating recovery of improper payments (Office
of Inspector General 2014).16
4 Performance
Table 3 summarizes the main areas in which the investigated incumbent operators
have put particular emphasis. If one operator is not marked in one area, this does not
necessarily mean that no measures were taken in the respective area; it only means
that the major strategic efforts have been concentrated in other areas.
The last two columns of Table 3 report the financial standing of the incumbent
operators. Consistently good margins have been achieved by the incumbent oper-
ators of AU, NZ, CH, and IT. It may be concluded that incumbent operators that put
an emphasis on financial services, perform particularly well in terms of EBIT and
revenue development. It would be false, however, to conclude that all POs should
diversify into financial services and banking. Having already been active in finan-
cial transaction services in the past and disposing of the necessary infrastructures
seem to be important prerequisites for successful business development in this
domain.
Generally, diversified postal incumbents appear to have developed well in
relative terms. Examples of successful diversification strategies among the case
studies discussed above are Australia Post with its community retail services
provided in post offices, New Zealand Post and Swiss Post with their banking
services, and Poste Italiane with its success in selling life insurances. In these cases,
diversification was based on leveraging existing infrastructure, competences, and
12
For example, better integrate information technologies to equip OIG staff with responsive,
mobile tools, enabling them to work on a variety of devices anywhere, anytime, thereby increasing
operational efficiencies and improving the effectiveness of OIG operations.
13
For example, further modernize OIG’s capability to synthesize data and information on postal
operations and turn it into useable knowledge and insight to identify the root causes and solutions
for weaknesses in postal operations.
14
For example, track social, technological, and industry trends capturing patterns of human
interaction, how businesses are organized and how transactions take place.
15
For example, help the Postal Service to self-police employee misconduct by sharing insights
from employee misconduct cases.
16
For example, strengthen internal controls on postal systems.
116 C. Jaag et al.
Table 3 Strategic efforts and financial development of selected incumbent operators

Financial
Major strategic efforts performance
Area 1 Area 2 Area 3 Area 4 Area 5 EBIT EBIT
Change
Financial Innovation EBIT
Multichannel and in parcel/ Increase Margin
vending and insurance mail prices/ Margin from 2007
retail services delivery cut costs Logistics 2012 to 2012
AU X X X 0.079 0.049
NZ X X 0.067 0.023
UK X X X X 0.032 0.019
CH X X X 0.106 0.006
US X X 0.075 0.004
IT X X 0.076 0.028
Source: Swiss Economics based on annual reports and IPC (2012)
reputation. Strategies that focus on a particular market, e.g., KEP markets or

logistics, succeed in compensating losses in the letter market. Overall margins,
however, appear to be more modest.
5 Generic Strategies
From the discussed case studies and performance indicators from above, generic
strategies can be derived which have proved to be successful or are promising due
to their impact on innovation and performance (see Fig. 1). These optimal strategies
are path dependent and ultimately result from the regulatory framework. The
strategies result in specific types of innovation and are reflected in business
performance (see Fig. 1). The reasoning of the derivation of generic strategies is
similar to the structure-conduct-performance paradigm developed by Bain (1959).
Strategy, innovation, and performance in turn can influence regulation, as indicated
by the dotted lines. Whereas regulation strategies aim at influencing regulation
directly, innovations may lead to changes of regulations, for example, based on new
possibilities to fulfill the USO. Performance on the other hand may limit or extend
the scope of fundable USO.
The dependency on the legal and regulatory environment of the successful
strategies and the posts’ ability to adopt them results from a number of interaction
channels: First, the state ownership and the associated principle of legality deter-
mine the legal boundaries of a PO’s scope of business. Second, the scope of the
(past and current) USO determines the USP’s cost structure and its assets for new
business development. Third, the status of market opening determines the degree of
(potential) competition and the USP’s ability to finance its obligation with own
funds (see e.g., Dietl et al. 2005). Fourth, the past definition of the USO and the
Regulation Strategy Innovation Performance
Fig. 1 From regulation to performance
Table 4 Post office network regulations and development

Regulation Strategy Innovation Performance
AU Maintain 4000 retail Bring value to post office IT platform facili- Strong growth
outlets, including network by providing tating high value and profitable
2500 in rural/remote third party services, e.g., retailing with third post office
areas banks parties network
NZ Provide 880 retail Bring value to remaining Launch of Strong growth
outlets, thereof post offices KiwiBank and profitabil-
240 post offices ity of bank
UK Full market opening, Compete for letters and Process Limited
structural separation parcels innovations growth and
of post office EBIT margin
network
CH Minimum accessi- Develop all business Process and prod- Solid growth,
bility, financial divisions and in particu- uct innovations strong devel-
transactions in USO lar PostFinance opment of
PostFinance
US Post office require- Develop mail and parcels Process Substantially
ments, no diversifi- business innovations negative
cation allowed margins
IT Full market opening, Diversify Launch of Strong growth
post offices for 96 % BancoPosta, of new
of municipalities PosteAssicure, services
PosteVita, and
more
performance of the USP determine its reputation and thereby its potential to
leverage it into new business fields where trust is key (see Dietl and Jaag 2011).
Fifth, the scope of the USO determines the viability of (the USP’s and its compet-
itors’) postal and financial services due to the strategic deterrence effect (see Jaag
2011). Sixth, the scope of the USO with associated mandatory services also
determines whether new services must be economically viable on their own or
whether they may benefit from cost savings due to the opportunity to substitute
traditional processes (e.g., physical delivery, cf. Jaag et al. 2015).
Table 4 illustrates the interdependencies based on the case studies from Sect. 3
with emphasis on regulations concerning the post office network. The figure reveals
substantial differences in the evolution of post office networks that can be explained
by different initial regulations.
The interdependencies also determine the optimal type of innovation: the more
flexible a post is in terms of its infrastructures, the more attractive investment in
118 C. Jaag et al.
Table 5 Generic strategies

Innovation Performance Lead
Regulation Strategies type effect example
No infrastructure • Exploit synergies Process Cost efficiency SE
requirements • Increase efficiency innovation
• Focus
Financial transac- • Exploit synergies Process Cost efficiency NZ
tions in USO innovation
• Increase efficiency Product USP in retail
• Diversify into banking innovation banking
services
Infrastructure • Exploit synergies Product USP in commu- AU
requirements • Increase efficiency innovation nity
• Diversify/bring services/high
value into network value retailing
process innovation is. Conversely, e.g., if the USO is strongly input oriented,
product innovation aiming to utilize these assets is more attractive.
Three generic strategies are shown in Table 5. (1) In a regulatory environment
with no infrastructure requirements as part of the USO (i.e., there is no minimum
number of self-run access points or prescribed delivery to every receiver’s door-
step), there is a large potential to increase efficiency by reducing and outsourcing
access points and relocate delivery to centralized postboxes. (2) In a regulatory
environment with a diverse USO including financial transaction services, a hybrid
approach is possible: To diversify into financial and banking services. This also
allows to exploit synergies between the different services. (3) If there are infra-
structure requirements in place, bringing value into the network by diversifying into
services which make use of these infrastructures, is the way to go.
Given the diverse legal and regulatory environment, Table 5 shows that one size
of a corporate strategy does not fit all. This also implies that there are things which
POs can learn from each other in terms of new business development, but only to
the extent that their regulatory frameworks support similar strategic choices.
6 Conclusion
Many POs have significantly and successfully moved away from their traditional
core businesses. In particular, operators with banking services have been able to
achieve growth and above average margins. However, the optimal business strat-
egies are not the same for all POs since they are often only second best, given the
specific regulatory framework. Hence, regulation is relevant for understanding
business strategies and the great policy challenge consists in accomplishing the
optimal coevolution of regulation and its institutions with market developments.
Three successful generic postal strategies have been identified in this chapter. It
is important that postal regulation allows POs to pursue one of these generic
strategies. For example, with binding regulations in place (e.g., USO prescribing
infrastructures or processes), these regulations should be complemented with legal
and regulatory support the following dimensions: First, there needs to be a legal
foundation for diversification into new business to better utilize the necessary
infrastructures or processes. Second, regulated services must be able to be
substituted over time (i.e., discontinued) if there is an innovative alternative in
place. Third, any cross-subsidization of existing or new services should be limited
in order to allow new services to compete on a level playing field. As an example,
regulations that prescribe a minimum number of post offices need to go along with a
corresponding legality and the ability to diversify to bring value into the existing
post office network.
References
Australia Post (2010) Annual report 2009–10

Australia Post (2015) Annual report 2014
Bailly D, Meidinger M (2013) Human capital and diversification choices for postal operators. In:
Crew MA, Kleindorfer PR, Kleindorfer PR (eds) Reforming the postal sector in the face of
electronic competition. Edward Elgar, Cheltenham, pp 350–360
Bain JS (1959) Industrial organization: a treatise, 2nd edn. Wiley, New York
Buser M, Jaag C, Trinkner U (2008) Economics of post office networks: strategic issues and the
impact on mail demand. In: Crew MA, Kleindorfer PR, Campbell JI Jr (eds) Handbook of
worldwide postal reform. Edward Elgar, Cheltenham, pp 80–97
Crew MA, Geddes R (2014) A business model for USPS. In: Crew MA, Brennan TJ (eds) The role
of the postal and delivery sector in a digital age. Edward Elgar, Cheltenham, pp 16–27
Dietl H, Jaag C (2011) Postal and regulatory reform in intermodal competition. In: Sund K, Osborn
D (eds) The future is in the post vol 2: perspectives on transformation in the postal industry.
Libri, Faringdon, pp 49–56
Dietl H, Trinkner U, Bleisch R (2005) Liberalization and regulation of the Swiss letter market. In:
Crew MA, Kleindorfer PR (eds) Regulatory and economic challenges in the postal and delivery
sector. Kluwer Academic, Boston, pp 53–72. ISBN 1-402-07972-9 (print), 0-387-23637-6
(electronic)
Greening M, Dargue M, Chiam H, Appleton S (2013) Postal Operators: Delivering Success
through Telecom Strategies, mimeo, Cartesian
Howard J (2015) Digital post: business transformation and the future sustainability of the Australia
Post. McKell Institute, Sydney
International Post Corporation and the Boston Consulting Group (2012) Focus on the future:
building on a new compelling position for posts
Jaag C (2011) Entry deterrence and the calculation of the net cost of universal service obligations.
Rev Network Econ 10(1)
Jaag C, Maegli M, Morel D (2015) Commercial and regulatory challenges for postal E-services in
Switzerland. Swiss Economics working paper
New Zealand Post (2015) Annual report 2014
Office of Inspector General (2014) Five year strategic plan FY2014-FY2018. https://www.uspsoig.
gov/sites/default/files/document-library-files/2014/oig_strategic_plan_fy14-fy18_0.pdf
120 C. Jaag et al.
Patrovic V (2005) Postal system strategy selections by applying multicriteria analysis. Yugoslav J
Oper Res 16(2):235–251
Poste Italiane (2014) Annual report 2013
Poste Italiane (2015) Annual report 2014
Royal Mail (2015) Annual report 2014
Swiss Post (2014) Swiss post annual report 2013
Swiss Post (2015) Swiss post annual report 2014
United States Postal Services (2012) Five-year business plan
United States Postal Services (2014) Annual report to congress
Universal Postal Union (2014) Development strategies for the postal sector: an economic
perspective
Case Studies in End-to-End Delivery
Competition
Philip Groves and Steven Cape
1 Introduction
The EU letters market was finally opened to full end-to-end (E2E) delivery com-
petition in all EU Member States on 1 January 2013. Nevertheless, due to the
diversity of market conditions, E2E delivery competition had already emerged in
several Member States prior to initial EU market opening.
In 2014, the European Regulators’ Group for Post (ERGP) published its report
on E2E competition and access in European postal markets. This report provided
both an overview of competition developments in all EU countries and an exam-
ination of the development of E2E competition, looking for common factors. While
identifying some possible influencing factors, this report concluded that few pat-
terns emerged and that the development of E2E competition in certain countries
(and not in others) suggested rather that country or time-specific factors may have
been more decisive (ERGP 2014a).
This chapter seeks to build on that work by examining different models of E2E
competition in selected countries with already established E2E competition, includ-
ing enabling or hindering factors, similarities, and differences. It is the result of
desk research, a review of literature and questionnaires and interviews with key
actors, e.g., regulators, incumbents, and/or competitors. The chapter examines only
those countries, where E2E competitors became established players (Germany,
Italy, the Netherlands, Poland, Spain, and Sweden), plus the UK, where it seemed
E2E competition was emerging in 2014, and New Zealand, which although outside
the EU, has had full market opening since 1998 (Appendix).
The chapter considers common themes potentially impacting E2E competition
including the liberalization process, access to the incumbent’s postal network
P. Groves (*) • S. Cape

Ofcom, London, UK
e-mail: Philip.groves@ofcom.org.uk

122 P. Groves and S. Cape
and/or infrastructure, preexisting E2E competition, the entry strategies of new

entrants, incumbent responses, and overall volume growth or decline.
The remainder of this chapter is structured as follows. Section 2 presents the case
studies of E2E competition, briefly assessing each in light of the above factors.
Each mini case study highlights how competition emerged, what E2E competition
exists now, and related issues including factors helping or hindering it. Section 3
draws tentative conclusions from the country analysis to highlight similarities and
differences between them. The appendix is a summary table which illustrates the
heterogeneity of the countries considered along side similarities for factors poten-
tially related to E2E competition.
2 Country Mini Case Studies
2.1 Germany
Liberalization in Germany began in 1989 when some items were removed from the
reserved area. Gradual liberalization continued until 2008 when the market was
fully liberalized. By 2008, access to the incumbent’s network, PO boxes, and
change of address information was made available. Small-scale competition has
existed in Germany since the early 1990s, mainly in delivering publications and
direct mail. Two significant competitors emerged by the 2000s, PIN Group AG and
EP Europost (later TNT Post and then renamed Postcon) who pursued a strategy of
acquisition and collaboration with other letter mail operators to extend their cov-
erage. PIN Group AG went bankrupt following the decision to introduce minimum
wages for the postal industry in 2008. This left one competitor with significant
national coverage. Competition was focused on price and a range of service
specifications.
E2E competition now represents around 10 % of letter volumes, with Postcon
offering business and advertising mail products. Other smaller operators increas-
ingly collaborate with each other to extend their coverage. Around 11 % of letter
volumes go through access agreements. Many of the companies using access to
Deutsche Post’s network are consolidators or end-to-end competitors using access
to deliver to areas not covered by them, which was facilitated by the regulator.
Germany has no designated Universal Service Provider, based on an assessment
that market forces can provide the universal service, supported by the high national
quality of service provided, exceeding regulatory targets.
As competition developed in Germany, numerous disputes arose. Ecorys (2005)
noted that disputes around network access and elements of postal infrastructure,
including PO boxes, change of address information, and the redirection of mail, had
been adjudicated by the NRA. The NRA also dealt with a dispute relating to a
subsidiary of Deutsche Post, First Mail, in 2011—concluding that its pricing
policies were discriminatory and anticompetitive. First Mail was obliged to set
prices no lower than the prices for Deutsche Post’s downstream access and
Case Studies in End-to-End Delivery Competition 123
subsequently ceased operations in 2011. The diverse nature of competition—both

end to end and access—may have been helped by the fact that mail volumes in
Germany were less affected by the financial crisis and e-substitution trends than in
many other countries, declining by 1.7 % a year in recent years (WIK 2013)
(Appendix).
2.2 Italy
Italy experienced major volume reductions in recent years due to accelerated

e-substitution and economic crisis. Between 2007 and 2013, postal volumes in
Italy decreased by 6 % per year while revenues declined by 5 % annually
(Appendix). Nevertheless, it has seen entry. Before 2007, TNT had a strong
express mail presence and operated registered mail services. TNT (rebranded as
Nexive in 2014) launched its “Formula Certa” service in 2007 based on a strategy
of offering value-added letter services which were guaranteed and tracked. The
commercial rationale was the need for senders, especially government agencies
and utility companies, to be sure that their communications, especially bills,
had been received by a given date and time. Most of such mail was formerly
unregistered bulk mail.
Since 2007, TNT progressively expanded its service to cover three quarters of
the national territory and 24 million households. Nexive offers a B2C proposition
with a D + 3 quality standard for mail collected and delivered within a single
metropolitan area and D + 5 for mail delivered to a neighboring metropolitan area
(both mail streams currently achieving 97 % quality of service) with compensation
payable for the 3 % of mail which does not hit the required standards. Nexive
performs little manual sortation due to its investment in walk sequencing machines
and delivery is via scooters, bikes, vans, and on foot. Only 25 % of its employees
are direct employees with the remainder employed through partner agencies.
Nexive’s more flexible model enables it to offer prices at 20 % below Poste
Italiane’s VAT exempt price. It also offers other mail-related services such as
database optimization, printing, and enveloping and returns management.
Poste Italiane responded to E2E competition by taking steps to modernize its
mail network and diversify into new business areas. It reached an 85 % level of mail
sorting automation and upgraded its transport networks, offering new tracked,
guaranteed services, and digital services. It also developed “Posta Time,” a product
it positioned as an alternative to Nexive’s Formula Certa. TNT/Nexive brought
various competition complaints to the Italian Anti-Trust Authority regarding Poste
Italiane’s VAT’s exemption, the Posta Time product, alleged predatory pricing, and
alleged interference by Poste Italiane with the mail of its clients to then target them
with Poste Italiane’s competing offers. The Anti-Trust Authority found in Nexive’s
favor regarding the complaint about “Posta Time.” However, the Administrative
Court later reversed this decision finding that Poste Italiane was not dominant in
that value-added market as its product had fewer features than Nexive’s and finding
that the Anti-Trust’s analysis of Poste Italiane’s costs was flawed. On the VAT
issue, the Anti-Trust Authority decided to align national law with the European
Court of Justice Decision in September 2013 and Poste Italiane must now charge
VAT to customers with individually negotiated contracts.
Other reported competition issues related to elements of postal infrastructure
such as control of postal franking machines, the handling of competitor mail
mistakenly posted into Poste Italiane’s network, changes to the postcode system,
and the maintenance of a reserved area for fines and judicial acts. These are still
under discussion.
2.3 The Netherlands
Distribution of publications in the Netherlands was always open to competition.

Since 1989, “identical items” and letter mail above 500 g were excluded from the
monopoly. The reserved limit was reduced to 100 g in 2000. In 2001, access to the
incumbent’s PO boxes was regulated, with the prices and terms decided through
negotiations between the incumbent and competitors.
In 1989, Publisher VNU Group’s Medianet was already delivering its own
consumer magazines and other periodicals through its own delivery network.
Other smaller companies delivered advertising mail regionally. By 2005, there
were a number of competitors delivering publications and direct mail. The two
largest competitors, Sandd and Selekt Mail (a subsidiary of Deutsche Post) had
nationwide coverage and around 4 % of total mail volumes (Ecorys 2005).
Sandd and Selekt Mail initially won customers on price, offering delivery of
pre-sorted bulk mail on 2–3 days each week for substantially less than the incum-
bent using a low cost model, of delivery 2 days a week, by temporary and part-time
delivery staff paid at a lower hourly rate (van der Lijn and Meijer 2003). TNT
reacted by lowering its bulk mail prices and offering longer term fixed rate
contracts. The incumbent also created a subsidiary, Netwerk VSP, to offer a low
cost bulk mail service, delivering once a week and primarily handling unaddressed
and direct mail (ITA Consulting and WIK 2009).
Consolidation was a feature of the Dutch letters market. TNT (later PostNL)
made acquisitions to support its position in the market including Euro Mail, a mail
preparation company, in 2005, and Sandd acquired Selekt Mail in 2011. At the
time, the Dutch competition authority said that regardless of the takeover, Selekt
Mail would exit the market and that it envisaged no more than two national
operators. (NMa 2011) Mail volume decline in the Netherlands was significant
in recent years (around 5 % per annum) which may have encouraged market
consolidation (Appendix). Sandd is now the only national E2E competitor and
retains 20–25 % of the business mail market, delivering bulk transactional mail,
addressed advertising mail and periodicals twice a week. There are also a number of
smaller postal companies, some of which offer a 5-day-a-week service in certain
regions.
These regional operators may be able to provide nationwide services in future.

The Authority for Consumers and Markets (ACM)1 published a draft decision that
Post NL must provide regional postal firms with nondiscriminatory access to its
network of delivery points and sorting centers without favoring the mail it has
collected itself. The access service Post NL offers needs to mirror every 24 h bulk
mail service which Post NL offers its retail business customers. ACM established
that, with a market share of 80–90 %, Post NL has a dominant position in the
business mail market. It further noted that for some businesses Post NL is the only
operator with which they are able to do business and that Sandd and regional
operators are dependent on Post NL to offer a national service to their business
customers.
2.4 New Zealand
The New Zealand market has been fully liberalized since 1998 and DX Mail is the
only major E2E competitor, with a 3 % volume share. DX Mail’s 6-day urban
delivery network uses NZ Post’s network for delivery to suburban and rural
networks. It has a next day intra-city delivery standard and a 2–3 day standard for
national mail. It has a retail presence through its sister company’s express delivery
branches2 in major cities and 13 sorting centers around the country. It competes
primarily on price. A new entrant, Whitestone Post, based in Oamuru, also recently
emerged, operating in a single city and looking to compete on speed and price for
intra-city mail. Its offer may be a local response to the reduced speed of delivery
and the closure of local sorting offices by NZ Post.
There are 24 postal operators in New Zealand, with most of the competition
based on access. Access competition is around 20 % of the letters market by
volume. There are two types of access agreement in New Zealand. One is a prepaid
service allowing access to NZ Post pillar boxes, such as that used by Pete’s Post
(among others). This is a form of resale of the NZ Post service, where the customers
of competitors to NZ Post can send prepaid envelopes and NZ Post recognized
stamped mail in NZ Post pillar boxes. The other is known as a “lodgment access”
agreement where an access operator can issue its own stamps, collect its customers
mail directly and then hand it to NZ Post for outward sorting and delivery.
Although not wholly due to competition, NZ Post faced reduced letter volumes,
particularly full-rate letters, for several years. It proposed raising prices to com-
pensate for revenue loss. While it was able to raise retail prices in the absence of
price controls, its proposals to increase access prices prompted several complaints
1
In January 2014, ACM took over the regulation of postal services in the Netherlands—its main
task is to ensure compliance with the 2009 Postal Act.
2
DX Mail is part of the Freightways Group, which includes a number of express delivery and
document management companies.
to the Commerce Commission, New Zealand’s competition authority, and were

subsequently abandoned. Between 2010 and 2014, prices for full-rate letters
increased by 60 %, while bulk mail prices increased by only around 12 %.
Unlike the EU countries that we have assessed in this chapter, the USO in
New Zealand is not written into legislation. Instead, it is agreed on a contractual
basis between NZ Post and the government, and is known as “The Deed of
Understanding” (OIG 2014). To deal with the decline in volumes, a new deed for
the USO was negotiated in 2013 on delivery frequency which came into effect in
July 2015. NZ Post is now required to deliver to urban addresses not less than 3 days
a week, and not less than 5 days a week to rural addresses and PO boxes. NZ Post
could have asked for more general regulatory relief (such as a reimposition of a
monopoly) or relief in access regulation (Clarke 2014) but it instead asked for a
USO reduction. This might indicate that it does not feel unduly threatened by
current competition, whether access or E2E. The introduction of alternate day
delivery services in urban areas may present a commercial opportunity for DX
Mail given that it delivers 6 days a week.
2.5 Poland
Full market opening took place on 1 January 2013 preceded by the gradual
emergence of competition. Around 250 postal operators already existed in 2011
which had risen to 295 operators by 2014. There are currently four competitors:
Inpost, Speedmail, Distribution Poland, and ABC Direct3 (Appendix).
Inpost, a letters and parcels delivery company, acquired its former partner PGP
in January 2015. It was until recently hopeful of becoming the designated USO
operator from 2016 onwards but recently lost out to the designated operator Poczta
Polska, pending an appeal. Speedmail has an established network enabling it to
reach 75 % of postcodes. It has a strategic focus on quality, price, flexibility,
reporting and track and trace, offering standard, and value-added mail and mail-
related services.
The Polish postal market is stable both in terms of revenues generated by postal
operators and volumes delivered. Addressed mail remains dominated by Poczta
Polska. Competitors gained certain non-USO contracts for bulk mail, especially
transactional mail from utility companies. Key tenders for delivery of judicial items
and administrative mail, including for Poland’s courts and prosecutors’ office, were
won by E2E competitors, sometimes acting in cooperation with each other.
Operators now look to diversify their businesses—Inpost recently looked into
the option of buying the mailing house Inforsys, which provides direct mail, hybrid
mail, and e-commerce services. Poczta Polska finds it difficult to compete on price,
3
Distribution Poland and ABC Direct are mainly active in unaddressed delivery.
given new entrants’ increased flexibility and cheaper labor costs, including tempo-
rary contracts and outsourcing.
Access to all the main elements of the incumbent’s postal infrastructure is
available. While the interested parties negotiated such access, occasional issues
remain. In April 2015, Poczta Polska signed an agreement granting access to
roadside mailboxes used by rural communities. In addition, the incumbent has
agreed on access to other parts of its infrastructure with seven other postal opera-
tors. But there is no postal network access.
2.6 Spain
Alternative E2E competitors are still operating mainly locally or regionally and in
the niche market of business bulk mail including direct mail in Spain. Mail volumes
per capita are low (87 items) (Appendix). They are seen as successful in those
markets albeit with low margins. Traditionally, domestic intra-city mail and direct
mail were not part of the reserved area in Spain. A large proportion of mail is
generated in the two major cities, Madrid and Barcelona. Alternative E2E compet-
itors rely on access to Correos’ network in areas, where they do not deliver.
Correos’ standard delivery is a D + 3 service although in 2013 it delivered most
letters sooner than D + 3.
There are two main alternative E2E competitors in Spain: Unipost, which pro-
vides E2E services mainly in Madrid and Barcelona, and Akropost in the Basque
country. They both provide full national and international coverage through access
to the incumbent’s network and, in the case of Unipost, a partnership with DHL
Global Mail. They also provide added value services to their customers (for
example, planning direct mail campaigns). In 2013, Correos retained 87 % of
mail volumes compared with 11 % for Unipost (Akropost is much smaller), and
92 % of revenues (compared with 7 % for Unipost). However, the Unipost figures
include its access mail.
Correos reacted to competition by denying the access of administrative notifi-
cations handled by competitors to its network—resulting in a competition case, and
sanction, against Correos.4 It also offered higher access prices to alternative E2E
competitors than those offered to its other large clients—conduct sanctioned in an
antitrust margin squeeze decision. Correos subsequently submitted an access model
contract, yet to be analyzed by the regulator CNMC, which it claims will be equally
offered to alternative E2E competitors and the remainder of its big clients.5 Correos
allegedly retained items stamped by alternative operators found in its network, and
a complaint is with the regulator.
4
Decision of 22 April 2011 (S/034/11 Correos).
5
Decision of 21 January 2014 (S/0373/11 Correos 2).
2.7 Sweden
The Swedish letters market was fully opened up to competition in 1993 (Appendix).
The early years were marked by protracted legal disputes between Citymail (later
renamed Bring Citymail AG) and the incumbent Sweden Post (later renamed
PostNord). One of the issues was the time it took to get final regulatory decisions
due mainly to lengthy appeals in the market courts. It took over a decade for
Sweden Post to abandon a pricing structure which Bring Citymail felt targeted its
delivery operations on non-cost-based grounds in order to try to prevent its viabil-
ity. Another issue for Bring Citymail was and remains the perceived lack of
transparency of PostNord’s pricing structure. PTS recently made a ruling to urge
that PostNord publish its tariffs in full which PostNord challenged in the courts
and lost.
Bring Citymail AG, owned by Post Norway, is the main competitor and is
mainly active in the bulk “industrialized mail” segment. It covers around 54 % of
delivery points with 13.4 % of delivered volumes or 6.6 % by value (PTS 2015).
There are also around 30 mostly small-scale local E2E operators. Despite an overall
decline in letter volumes of 29 % between 2000 and 2014, competitors to the
incumbent increased their market share to just over 15 % of total mail volumes
by 2014 (PTS 2015) The corresponding revenue share is 7.7 %, reflecting the low
value nature of the bulk mail they are delivering.
Bring Citymail mainly competes in the market for pre-sorted economy class
bulk mail. PostNord has a de facto monopoly on overnight nationwide bulk mail
due to the need for a costly preexisting network to provide it. Bring Citymail’s
business model is based on customers producing and pre-sorting mostly bills,
statements, magazines, and advertising mail at the printing stage to the first four
of five postcode digits. This system enables the final walk sortation to be carried
out quickly by hand also removing the need for expensive walk sequencing
machines.
Bring Citymail has two zones based mainly on the delivery point density with
savings on PostNord’s prices of around 35 % per letter item. Delivery is on a
defined day every third day and each delivery zone is split into three sections A, B,
and C with one third of the mail being delivered to each section of the zone
depending on the delivery day. In the submarket of economy bulk mail, PTS
(2015) observes that Bring Citymail shows clear growth in volume (3.7 %) in a
generally declining market. Bring Citymail has carried out cost rationalization
initiatives and expanded its offer to customers recently, including its merger with
a newspaper distribution company in Stockholm, which is helping it to improve its
financial situation.
Bring Citymail’s staffing model is similar to PostNord’s in that the same union
represents its employees and joint negotiations take place on terms and conditions.
Most of its staff are full-time employees with some part timers. Sweden has some of
the lowest bulk mail prices in the European Union, for example, 20 euro cents for a
20 g bulk mail non-priority item. (ERGP 2014b) Partly as a result, e-substitution has
been significantly less in Sweden than in Denmark, though Denmark experienced a

strong Government drive to e-billing not experienced in Sweden.
Competition enabled both PostNord and Bring Citymail to become more sales
and customer focused. PostNord reduced the cost of its counters outlets, changed its
relations with the union, automated its mail centers and reformed its delivery
routes, including separating company and residential deliveries where possible.
PostNord reduced its delivery staff numbers and now makes a positive return on
providing the USO. Bring Citymail expanded into home delivery of packets/small
parcels. Regulation of infrastructure issues was essential to Bring Citymail’s
delivery operation especially in setting up a joint management with PostNord for
a change of address database. This agreement was reached voluntarily but with the
regulator ready to intervene. However, no network access exists.
Any further substantial letters volume growth by Bring Citymail is likely partly
to depend on its ability to expand its geographic area of activity. Yet that is difficult
because of the different forms of volume discounts offered by PostNord (up to 15 %
for the highest volumes, though formerly up to 25 %) which helps it to retain larger
contracts. Further expansion may also be possible through Bring Citymail attracting
higher volumes by diversifying further into the (mainly unsorted) office mail
market. The costs would be higher here and require new investment in sortation
and collection.
2.8 The United Kingdom
The letters market in the UK began to be opened to competition in 2000, at a time

when UK mail volumes were relatively high and still growing (Appendix), when a
licensing regime was introduced. Courier services, document exchange, and out-
going international mail were excluded from the license requirement, reflecting
existing competition. Licenses to competitors in the reserved area were issued
subject to maintenance of the USO and were short term, specific, and included
restrictions on volumes.6
Three types of licenses were introduced in 2003 and the reserved area was
reduced from 1 January 2003 to 100 g, in line with the Postal Directive. Licenses
were available for carrying bulk mail—single mailings of 4000 identical format
items from a single UK site, for consolidation activities, for collection and sorting
mail, and for defined, specific activities, within the reserved area. On 1 January
2006, the market was fully opened. Access mail is the predominant form of
competition. Royal Mail’s license included a condition which required it to nego-
tiate in good faith to agree terms for access. The first negotiation, between Royal
Mail and UK Mail, broke down and UK Mail requested a determination from
6
For example, Deutsche Post Global Mail (UK) Ltd was permitted to carry 40 million items per
year. Express Dairies Ltd was permitted to carry 4.6 million items per year.
Postcomm. However, its proposed notice was unacceptable to both parties who
instead came to an agreement, leading to downstream access competition in 2004.
This agreement provided the model for other access agreements.
The introduction of downstream access probably discouraged any significant
E2E competition despite earlier formal market opening. The liberalization of bulk
mail did not lead to an increase in delivery competition, and WIK noted that
mailings of 4000 items or more would be addressed to consumers and would
“require a well-developed delivery network covering significant portions of the
country” (WIK 2006). Prices for access mail were also low, reflecting the lower
costs of access operators, and helped by Postcomm’s price control mandating the
price differential between the bulk mail retail and wholesale prices. Until recently,
it looked as though significant E2E competition in the UK could still emerge despite
extensive access volumes. Whistl (formerly TNT Post, a subsidiary of the Dutch
incumbent Post NL, and a major access operator) began to roll out an E2E network
in urban areas after trialing deliveries in West London in April 2012. Whistl
intended to cover 42 % of UK addresses (8.5 % of the UK by area) by 2017,
using a 2–3 day delivery model and employing 20,000 postal workers (CWU 2014).
It handled business to consumer mail, such as bills, statements, and direct mail as
well as some smaller parcels.
Whistl rolled out its network to deliver in areas of London, Manchester, and
Liverpool, covering two million addresses, or 7.4 % of UK households (Post NL,
2015a), and for 2013–14 had a market share by volume of 0.5 % of the letters
market (Ofcom 2014a). However, in May 2015, Whistl suspended its E2E opera-
tions after Lloyds Development Capital (LDC), its private equity partner decided
against investing in Whistl. LDC stated that this was due to “ongoing changes in
UK postal market dynamics and the complexity of the regulatory landscape”
(Post NL, 2015b). Whistl then announced in June 2015 that it would cease these
operations entirely. It remains a significant access provider. There are also smaller,
local E2E operators in the UK, primarily serving local businesses.
In response to Whistl, Royal Mail had introduced changes to its access pricing
and terms in January 2014. These included a 1.2 % price differential between its
price plans that included forecasting of future volumes and those that did not, and
changes to the prices for delivering to its access zones, including large price
reductions in the competitive zones. Royal Mail stated that these differentials
were “an important part of Royal Mail’s commercial response to changing market
conditions, including the expansion of direct delivery [E2E] competition” (Royal
Mail Group 2014).
Following a complaint about these changes by Whistl, Ofcom opened an inves-
tigation into the price changes under competition law. Whistl also suspended its
rollout plan at this time as a result of the announced access prices. Alongside the
competition law investigation, Ofcom began a review of Royal Mail’s access
pricing. Following the withdrawal of Whistl’s E2E operations in June 2015,
Ofcom announced a fundamental review of the regulation of Royal Mail in light
of the loss of significant E2E competition for the incumbent.
3 Assessment of Similarities and Differences Between

Country Case Studies
There are several similarities in the market entry models in the eight mini case
studies. In most cases, there is at least one larger, more established E2E competitor
alongside many small, local operators. Competitors typically deliver pre-sorted
bulk mail on fewer days, using a flexible labor model with fewer full-time staff.
However, there are exceptions to this model, notably Sweden, where the main
competitor has similar labor conditions to the incumbent, and New Zealand where
one competitor delivers 6 days per week.
The incumbents’ responses tended to be similar in targeting the areas served by
the new entrant either through a direct pricing response, on bulk mail, or through
making access terms more attractive in the short term. In some cases, there was a
replication of competitor services, such as Posta Time in Italy or Netwerk VSP in
the Netherlands. In countries where E2E competition emerged, the incumbents
have all modernized their networks to varying degrees, partly in response to E2E
competition and e-substitution. Arrangements to access the incumbent’s postal
infrastructure were needed and have generally required regulatory oversight. Reg-
ulatory uncertainty or lengthy disputes appear to be inhibiting factors to E2E
service expansion, for example, in Sweden.
There are few examples of long standing E2E competition, mainly due to the fact
that the EU regulatory framework evolved only gradually towards full market
opening. Some Member States liberalized to a greater extent than the minimum
substantially before this date, notably Germany, the Netherlands, Spain, Sweden,
and the UK. To date, it appears that the volume market share taken by competitors
in such countries appeared to stabilize at 10–15 % of letter volumes in most cases
(UK and New Zealand are exceptions).
The more enduring E2E competition emerged when volumes were either stable
or increasing. But its fragility can be seen in the market exit of Whistl in the UK,
and the merger of Sandd and Selekt Mail in the Netherlands, where the competition
authority noted that it would be unlikely to see the multiple nationwide postal
operators continue to exist while volumes decline. Overall, it seems probable that
the timing of full EU market opening which accompanied volume declines in most
Member States affected the extent and nature of market entry. Even so, some
competitors have entered the market or expanded while volumes are declining,
for example, Nexive in Italy, or the main competitors in Spain and Poland,
notwithstanding regulatory issues. In all countries, many highly local competitors
do business with very low volumes.
But there are notable differences across countries. E2E competition has occurred
in countries with very different characteristics such as population levels, geogra-
phies, population densities, levels of urbanization, and mail volumes per capita.
This suggests that national circumstances are an important factor, especially in the
following areas:
Competition already existed in Germany, Spain, Sweden, and the Netherlands

(also New Zealand outside the EU) independently of the gradual and controlled
liberalization course adopted by the three Postal Directives. In Sweden, it devel-
oped out of meeting a business need for cheaper, lower specification delivery
services for pre-sorted bulk mail. In Italy, it developed rather out of a need for
tracked day and time certain delivery services. In Spain, Germany, the Netherlands,
and Sweden, preexisting competition in niche areas was able to develop and expand
as the market opened more fully. But there was no set pattern to the emergence of
E2E competition.
In Italy, a perceived gap in the market was exploited by the main E2E competitor
to offer a tracked and guaranteed service with compensation offered for service
failures. This was also the model of competition that emerged in Germany despite
the perceived higher quality of service of its incumbent operator. There, compet-
itive entry was secured through a system of regional licenses for particular types of
mail, which senders were able—after regulatory intervention—to combine with
access to Deutsche Post’s work sharing discounts. However, in other countries,
such as the Netherlands and the UK, E2E competition was not accompanied by
value-added services.
The preexistence of network access or access regulation played a role in market
entry in different ways. The early access agreement between Royal Mail and UK
Mail may have delayed the emergence of E2E competition due to the favorable
access rates negotiated leading to access volumes growing substantially. The
different access regime in New Zealand, allowing competitors to sell preprinted
envelopes and stamps then deposited by the sending customer in New Zealand
Post’s pillar boxes, may also have discouraged more widespread E2E competition.
On the other hand, the emergence of E2E competition on a regional basis in the
Netherlands, Germany, and Spain stimulated regulatory steps to ensure nondiscri-
minatory network access which had not previously existed so challenger operators
could offer national coverage.
Some incumbents such as Post NL tried to reconfigure the universal service
network (for example, consumer access points such as pillar boxes or post office
branches). However, this was possibly more in response to declining volumes than
to E2E competition given the volumes lost in each case. However, it made the
incumbents more competitive domestically. However, in New Zealand, which saw
a major USO modification, the USO changes may have benefited the challenger
operators as NZ Post no longer delivers daily in urban areas, unlike one of its
competitors, and this provides an opportunity for the competitor to differentiate its
services.
In countries where the alternative operator does not deliver daily to each address,
such as Italy, there is the possibility of certainty on the latest day/time of delivery,
enabling better targeting of mail delivery times. If, as is currently envisaged in Italy
to secure the financial sustainability of the USO, daily USO delivery were to cease
for a significant portion of addresses, then this might further increase the attrac-
tiveness of targeted alternative offerings with value-added elements.
To conclude, the historical regulatory situation appears to have been a main

contributory factor to the emergence of E2E competition. Market entry has
occurred in countries where a type of mail (such as publishing or direct mail) was
excluded from the monopoly area prior to more extensive EU market opening, and
where letter mail volumes were growing. Volume declines, although not necessar-
ily low mail volumes to start with, have discouraged entry. Regulatory oversight
and the implementation of different access regimes has also had an impact on the
development of E2E competition, with access to postal infrastructure, such as PO
boxes, encouraging competition, and the existence of attractive network access
terms acting as an inhibiting factor. On the other hand, network access has also
worked as an enabler, allowing competitors to offer a national service while
undertaking deliveries on a smaller scale. Finally, several case studies have indi-
cated the fragility of E2E competition which can result from prolonged regulatory
uncertainty or lengthy legal disputes especially when combined with problematic
behaviour by the incumbent. As volumes continue to decline competitive pressure,
legislative and regulatory changes or the actions of the incumbent can also lead to
market exit of significant competitors.
Acknowledgements The authors would like to thank those in the countries that we studied who
were able to talk with them for their assistance and expertise. Their help was much appreciated.
The views expressed in the chapter are the personal views of the authors, and not those of the
interviewees or of Ofcom.
Appendix: Summary Table of Key Country Indicators

Including Factors Potentially Affecting E2E Competition
The
134
Germany Italy Netherlands New Zealand Poland Spain Sweden UK

Population size/density 81 million/ 62 million/ 17 million/ 4.5 million/ 38 million/ 48 million/ 10 million/ 64 million/
(2014) (CIA 2015) 232 per km2 210 per km2 498 per km2 16 per km2 126 per km2 96 per km2 24 per km2 263 per km2
Urban population 75 % 69 % 90 % 86 % 61 % 79 % 86 % 82 %
(2014) (CIA 2015)
Full market opening 2007 2011 2009 1998 2013 2011 1993 2006
Existing competition Direct mail, Registered Direct mail, Document Document Intra-city Document Document
pre-liberalization catalogue and mail publications exchange exchange mail and exchange exchange
magazine subcontracted direct mail
distribution by incumbent
Trend in mail volume Broadly Slight growth Stability Decline Growth but Decline Decline Slight growth
since 2002 stable followed by followed by slowing followed by
decline decline decline
Average 2002–2007 0.8 % 0.9 % 0% n/a 2.2 % n/a 0.9 % 0.2 %
mail vol- (ITA Con-
ume sulting and
growth WIK 2009)
2007–2001 1.7 % 6.0 % 5.4 % n/a 0.7 % 6.6 % 3.2 % 6.4 %
(WIK
2013)
Mail volume per capita 196.1 items 59.4 items 228.8 items 178 itemsa 47.1 items 87.1 items 265.9 items 230.4 items
(2013) (Ofcom 2014b)
Scope of USO (WIK Single-piece Single-piece Single-piece Single-piece Single-piece Single- Single-piece Single-piece
2013) letters and letters, publi- letters and letters and letters and piece and letters and letters 5 days
parcels, cations, and parcels, parcels, parcels, bulk letters parcels per week,
6 days per parcels, 5 days 5 days per 3 days per 5 days per and parcels 5 days per singlepiece
week per week week week urban week 5 days per week parcels
areas, 5 days week 6 days per
per week week
rural areasb
P. Groves and S. Cape
Incumbent D + 1 QoS 80 %/91.2 % 89 %/90.4 % 95 %/95.8 % n/a 82 %/66.7 %c 93 %/ 85 %/94.9 % 93 %/93.2 %
2013 (target/achieved) 96.6 %d
(ERGP 2014c)
Number of significant 1 significant 1—Nexive 1 nationwide 1—DX Mail 4 (Inpost, 1 (Red 1—(Bring 1 former sig-
e2e competitors competitor operator Speedmail, Unipost) Citymail nificant com-
(Postcon) (Sandd) plus Distribution plus AB) plus petitor
plus multiple regional and Poland, ABC smaller around (Whistl)
regional and local Direct) plus regional 30 others recently
local compet- operators smaller oper- competitors ceased e2e
itors. 1200 ators. services.
postal 274 licensed Some small-
licenses at operators in scale local
end of 2013 2013 operators
Competitors market c.10 % c.10 % c.10–20 % c.3 % c.10 % c.15 % c.13 % c.1 %
share (volume)
Case Studies in End-to-End Delivery Competition
a
2011 figure: calculated from Clarke (2014)
b
As of July 2015, Clarke (2014)
c
Although QoS is below the target, improvement has been made in recent years. The achieved standard in 2010 was 53 %, and in 2011 it was 63 %
d
D+3
135
References
Central Intelligence Agency (CIA) (2015) The world factbook. https://www.cia.gov/library/publi

cations/resources/the-world-factbook/. Accessed 14 June 2015
Clarke R (2014) Legalising alternate day delivery: the reshaping of the universal postal service in
New Zealand. Prepared for Comité européen des régulateurs postaux (CERP), for discussion at
its 49th plenary meeting, 29–30 May 2014
Communication Workers Union (CWU) (2014) CWU submission to Ofcom: the threat to the
universal postal service from direct delivery competition. Citing a presentation from
Marketforce’s 12th annual conference on the future of UK postal services, September 2013
Ecorys (2005) Development of competition in the European postal sector
ERGP (2014a) ERGP Report on end-to-end competition and access in European postal markets
ERGP (2014b) ERGP(14)(23) report on benchmarking of universal tariffs, Fig 3-20
ERGP (2014c) ERGP report 2014 on the quality of service and end-user satisfaction, pp 23–24
ITA Consulting & WIK (2009) The evolution of the European postal market since 1997
NMa (2011) Sandd may take Selektmail, NMa. https://www.acm.nl/nl/publicaties/publicatie/
10750/NMa-Sandd-mag-Selekt-Mail-overnemen/. Accessed 7 July 2015
Ofcom (2014a) Annual monitoring report on the postal market: financial year 2013–14
Ofcom (2014b) International communications market report 2014, p 315
Office of Inspector General, United States Postal Service (OIG) (2014) Guiding principles for a
new universal service obligation: RARC report, p 35
Post NL (2015a) Annual report 2014. http://www.postnl.nl/en/Images/2014-postnl-annual-report_
tcm9-22091.pdf
Post NL (2015b) LDC concludes discussion with PostNL on investment in Whistl UK E2E
operations. http://www.postnl.nl/en/about-postnl/press-news/press-releases/2015/april/ldc-con
cludes-discussion-with-postnl-on-investment-in-whistl-uk-e2e-operations.html
PTS, Svensk Postmarknad (2015)
Royal Mail Group (2014) Ofcom investigation into Access contract changes. http://www.
royalmailgroup.com/ofcom-investigation-access-contract-changes
van der Lijn N, Meijer A (2003) Is mandatory access in the postal sector the key to success? In:
Crew MA, Kleindorfer PR (eds) Competitive transformation of the postal and delivery sector.
Springer, New York
WIK (2006) Main developments in the postal sector (2004–2006), p 184
WIK (2013) Main developments in the postal sector (2010–2013): country reports
Promoting Competition at the Digital Age
with an Application to Belgium
Pierre Copée, Axel Gautier, and Mélanie Lefèvre
1 Introduction
The postal sector has undergone dramatic changes over the recent years under the
double effect of liberalization and increased competition from alternative commu-
nication channels (e-substitution). As a result, the mail volume handled by the
historical operator has declined sharply (Nikali 2008; Fève et al. 2010; Meschi
et al. 2010). In the long run, declining volumes may affect both the extent of postal
competition in the market and the sustainability of the universal postal service
(Crew and Kleindorfer 2005) with the two dimensions being intrinsically linked
(Gautier and Wauthy 2012).
In most of Europe, postal markets have been fully liberalized since 2010.
Alternative postal operators, provided that they satisfy licensing requirements,
can offer products and services without any restriction. Newcomers in the postal
markets have adopted an alternative business model. They target commercial mail
of large senders for which collection and sorting costs are limited. They deliver
mail less frequently than the incumbent operators (usually 2 or 3 days a week) and
they may not cover all the national territory. So far, the development of competition
is unequal among Member States. Some countries (the Netherlands, Sweden,
P. Copée
University of Liege (ULg), HEC Management School, Liege, Belgium
A. Gautier
LCII (ULg) and CORE (UCL), Liege, Belgium
e-mail: agautier@ulg.ac.be
M. Lefèvre (*)
LICOS (KULeuven), KU Leuven, Belgium
e-mail: melanie.lefevre@ulg.ac.be

138 P. Copée et al.
Germany) experience large-scale competition from alternative operators, while

others have minimal competition.
The development of competition—from postal or other digital operators—is
seen as a threat for the sustainability of the universal service. There are two reasons
for that. Firstly, competitors target mainly the profitable market segments, leaving
the less profitable ones to the universal service provider that must serve them as part
of the Universal Service Obligations (USO) (a phenomenon known as “cherry-
picking”). Secondly, there are important economies of scales in mail delivery. If
competition from postal competitors or electronic communication erodes mail
volumes, the average delivery cost increases for a given delivery frequency. For
these reasons, the question of entry cannot be totally separated from the question of
designing and financing the universal service.1
Our objective in this chapter is to develop a model of competition to assess its
extent in a context of declining mail volumes. Our focus will be on the Belgian
market, interesting for two reasons. Firstly, despite a high population density, postal
competition remains limited, compared for instance to the Netherlands, with cur-
rently one active licensed small-scale operator (TBC-POST). Secondly, licensing
requirements imposed on postal operators are strong. Licensed operators have the
obligation to deliver mail twice a week on 80 % of the territory. More precisely they
have to cover 80 % of the territory in each of the three regions of Belgium
(Flanders, Wallonia, and Brussels) 5 years after they begin to operate, according
to the following timeline: first year: 10 %, second year: 20 %, third year: 40 %,
fourth year: 60 %, and fifth year 80 %. These strong licensing requirements may
deter entry.
The first step in the analysis is to assess the extent of unconstrained entry to
evaluate whether or not the licensing obligations act as a constraint. In a second
step, we can evaluate the cost of meeting the licensing obligations. If this cost is
prohibitive, the coverage constraint is a barrier to entry. Last, we can evaluate the
impact of entry on the sustainability of the universal service.2
To that end, we develop a fully fledged model of postal competition. The cost
structure of the postal operators is based on Roy (1999) and the competition model
is based on d’Alcantara and Gautier (2008) and Gautier and Paolini (2011).
Section 2 is devoted to the cost model, while the market model is exposed in
Sect. 3. Simulation results are shown in Sect. 4. We present two scenarios, one with
1
The universal service and its financing are not competitively neutral (Borsenberger et al. 2010;
Gautier and Wauthy 2012), implying that a change in the scope of the USO is likely to have an
impact on the extent of competition on the market.
2
To maintain the sustainability of the universal service, Member States have on the one hand,
safeguarded the USO financing by installing dedicated funding for the USO (compensation fund
and/or state aid) and, on the other hand, reformed the USO itself. The USO reforms concern all the
dimensions of the universal service: the product bundle included in the definition of the USO, the
pricing constraints applied to the universal service providers (uniform pricing, special rates,
commercial freedom) and the definition of the universal service itself (accessibility of contact
points, doorstep delivery, delivery frequency).
Promoting Competition at the Digital Age with an Application to Belgium 139
the current mail volumes, the other where volumes decline by 40 %. Section 5
analyses the impact of the licensing requirements imposed in Belgium. At current
mail volumes, we estimate that the cost of licensing requirements is moderate but if
volumes fall by 40 % this cost becomes prohibitive for the entrant. In the context of
e-substitution, there appears to be room for profitable entry unless the current
licensing requirements are maintained. We also show that competition and
e-substitution have a strong impact on the incumbent’s average delivery cost
which may in turn weaken its ability to sustain the universal service. This is
discussed in Sect. 6. Finally, Sect. 7 concludes.
2 Cost Model
2.1 Cost of Outdoor Delivery Work
We compute the cost of the outdoor delivery work (ODW) following the approach
of Roy (1999), which is based on Jasinski and Steggles (1977) and Cohen and Chu
(1997). The delivery cost is broken down into four categories: travel (active route),
stopping, delivery, and loading. Due to a lack of available data, the cost of travel to
and from the delivery unit to the route is ignored.
Three processes j may be used for delivery: foot, bike, or car. The daily cost3 of
outdoor delivery work on a given area i using process j is given by:
Ci j ¼ wT ij þ v j Li ð1Þ
where w is the hourly salary of the carrier, Tij is the time of the delivery round in
area i using process j, vj is the vehicle cost per kilometer with process j, and Li is the
length of delivery round in area i. Note that vj is assumed to be zero when the
delivery is done by foot or by bike.
The time for the route is measured as:

T ij ¼ si intrastop timeij þ si stoptime j þ Qi ðdelivery timeÞ
Qi
þ ðloading timeÞ ð2Þ
qmax j
The first element corresponds to active route time, that is, the time used by the
postman to travel in the delivery area i where si is the number of stops the delivery
man has to do in the area and (intrastop timeij) is the average time necessary to
travel the distance between two stops in this area using process j. Let speedij be the
speed of vehicle j on area i, and Li, the length of roads in area i, we have:
3
Precisely, Cij is the cost of the outdoor delivery work per delivery day, such that the weekly cost
is Cijd, where d is the delivery frequency per week.
Li =si
intrastop timeij ¼ ð3Þ
speed ij
The second element of (2) is the stopping time. It is the product of the numbers of
stops in the area si and the time spent at each stop using process j (entering a
building, walking up a path, parking the vehicle, etc.).
The third element is related to the delivery (or drop): delivery time is the time
necessary to remove the mail from the bag and posting it through the mailbox.
While it may not be exactly the case in practice, we assume, following Roy (1999),
that this operation generates the same average individual time for each object. Qi is
the number of items delivered in the area i and is obtained by multiplying the
number of items delivered per inhabitant per day q by the population of the area
popi.
The last element of (2) comes for the use of relay boxes when the process used
implies capacity constraints. It is typically the case for deliveries by foot or by bike.
qmax j is the maximum number of items that can be transported at once with process
j, and loading time is the time necessary to remove the mail from the relay box and
putting it into the delivery bag.
The number of stops si in area i depends on the number of delivery points and the
probability of distributing mails at a given delivery point, i.e., the probability of
making a stop at a delivery point. The number of delivery points is measured by the
number of buildings ni in area i. The probability of delivering a mail at a given stop
point depends on the grouping index gi which is the ratio of the number households
hi to the number of buildings: gi ¼ hi =ni (Boldron et al. 2007) and the total mail
volume delivered in the area. We follow Roy (1999) and we model the probability
that a household receives a mail item as a Poisson process. If at least one household
in a building receives a mail, then the delivery man has to make a stop; with a
Qi
Poisson process, this probability is given by exp hi gi and it increases with the
grouping index and the total mail volume. The number of stops is then given by:

Qi
si ¼ ni 1 exp gi ð4Þ
hi
Using data for Belgium, we calculate the daily cost Cij for each of the three possible
delivery processes. Then, we choose the process with the minimum cost. The
chosen process is assumed to be the same for the entire area i while it can differ
from one area to the other. Hence, the daily cost of outdoor delivery work on a given
area i is:
Ci ¼ Min j Cij ð5Þ
Using this cost model, we compute the daily cost of outdoor delivery work for
different values of the delivered items per inhabitant q. We then simulate the effect
of a change of the number of items delivered each day in the area Qi on the daily
cost Ci . Using these simulated costs, we estimate a linear cost function for the
outside delivery work in area i:
e i ¼ α þ βQi þ γXi þ Ei
C ð6Þ
where α and β are parameters to be estimated, γ is a vector of parameters to be

estimated, Xi is a vector of variables representing the characteristics of area
i (density, length of roads, number of households, number of buildings, etc.), and
εi is an error term. The estimated coefficient β^ can be interpreted as the marginal
cost of mail.
2.2 Data
There are no publicly available data on postal routes. We therefore use, as a first
approximation, data on Belgian municipalities as the reference point for defining the
delivery areas i. Belgium is divided into 589 municipalities ranging from 1 to 214 km2.
Most municipalities (88 %) have a density lower than 1000 inhabitants per km2, but
some are densely populated (up to 22,048 inhabitants per km2). For each municipality,
we collected administrative demographic and geographic data from several public
services websites. Descriptive statistics and sources are given in Table 1.
In order to calibrate the cost model, we formulate a number of hypotheses (see
Appendix).
2.3 Cost Estimation
Equation (5) is calibrated for the 589 Belgian municipalities using data and
hypotheses presented in Table 1 and Appendix. For the annual number of items
per capita x, we use the figures provided by WIK (2013) for the year 2011:
Table 1 Descriptive statistics

Variable Mean s.d. Source
Density (inhabitants/km2) 727 1957.44 SPF Economiea
Population (inhabitants) popi 18,257 29,021.9 SPF Economiea
Length of roads (km) Li 198.1 143.53 SPF Mobilitéb
Number of households hi 7832 13791.2 SPF Economiea
Number of buildings ni 6049.6 7734.51 Statbela
Grouping index gi 1.19 0.445 gi ¼ hi/ni
Number of stops si 5577.2 7329.43 Equation (4)
Number of observations: 589
a
Data 2009
b
Data 2013
233 items per capita. We divided it by 5 working days and 52 weeks to obtain the
number of items delivered by inhabitant per day (0.9) that is considered to be the
same in all areas.
Figure 1 shows the results for the unit cost ( Ci =Qi ) per municipality, where
municipalities are presented from the most to the least densely populated. Average
unit cost of outdoor delivery work is 0.147€ per item. From Fig. 1, we can see that
the country is somewhat homogenous with few very densely or sparely populated
municipalities. Indeed, the average delivery cost per quintile is almost identical for
Q2, Q3, and Q4 (¼0.157, 0.164, and 0.174), while it is substantially lower for Q1
(0.118) and higher for Q5 (0.235). Given that municipalities have similar costs in
the three middle quintiles, it is expected that the entrant will either cover all or none
of them (d’Alcantara and Gautier 2008).
We then calibrate this Eq. (5) for the 589 Belgian municipalities allowing
changes in the number of items delivered by inhabitant, using x ¼ {100,
233, 400}. This provides us with 1767 observations from which we can estimate
Eq. (6). Table 2 shows the results of the estimation.
0.6 €
Unit cost in each municipality
Average Cost
0.5 €
Average Cost per quitile
0.4 €
Unit cost (euros)
0.3 €
0.2 €
0.1 €
0.0 €
0.0 0.2 0.4 0.6 0.8 1.0
Percentile of density (high to low)
Fig. 1 Unit cost of outdoor delivery work
Table 2 OLS estimation: Coefficient SE

daily cost of outdoor
Items delivered in the area (Qi) 0.0467*** 0.0004
delivery work
Length of roads (li) 2.3048*** 0.0617
Number of buildings (ni) 0.1580*** 0.0026
Number of households (hi) 0.0139*** 0.0017
Density 0.0364*** 0.0040
Constant 44.5569*** 12.8923
N 1767
R2 0.992
***Significant at 1 % level
This estimation strategy allows us to disentangle the delivery cost in area i into
(a) a marginal cost of outdoor delivery per item and (b) a fixed cost of delivery in
area i that can be reconstructed using the characteristics of area i and the coeffi-
cients of Table 2. The marginal cost is estimated at 0.0467€ per item. The drivers of
the delivery costs have the expected sign. The cost decreases with the density and
increases with the length of roads, the number of buildings and households.
3 Market Analysis
3.1 The Market Game
As a next step, we use the simulated cost model to construct market scenarios. We
consider two scenarios where an entrant (E) competes with the incumbent (I) on the
bulk mail market. In our market simulations, we consider that the entrant is equally
efficient as the incumbent. In particular, we consider that the two operators use the
same delivery cost technology represented by our simulated cost model.
Even if the two operators use the same technology, they have different business
models. The incumbent, which is considered to be the universal service provider,
must deliver standard and bulk mails five days a week. The entrant has a business
model that differs in three respects from the incumbent’s one. First, the entrant
concentrates its operation on the bulk mail market where sorting and collecting
costs are lower. This means that the incumbent remains the unique provider of
traditional mail. Second, the entrant has a lower delivery frequency (d ) of 2 days a
week. Last, the entrant is not committed to cover the whole territory, and it may
cover only the areas i where it is profitable.
To model competition on the bulk mail market, we assume demand functions for
the incumbent and the entrant, derived from the maximization of the utility of a
representative sender defined as:
x2I x2
U ðxI ; xE Þ ¼ αI xI þ αE xE β β E δxI xE ð7Þ
2 2
where xI and xE are the number of (bulk) items sent with the incumbent and the
entrant, respectively. The asymmetry in business models is reflected by the differ-
ence between αI and αE. All else equal the representative sender is willing to pay an
extra αI – αE, for the use of the incumbent’s service.4 Demand levels xI and xE are
expressed as mail received per inhabitant per year. From Eq. (7), we can derive
4
In an attempt to estimate the willingness to pay for different service attributes, Rohr et al. (2013)
found that, for business clients there is a substantial difference between the willingness to pay for a
next-day delivery service compared to a delivery within 2 or 3 days.
demand functions corresponding to a competitive situation (if area i is covered by

both I and E) and a monopolistic situation (if area i is only covered by I).
The timing of the game is the following. The incumbent first sets its price for
bulk and standard mails. Then, the entrant decides on its price and on its area
coverage. With sequential price choices, there is no strategic limitation of the
coverage (Valletti et al. 2002) and the price equilibrium always exists (Gautier
and Wauthy 2010). Given prices ( pI, pE), the mail volume captured by the entrant is
equal to xE( pI, pE) per inhabitant. The entrant’s coverage will consist of all the areas
i where operations are profitable, that is all areas i satisfying:
e i ðQEi Þ 0
π Ei ¼ ð pE cu ÞxE po pi 52dC
where QEi ¼ xE52d

po pi
is the number of items delivered by the entrant on area i per
delivery day and cu the upstream cost of mails (collection, transport, and sorting).
3.2 Calibration of the Model
To calibrate these demand functions, our starting point is a reference monopoly

scenario in which the incumbent operator is the unique postal operator. In this
monopolistic case, the price of bulk mail is pI ¼ 0.5€ and at this price, the bulk mail
volume is equal to 200 items per inhabitant per year. The price of the standard mail
is 0.72€ and, at this price, the mail volume is equal to 33 mail per inhabitant per
year. The prices used are representative of the ones currently observed in Belgium
and the repartition between B2X and C2X mailed items roughly corresponds to
the figures provided by Copenhagen Economics (2010): 88 % for B2X and 12 %
for C2X.
We then consider that the entrant would capture 10 % of the market for bulk mail
if it applies the same price as the incumbent. It would capture 50 % of the market
if it offers a 20 % discount with respect to the incumbent’s price. We suppose
a direct price elasticity of 0.4 and a displacement ratio of 0.9. This last ratio is
σ ¼ ðdxI =d pE Þ=ðdxE =d pE Þ and indicates that 90 % of the items sent by the
entrant are captured from the incumbent. These parameters are similar to the ones
used by De Donder et al. (2006) and d’Alcantara and Gautier (2008).
The demand functions addressed to the two firms are not totally symmetric as, at
equal prices, the incumbent keeps a higher market share than the entrant. To have
equal market shares, the entrant should offer a 20 % discount. This specification
aims at reflecting the fact that the two firms have a different business model, in
particular their coverage and delivery frequency differ. For that reason, even if the
entrant offers a better rate than the incumbent, all the consumers do not immedi-
ately switch to the entrant. In our simulation, the cross-price elasticity evaluated at
pI ¼ pE ¼ 0.5€ is equal to 0.72. This cross-price elasticity is in our specification of
the demand function independent of the entrant’s coverage, which is of course a
restriction of the demand model.
The upstream cost cu is set to 0.20€ for bulk mail and 0.25€ for standard mail,
both for the incumbent and the entrant. Other costs (e.g., post offices) are not
included in our modeling. For this reason, we will focus mainly in the change in
the incumbent’s profit after entry and the absolute value for the profit should be
interpreted with caution as it does not include omitted cost.
4 Simulation Results
We consider two scenarios. In the first, the demands are calibrated to corresponds to
actual mail volumes (200 bulk items at a price pI ¼ 0.5 for a monopolistic incum-
bent) and the entrant delivers mails twice a week on part of the territory. In this
scenario, a hypothetical monopolistic incumbent realizes a profit equal to 459.05
millions €, and it has an average cost of ODW equal to 0.141€. In the second
scenario, the mail volume declines by 40 % compared to the current level.
4.1 Market Simulations: Current Mail Volume
The results for the current mail volume scenario are presented in Table 3. In the first
column, we report the results assuming that the incumbent continues to charge the
current monopoly price after entry; in the second column, the incumbent sets the
Table 3 Market simulations, current mail volume

Passive incumbent Profit-maximizing incumbent
Incumbent’s price ( pI) 0.5 0.45
Entrant’s price ( pE) 0.4 0.383
Municipalities covered (%) 84.38 65.87
Bulk volume for incumbent (xI) 105 131
Bulk volume for entrant (xE) 105 86
Incumbent’s average cost of ODW (€) 0.200 0.173
Entrant’s average cost of ODW (€) 0.125 0.137
Incumbent’s profit (106 €) 212.80 220.70
Entrant’s profit (106 €) 81.10 36.95
Territory covered (%) (Belgium) 73.41 54.17
Territory covered (%) (Brussels) 100 100
Territory covered (%) (Flanders) 95.48 83.34
Territory covered (%) (Wallonia) 52.67 26.64
Cost of licensing conditions (106 €) 1.02 2.81
Cost of licensing conditions (% of π E) 1.25 7.60
Note: Incumbent’s price in the monopoly case is 0.5 for bulk mail and 0.72 for standard mail;
volume (number of items delivered per capita per year in the monopoly case) is 200 for bulk mail
and 33 for standard mail; upstream cost is 0.20 for bulk mail and 0.25 for standard mail; price
elasticity is 0.4; displacement ratio is 0.9; the number of deliveries per week (d ) is 5 for the
incumbent and 2 for the entrant
profit-maximizing prices. For each simulation, we report prices, volumes, coverage

by the entrant, profits, and the average costs of the ODW.
The first thing that should be noticed is the extent of the entrant’s territorial
coverage. At current prices, the entrant would cover 84 % of the municipalities and
66 % if the incumbent sets the profit-maximizing price assuming entry will occur.
The homogenous nature of the country with a lot of municipalities with similar cost
conditions explains the importance of entry.
After entry, the incumbent optimally decreases its price by 10 %, with its
corresponding market share for the bulk mail being equal to 65 %. This large
market share is explained by the facts that, in the municipalities where both firms
deliver, the incumbent market share remains substantial (60 %) and the incumbent
is the only firm delivering mail in the non-covered municipalities.
Entry has a significant impact on the incumbent’s profit. Compared to the
monopolistic situation, the profit declines by approximately 52 %. This sharp
decrease in the profit results from both the business-stealing effect of entry and
the increase in the average cost of delivery due to lost scale economies. The
incumbent’s average cost of ODW cost increases from 0.141 in the monopoly
situation to 0.173€. Due to its lower delivery frequency and its selective entry in
the least costly areas, the entrant has a lower average cost of ODW than the
incumbent. For this reason, our simulations show that there is room for profitable
entry and that a profit-maximizing entrant will limit its territorial coverage, with
only 65 % of the municipalities covered mainly in Brussels and in Flanders.
4.2 Market Simulation: Declining Mail Volume
In the second scenario, we consider a decrease in the mail volume compared to the
current situation. More in particular, we suppose that at the current incumbent price,
both the bulk and the standard mail volume drop by 40 %. This decrease in volume
is in our view a reasonable scenario to capture the long-term impact of
e-substitution. Except for this volume change, the other parameters used to calibrate
the demand functions are similar.
With declining mail volumes, the coverage of the entrant falls drastically: it is
only profitable to cover 7.5 % of the municipalities. Notice also that, despite a
limited coverage, market competition remains intense as the price charged by the
incumbent is similar to the above case. The reason is the homogenous nature of the
country. At a higher price, the entrant would have a much larger coverage, as it is
illustrated by the passive incumbent case in the first column of Table 4: without a
price reaction, the entrant would cover more than 40 % of the municipalities. The
optimal price set by the incumbent can be seen as a limit price that prevents larger
scale entry by the challenger. Thus, even if the incumbent’s market share remains
important (91 %), the market can be considered as very competitive. The incumbent
profit falls drastically (55 % compared to the first scenario) as a result of the
Table 4 Market simulations, declining mail volume

Passive incumbent Profit-maximizing incumbent
Incumbent’s price ( pI) 0.5 0.45
Entrant’s price ( pE) 0.4 0.383
Municipalities covered (%) 43.97 7.47
Bulk volume for incumbent (xI) 63 79
Bulk volume for entrant (xE) 63 51
Incumbent’s average cost of ODW (€) 0.267 0.214
Entrant’s average cost of ODW (€) 0.160 0.131
Incumbent’s profit (106 €) 71.32 97.46
Entrant’s profit (106 €) 19.34 5.75
Territory covered (%) (Belgium) 33.73 3.51
Territory covered (%) (Brussels) 100 100
Territory covered (%) (Flanders) 51.97 5.16
Territory covered (%) (Wallonia) 16.14 0.99
Cost of licensing conditions (106 €) 5.72 19.64
Cost of licensing conditions (% of π E) 29.59 341.66
Note: Incumbent’s price in the monopoly case is 0.5 for bulk mail and 0.72 for standard mail;
volume (number of items delivered per capita per year in the monopoly case) is 120 for bulk mail
and 20 for standard mail; upstream cost is 0.20 for bulk mail and 0.25 for standard mail; price
elasticity is 0.4; displacement ratio is 0.9; the number of deliveries per week (d ) is 5 for the
incumbent and 2 for the entrant
decline in mail volume and the increase in the average cost of ODW. With declining
mail volume, there is still room for profitable entry but at a much lower scale than at
the current mail volume.
5 Licensing Requirements
Belgium has quite unusual and substantive licensing obligations including, (i) the
obligation to deliver mails twice a week after 2 years of operations, (ii) territorial
coverage constraints with the obligation to cover 80 % of the territory in all three
regions after 5 years of operations, and (iii) the use of a uniform tariff. Currently,
there is only one licensed operator who started its operation in May 2013, and it has
a limited market share (~1 %). Belgium strong licensing obligations have been
criticized. The European Commission started an infringement procedure against
Belgium for imposing licensing conditions that are non-necessary and not justified.
A recent WIK report (WIK 2015) considers the licensing conditions to be a barrier
to entry and suggested to remove them. Following these critics, the Belgian postal
law is expected to be revised soon, particularly on this point.
Using our market scenarios, we can evaluate whether the entrant meets the
licensing requirements and, if not, the cost of matching them. In Tables 3 and 4,
we report the coverage in terms of both the percentage of municipalities covered

and the percentage of the territory covered, this last number being split by region.
Coverage constraints in the licensing obligations are expressed in terms of percent-
age of the territory with the 80 % threshold to be reach after 5 years in all the three
regions.
If the entrant fails to meet this coverage constraint (which is the case), we
evaluate the cost of meeting the licensing conditions, using the following algo-
rithm: the entrant has to cover loss-making municipalities till it reaches the 80 %
coverage constraint in the three regions, and it will cover firstly the municipalities
where the losses are the smallest.5 The cost of the licensing obligations is then
evaluated as the lost profit on the municipalities that the entrant is constrained to
serve. If, as a result, the entrant’s profit becomes negative, it is not possible to enter
profitably the market and to meet the licensing constraints. In this case, the license
is a barrier to entry.
At current mail volumes, the coverage of the Walloon region (the less densely
populated) is only 26 %, far below the 80 % obligations while the threshold is
reached in the other two regions. Increasing coverage in Wallonia costs 2.81 million
€ to the entrant, representing 7.6 % of its profit. Though substantial it cannot be
considered as a barrier to entry. Things are radically different if volumes decline. In
this case, the coverage constraint is not satisfied in Flanders (5 %) and in Wallonia
(1 %). We estimate the cost of reaching the 80 % threshold to 19.64 million €. This
amount is larger than the unconstrained profit, which means that with declining
volumes, the license constitutes a barrier to entry. We therefore conclude that a
decrease in the mail volume combined with the licensing requirements do not leave
enough space for the development of competition in the market.
6 Cost of Universal Service
Our model does not allow us to evaluate the cost of the universal service as this
would require the specification of a counterfactual scenario without USO. How-
ever, the model allows us to evaluate the joint impact of competition and
e-substitution on the incumbent’s cost of ODW, and this cost is substantial (see
Fig. 2). If the incumbent is the designated universal service provider and, as such, it
must deliver mails 5 days a week, it cannot compensate decreasing volumes by a
lower delivery frequency. It is therefore interesting to evaluate and to decompose
the evolution of the incumbent’s average cost of ODW. In the monopoly case, this
cost is equal to 0.141€; it increases to 0.173€ after entry and to 0.214€ if we
5
Notice that this does not preclude that there are cheapest way to satisfy the coverage constraint.
0.25
Incumbent Entrant
0.2
0.15
0.1
0.05
0
Monopoly Duopoly, Duopoly,
current volume reduced volume
Fig. 2 Average cost of outdoor delivery work. Note: Incumbent’s price in the monopoly case is
0.5 for bulk mail and 0.72 for standard mail; duopoly results and hypotheses are detailed in Table 3
column 2 for current volume and in Table 4 column 2 for reduced volume
combine entry and declining mail volumes. The total increase is thus equal to 0.073
€ (+52 %) which can be decomposed into an increased cost due to entry (0.032€,
+23 %) and e-substitution (0.041€, +29 %). It is then not a surprise that in a context
of declining volumes, the definition of the USO is under question (Gautier and
Poudou 2015).
7 Concluding Remarks
In this chapter, we developed a fully fledge model of competition on the bulk mail
market to evaluate the impact of licensing requirements and declining volume on
the extent of market competition and the cost of the universal service. The model is
constructed using publicly available data and could be refined further with more
precise data at the postal route level.
Our estimations show that there is room for profitable entry in the market. But
for that, the strong licensing requirements should be removed if volumes are
expected to decline further in the future. We considered a scenario where bulk
mail volume decline by 40 % compared to actual level and, in this case, the
licensing conditions are a barrier to entry. Without coverage constraints, the market
could become quite competitive even if the entrant has a limited territorial cover-
age. This is undoubtedly associated with the homogenous nature of the country with
a significant proportion of the municipalities characterized by a similar cost of
ODW, implying that a change in the market conditions could have an important
impact on the extent of entry.
Finally, we show that the combination of e-substitution and entry increases the
incumbent’s cost of ODW by 52 % because economies of scales in the delivery
activity can no longer be exploited. This figure is a threat for the future of the
universal service, and it would be interesting to assess the impact of modifying the
delivery frequency requirement (currently five deliveries per week) on both the
incumbent’s cost of ODW and the extent of entry by the competitor, but this is left
for future work. Such future work could also reflect the effects of different delivery
frequencies and different numbers of locations served on demand for the entrant’s
service.
Acknowledgements The authors are grateful to Michael Crew, Tim Brennan, Carla Pace, Axel
Desmedt, Robert Campbell, Ian Streule, and participants at the 23rd CRII conference on postal and
delivery economics.
Disclosure This work was funded by the University of Liege and the FNRS (Belgian National
Scientific Research Fund).
Appendix: Hypotheses
Parameter Hypothesis Reference

Postman salary per w ¼ 28.81 Statbel
hour (€)
Car cost per kilo- pcar ¼ 0.36 Tera Consul-
meter (€) tants (2013)
Length of the deliv- Li ¼ 1.5li Jasinski and
ery round (km) Steggles (1977)
Speed (km/h) speed car ¼ 4 if Li
si < 0:02 Roy (1999)

speed car ¼ 4 þ 31
0:298
Li
si 0:02 if 0:02 < Lsii < 0:3
speed car ¼ 35 if si > 0:3
Li
speed bike ¼ 4 if si < 0:02

Li

speed bike ¼ 4 þ 0:098
11 Li
si 0:02 if 0:02 < Lsii < 0:1
speed bike ¼ 15 if Li
si > 0:1
speed foot ¼ 4
Stop time (s) stop timecar ¼ 12 þ ðgi 1Þ19
9
Roy (1999)
stop timebike ¼ 20 þ ðgi 1Þ21
9
stop timefoot ¼ 40 þ ðgi 1Þ40
9
Delivery time (s) delivery time ¼ 3 Roy (1999)
Maximum capacity qmax,foot ¼ 400 Roy (1999)
qmax,bike ¼ 400
Loading time (s) loading time ¼ 600 Tera Consul-
tants (2013)
(continued)
Parameter Hypothesis Reference

Delivery frequency d¼5 WIK (2013)
(days/week)
Items delivered per x ¼ 233 WIK (2013)
capita par year
Items delivered per q ¼ x/(52 * d )
capita par day
Items delivered in Qi ¼ q popi
the area par day
References
Boldron F, Joram D, Martin L, Roy B (2007) From size of the box to the costs of universal service
obligation: a cross-country comparison. In: Crew MA, Kleindorfer PR (eds) Liberalization of
the postal and delivery sector. Edward Elgar, Cheltenham
Borsenberger C, Cremer H, De Donder P, Joram D, Roy B (2010) Funding the cost of universal
service in a liberalized postal sector. In: Crew MA, Kleindorfer PR (eds) Heightening compe-
tition in the postal and delivery sector. Edward Elgar, Cheltenham
Cohen RH, Chu EH (1997) A measure of scale economies for postal systems. In: Crew MA,
Kleindorfer PR (eds) Managing change in the postal and delivery industries. Kluwer Aca-
demic, New York
Copenhagen Economics (2010) Main developments in the postal sector (2008–2010). Study for
the European Commission, Directorate General for Internal Market, Copenhagen
Crew MA, Kleindorfer PR (2005) Competition, universal service and the graveyard spiral. In:
Crew MA, Kleindorfer PR (eds) Regulatory and economic challenges in the postal and delivery
sector. Kluwer Academic, New York
D’Alcantara G, Gautier A (2008) National postal strategies after a full postal market opening. In:
Crew MA, Kleindorfer PR (eds) Competition and regulation in the postal and delivery sector.
Edward Elgar, Cheltenham
De Donder P, Cremer H, Dudley P, Rodriguez F (2006) A welfare analysis of price controls with
end-to-end mail and access services. In: Crew MA, Kleindorfer PR (eds) Liberalization of the
postal and delivery sector. Edward Elgar, Cheltenham
Fève F, Florens JP, Rodriguez F, Soteri S (2010) Forecasting mail volumes in an evolving market
environment. In: Crew MA, Kleindorfer PR (eds) Heightening competition in the postal and
delivery sector. Edward Elgar, Cheltenham
Gautier A, Poudou J-C (2015) Reforming the postal universal service, Rev Network Econ
forthcoming
Gautier A, Paolini D (2011) Universal service financing in competitive postal markets: one size
does not fit all. Rev Network Econ 10(3)
Gautier A, Wauthy X (2010) Price competition under universal service obligations. Int J Econ
Theory 6:311–326
Gautier A, Wauthy X (2012) Competitively neutral universal service obligations. Inf Econ Policy
24:254–261
Jasinski KM, Steggles TJ (1977) Modeling letter delivery in town areas. Comput Oper Res 4
(4):287–294
Meschi M, Cherry M, Pace C, Petrova M (2010) Understanding the impact of e-substitution on
letter mail volumes: a multi-country panel study. In: Crew MA, Kleindorfer PR (eds)
Reinventing the postal sector in an electronic age. Edward Elgar, Cheltenham
Nikali H (2008) Substitution of letter mail for different sender-receiver segments. In: Crew MA,
Kleindorfer PR (eds) Competition and regulation in the postal and delivery sector. Edward
Elgar, Cheltenham
Rohr C, Trinkner U, Lawrence A, Kim C-W, Potoglou D, Sheldon R (2013) Measuring consumer
preferences for postal services. In: Crew MA, Kleindorfer PR (eds) Reforming the postal sector
in the face of electronic competition. Edward Elgar, Cheltenham
Roy B (1999) Technico-economic analysis of the costs of outside work in postal delivery. In: Crew
MA, Kleindorfer PR (eds) Emerging competition in postal and delivery services. Kluwer
Academic, New York
Tera Consultants (2013) Vérification du calcul du co^
ut net du service universel postal en Belgique.
Study for the BIPT, Brussels
Valletti T, Hoernig S, Barros P (2002) Universal service and entry: the role of uniform pricing and
coverage constraints. J Regul Econ 21(2):169–190
WIK (2013) Main developments in the postal sector (2010–2013). Study for the European
Commission, Directorate General for Internal Market and Services, Bad Honnef
WIK (2015) Review of the postal market three years after full market opening. Study for the BIPT,
Brussels
Compensation Fund Under EU Law: A
Suitable Solution for the Postal Market?
Alessandra Fratini
1 Introduction
As in the telecoms and electricity sectors, the EU Postal Directive1 allows Member
States to set up compensation funds to finance universal service obligations (USO).
Postal operators not responsible for the USO can be required to share the net cost of
the USO.
Provided for in the first Postal Directive (97/67/EC), the use of compensation
funds as a funding mechanism in the postal sector has been rather limited for some
time—the only (unsuccessful) example being Italy. More recently, in a period of
scarce financial resources and in the context of the ongoing privatization processes,
an increasing number of Member States have decided to implement such compen-
sation funds to secure the financing of the USO and, where applicable, maximize
the proceeds of the planned privatizations.
The European Commission considers that financing through compensation funds
is subject to State aid control. In its first decision on such schemes, which opened an
in-depth investigation on the compensation fund envisaged by Greece in favor of
ELTA, the Commission warned that it will ensure that the compensation fund
mechanism would strike the right balance between the objectives of securing the
USO on the one hand and allowing fair competition in the postal market.2
1
Directive 97/67/EC of 15 December 1997 on common rules for the development of the internal
market of Community postal services and the improvement of quality of service, OJ L
15, 21.1.1998, p. 14, as last amended by Directive 2008/6/EC of 20 February 2008 amending
Directive 97/67/EC with regard to the full accomplishment of the internal market of Community
postal services, OJ L 52, 27.2.2008, p. 3.
2
State aid SA.35608 (2014/C) (ex 2014/N)—Greece, Hellenic Post (ELTA)—Compensation for
the financing of the universal postal service, (2014) 5436 final of 01.08.2014.
A. Fratini (*)
FratiniVergano - European Lawyers, Brussels, Belgium
e-mail: a.fratini@fratinivergano.eu

154 A. Fratini
The chapter, after a brief overview of the relevant provisions under the telecoms
and electricity Directives (Sect. 2), will examine the design of the compensation
fund under the Postal Directive3 (Sects. 3 and 4), with a specific focus on the
transparency, nondiscrimination and proportionality requirements (Sect. 5). It will
also review the compensation fund from a State aid perspective, against the
background of the current EU postal scenario, taking into account the Commis-
sion’s emerging position (Sect. 6). Section 7 concludes.
2 The Compensation Fund as a USO Financing Mechanism

in the Telecoms and Electricity Sectors
As a general rule which applies across public services, for certain services of
general economic interest (SGEIs) to operate under conditions that enable them
to fulfil their missions, external financial support may prove necessary where
revenues accruing from the provision of the service do not allow the costs resulting
from the public service obligation to be covered.4 Member States enjoy the freedom
to decide the method of financing that is best adapted to their particular situation,
provided that they avoid any disproportionate distortions to the functioning of the
market.
Along with the possibility to compensate the SGEI provider via direct public
funds, some sector-specific frameworks explicitly provide for the option of intro-
ducing a “compensation fund” of the net costs of the USO, financed by the
undertakings active in the market.
That is the case of the electronic communications sector. Article 13 of the
Universal Service Directive5 provides that “[w]here, on the basis of the net cost
calculation [. . .], national regulatory authorities find that an undertaking is subject
to an unfair burden, Member States shall, upon request from a designated under-
taking, decide: (a) to introduce a mechanism to compensate that undertaking [. . .]
from public funds; and/or (b) to share the net cost of universal service obligations
between providers of electronic communications networks and services.” Under the
Universal Service Directive, the cost recovery “need not result in any distortion of
competition, provided that designated undertakings are compensated for the spe-
cific net cost involved and provided that the net cost burden is recovered in a
competitively neutral way.”6
3
Article 7(5) of the Postal Directive.
4
Communication from the Commission—European Union framework for State aid in the form of
public service compensation (2011), OJ C 8, 11.1.2012, p. 15.
5
Directive 2002/22/EC of 7 March 2002 on universal service and users’ rights relating to
electronic communications networks and services (Universal Service Directive), OJ L
108, 24.4.2002, p. 51.
6
Ibidem, Recital 4.
Compensation Fund Under EU Law: A Suitable Solution for the Postal Market? 155
In the case of cost recovery by means of levies on undertakings, Member States

should ensure that the sharing mechanism is based on objective and nondiscri-
minatory criteria and is in accordance with the principles of proportionality,
transparency and least market distortion.7 The principle of proportionality does
not prevent Member States from exempting new entrants that have not yet achieved
any significant market presence (the so-called de minimis threshold). At the same
time, the principle of least market distortion “means that contributions should be
recovered in a way that as far as possible minimizes the impact of the financial
burden falling on end users, for example, by spreading contributions as widely as
possible.”8 In order to respect the principle of proportionality, a number of Member
States have provided for a maximum ceiling on the individual contributions of
operators, set as a percentage of the operators’ national annual turnover from the
provision of electronic communication networks and services, with the balance of
the net cost financed from public funds.9
Article 13(2) of the Universal Service Directive further requires that the sharing
mechanism be administered by the national regulatory authority (NRA) or a body
independent from the beneficiaries under the supervision of the NRA, while an
entire provision is devoted to the transparency requirement in setting up and
administering the sharing mechanism.10
There are more limited provisions in the energy sector, with the Electricity
Directive merely requiring that “[w]here financial compensation, other forms of
compensation and exclusive rights which a Member State grants for the fulfillment
of the obligations [the USO] are provided, this shall be done in a nondiscriminatory
and transparent way.”11 The same Directive makes it clear12 that, to the extent that
measures taken by Member States to fulfil public service obligations constitute
State aid under the Treaty, there is an obligation to notify them to the
Commission.13
The required compliance with the principles of EU law, such as nondiscri-
mination, proportionality and administration by independent authorities, together
7
Article 13(3).
8
Recital 23.
9
Communication from the Commission, Universal service in e-communications: report on the
outcome of the public consultation and the third periodic review of the scope in accordance with
Article 15 of Directive 2002/22/EC, COM(2011)0795 final.
10
Article 14.
11
Directive 2009/72/EC of 13 July 2009 concerning common rules for the internal market in
electricity and repealing Directive 2003/54/EC, OJ L 211, 14.8.2009, p. 55, Article 3(8).
12
Recital 49 of the Electricity Directive.
13
See, for example, Commission Decision of 28 January 2009 on aid implemented by Luxem-
bourg in the form of the creation of a compensation fund for the organization of the electricity
market (Case C 43/02 (ex NN 75/01), OJ L 159, 20.6.2009, pp. 11–20), on the establishment of a
compensation fund to share the additional costs associated with the obligation to purchase “green”
electricity between final consumers.
156 A. Fratini
with State aid control where the measure involves State aid, similarly apply to USO
financing via compensation fund in the postal sector.
3 The Compensation Fund as Progressively Disciplined

Under The Postal Directives
The compensation fund was provided for already in the first Postal Directive14:
Article 9(4) of Directive 97/67/EC referred to the possibility of establishing a
compensation fund “in order to ensure that the universal service is safeguarded,”
by making the granting of authorizations subject to an obligation to contribute to
that fund. It also established the obligation for Member States to entrust an
independent body for the administration of the fund and to “ensure that the
principles of transparency, nondiscrimination and proportionality are respected in
establishing the compensation fund and when fixing the level of the financial
contributions.” The Directive made it clear that the compensation fund could only
be used to finance the activities falling within the scope of the universal service.
The actual implementation of that provision was at first limited: in 2002, in its
first Application Report,15 the Commission noted that “7 Member States have
introduced provisions for a compensation fund in their national regulatory frame-
work, of which only one has firm plans to activate such a fund.” That was confirmed
in 2006 by the European Parliament in its resolution in response to the Commission
second Application Report,16 where it stated that the funding models for USO used
that far in the Member States had not been very successful and that the tried and
tested funding instrument for universal services in the past had been the reserved
sector.
Later, in the 2008 Staff Working document accompanying the third Application
Report,17 the Commission acknowledged that while several Member States have
made provisions for setting up compensation fund arrangements, it was only Italy
that had actually established one. With the reserved area to be abolished by
31 December 2010 and thus no longer available to finance the US, the Commission
admitted that the issue of compensation funds would become more important and
“further consideration may be needed in the future to see how to make this optional
tool more operational.”18 That was restated in the proposal for the third Postal
14
Directive 97/67/EC, cit.
15
Report from the Commission on the application of the Postal Directive, COM(2002)0632 final.
16
European Parliament resolution on the application of the Postal Directive (Directive 97/67/EC,
as amended by Directive 2002/39/EC), 2005/2086(INI), 02.02.2006.
17
Commission Staff Working Document—accompanying document to the Report from the
Commission on the application of the Postal Directive (Directive 97/67/EC as amended by
Directive 2002/39/EC), COM(2008) 884 final, 22.12.2008.
18
Ibidem, p. 19.
Directive19: “[t]o date, experience with the compensation fund as a mechanism for
financing universal postal service has been very limited. However, this is not
surprising, as the substantial reserved area still in place in most Member States
was an obvious source of financing. Without the reserved area however, it is likely
that several Member States may revisit the need for a compensation fund, making it
a potential safeguard that should be maintained.”
Against the background of full market opening, it became appropriate to explic-
itly clarify in the third Postal Directive20 the alternatives available to Member
States to ensure the USO financing, where needed and adequately justified, while
leaving Member States the choice of the financing mechanisms to use. Article 7 of
the Postal Directive was entirely replaced. For what matters here, Article 7(3)
(b) provides that where a Member State determines that the USO entails a net
cost,21 and represents an unfair financial burden on the USP(s), it may introduce “a
mechanism for the sharing of the net cost of the universal service obligations
between providers of services and/or users.” Article 7(4) further provides that “a
compensation fund [. . .] may be funded by service providers and/or users’ fees, and
is administered for this purpose by a body independent of the beneficiary or
beneficiaries.” Member States may make the granting of authorizations to service
providers subject to an obligation to make a financial contribution to that fund or to
comply with universal service obligations. Under the newly amended Article 9(2),
in fact, for services which fall within the scope of the USO, Member States may
make the granting of authorizations subject, where appropriate, to an obligation to
make a financial contribution to the compensation fund.
With a view to making the fund “more operational,” it was established that, in
order to determine which undertakings may be required to contribute to a compen-
sation fund, Member States “should consider whether the services provided by such
undertakings may, from a user’s perspective, be regarded as services falling within
the scope of the universal service, as they display interchangeability to a sufficient
degree with the universal service, taking into account the characteristics of the
services, including added value features, as well as the intended use and the pricing.
These services do not necessarily have to cover all the features of the universal
service, such as daily delivery or complete national coverage22 (the so-called
interchangeable services). However, “[e]xcept in the case of undertakings that
have been designated as universal service providers [. . .], authorizations may not:
[. . .] for the same elements of the universal service or parts of the national territory,
19
Proposal for a Directive of the European Parliament and of the Council amending Directive
97/67/EC, COM(2006)0594 final.
20
Directive 2008/6/EC of 20 February 2008 amending Directive 97/67/EC with regard to the full
accomplishment of the internal market of Community postal services. OJ L 52, 27.2.2008, p. 3.
21
The USO net cost is the difference between the net cost for a designated USP of operating with
the USO and the same provider operating without it, taking into account, however, any intangible
benefits.
22
Recital 27 of Directive 2008/6/EC.
158 A. Fratini
impose universal service obligations and, at the same time, financial contributions
to a sharing mechanism.”
Article 7(5) recalls that the principles of transparency, nondiscrimination and
proportionality shall be complied with when establishing the compensation fund
and when fixing the level of the financial contributions. Any decision in that respect
“shall be based on objective and verifiable criteria and be made public.”23
Interestingly enough, the caveat as to the potential application of the rules on
State aid appeared in a recital of the second Postal Directive, which warned that
“the establishment of a compensation fund or any change in its operation or any
implementation of or payment from it, may involve aid granted by a Member State
or through State resources in any form whatsoever within the meaning of Article 87
(1) of the Treaty necessitating prior notification to the Commission pursuant to
Article 88(3) thereof.”24
4 Practical Implementation
The establishment and management of a compensation fund in the postal sector is a

complex exercise. Besides the net avoided cost and unfair burden verification,
which also apply to public funding, the competent authorities have to define the
services displaying “interchangeability to a sufficient degree” to the US,25 to
identify the operators requested to contribute to the fund and the method to
calculate their contribution. The actual administrative management of the fund
requires a preliminary scrutiny of postal services’ classification, and related cost
accounting,26 for all undertakings operating in the US and non-US postal markets,
with a view to detect biased revenue declarations from the providers.
To date, experience with the compensation fund as a mechanism for financing
the USO continues to be limited. In 2013, while 22 Member States27 had authorized
23
Further indications are provided in Annex I, Part C.
24
Recital 25 of Directive 2002/39/EC of the European Parliament and of the Council of 10 June
2002 amending Directive 97/67/EC with regard to the further opening to competition of Commu-
nity postal service, OJ L 176, 5.7.2002, p. 21.
25
With respect to the definition of interchangeable services, in 2008 the Commission excluded that
hybrid mail services, i.e., services involving a communication electronically transmitted to the
service provider, electronically processed and converted into a physical mail item delivered to the
addressee, be considered interchangeable to the US (decision C(2008) 5912 final of 7 October
2008—Case COMP/39.562, confirmed by the General Court on 25 March 2015, Case T-556/08,
Slovensk a pošta v Commission, not yet published).
26
In this respect, Member States may require those providers called to contribute to a compensa-
tion fund “to introduce an appropriate accounting separation to ensure the functioning of the fund”
(Article 14(10) of the Postal Directive).
27
Austria, Cyprus, Czech Republic, German, Denmark, Estonia, Greece, Spain, France, Croatia,
Hungary, Italy, Luxembourg, Latvia, Malta, Poland, Portugal, Romania, Slovenia, Slovakia, the
United Kingdom, Ireland.
the establishment of a compensation fund, only four28 had actually established

it. According to available figures, the funds appear to have covered an immaterial
share of the USO cost.29 In Italy, where authorized postal services providers
(license and authorization holders) may be required to contribute to the fund within
a maximum of 10 % of the gross revenues from the US and interchangeable
services30—the level of contribution having been regularly set by the regulatory
authority at 3 %—compensation collected via the fund in the 2000–2010 period
accounted for 0.01 % of the USO net cost.31 In Slovakia, contributions by providers
of interchangeable postal services, set by the regulatory authority within a maxi-
mum of 3 % of their turnover from the provision of interchangeable services,32 shall
correspond to the market share of the provider, determined as the ratio of its
interchangeable services’ turnover to the total turnover of the other providers of
interchangeable services, plus the turnover of the USP from provision of the USO.
Contributions into the fund for 2012 were worth less than 1 % of the USO net
cost.33
Even so, as noted by the Commission, “[i]n a period characterized by scarce
financial resources, some Member States have recently expressed a desire to
implement such compensation funds for the financing of the USO”34 also in
connection with privatization plans. That makes the compatibility requirements,
under both the Postal Directive and State aid law, worth looking at in more detail.
5 The Compatibility Requirements Under the Postal

Directive
The Postal Directive leaves to Member States the conditions governing the estab-
lishment of the compensation fund as well as the level of the financial contributions,
but it does require that Member States respect the principles of transparency,
nondiscrimination, and proportionality.35 Compliance with these principles repre-
sents the key benchmark for the assessment of the compatibility of a compensation
fund with the Postal Directive. To date, there is no precedent in that respect in the
Commission’s decision making practice or in the Application Reports.
28
Cyprus, Estonia, Italy, Slovakia (Wik-Consult 2013).
29
“The Role of Regulators in a More Competitive Postal Market”, cit., p. 124.
30
Article 10, Decreto Legislativo 22 luglio 1999, n. 261, GU n. 182, 5.8.1999, as subsequently
modified. A Ministerial Decree of 17.04.2000 provides the criteria for the functioning of the fund.
31
Poste Italiane’s calculations.
32
Article 58(1), Act on Postal Services of 14.09.2011 (n.324/2011).
33
Vestnı́k Úradu pre regul aciu elektronickyćh komunik
aciı́ a poštovyćh sluzˇieb—č. 5/2014 z
20.6.2014—Prı́jmy a vy´davky kompenzačného fondu za rok 2013.
34
Competition policy brief “High quality and competitive postal services for citizens and busi-
nesses—State aid control in the postal sector”, Issue 6, May 2014.
35
Article 7(5).
160 A. Fratini
5.1 The Transparency Principle
The principle of transparency is understood, amongst other, as imposing a require-

ment for prior publication of unambiguous, objective criteria on which the exercise
of administrative discretion affecting rights and freedoms under the Treaties will be
based. Accordingly, the Postal Directive requires that the decision to establish a
compensation fund be based on objective and verifiable criteria and be made public.
Compliance with this requirement appears relatively undemanding, as it implies the
publication of the procedure, and the conditions thereof, for the establishment of the
fund, as well as of the criteria governing its functioning.36
5.2 The Nondiscrimination Principle
The principle of nondiscrimination, instead, triggers further reflections. This gen-

eral principle of EU law, also referred to as “equal treatment,” requires that
comparable situations shall not be treated differently and that situations which are
not comparable shall not be treated in the same way, unless such different treatment
is objectively justified.
In connection with the compensation fund, the nondiscrimination requirement
under Article 7(5) arguably entails that the USP(s) should also contribute to the
fund proportionally to its revenues generated by the provision of both universal
services and interchangeable services. The principle is thus relevant under two
perspectives, i.e., the participation of the USP(s) to the fund and the extent of its
participation.
As regards the very participation of the USP(s) to the compensation fund, Article
7(4) of the Postal Directive, seen above, makes it clear that the granting of
authorizations to service providers under Article 9(2) may be subject “to an
obligation to make a financial contribution to that fund or to comply with universal
service obligations” (italic added). In other terms, providers of services falling
within the scope of the US either provide the US (or part of it) or contribute to
financing its cost. In the preparatory works of the third Postal Directive,37 the
Commission ruled out any “concurrent requirement to contribute to a sharing
mechanism and the imposition of US or quality obligations for the same quality,
availability, and performance requirements. The universal service obligations
spread the uneconomic costs of the service between several companies while the
compensation fund imposes a levy on companies to finance the uneconomic cost
incurred by one company, making its combined application unjustified.”
36
Annex I requires a “transparent and neutral mechanism for collecting contributions.”
37
Proposal for a Directive of the European Parliament and of the Council amending Directive
97/67/EC (COM/2006/594 final), § 3.3.1.
Accordingly, the combined application of a USO and of a contribution to the

compensation fund for the same part of the USO would appear unjustified under
the Directive.
However, the same Article 9(2) explicitly excepts the USP(s) from the above
(“[e]xcept in the case of undertakings that have been designated as universal service
providers in accordance with Article 4, authorizations may not: [. . .] for the same
elements of the universal service or parts of the national territory, impose universal
service obligations and, at the same time, financial contributions to a sharing
mechanism [. . .]”). While it may be argued that Article 9(2) should be interpreted
in the light of Article 4(1) of the Postal Directive, which allows Member States to
designate one or more undertakings as USP(s)38—in a context of several designated
USPs, where one or some of the USPs incur in an unfair financial burden, the other
USP(s) may well be required to contribute to the financing of that unfair burden via
the compensation fund, thus being subject to both the provision of (part of) the USO
and the contribution to compensate the net cost of (the other part of) the USO—it
shall be concluded that excluding the USP(s) from the contribution to the financing
of the unfair burden would be discriminatory, as it would totally relieve it from its
equivalent burden and ultimately shift the latter entirely onto the USP’s competitors
in the same market.
In connection with the extent of the USP’s participation, pursuant to the equal
treatment principle the USP should also contribute to the fund proportionally to its
revenues generated by the provision of interchangeable services. That was
established already by the Court of Justice in the 1999 TNT Traco case.39 The
case concerned the Italian provisions that made the supply of express mail services
by undertakings other than the USP subject to payment of the postal dues ordinarily
applicable to the universal service and allocated the proceeds of those dues to the
USP. The judgment predates the Postal Directive and concerns the application of
Articles 86 and 90 EC (now Articles 102 and 106 TFEU), still it can be usefully
relied upon to justify the requirement that USP’s revenues from interchangeable
services be subject to the contribution to the compensation fund. The Court in fact
held that “[i]n those circumstances, the undertaking responsible for the universal
postal service must also be required, when itself supplying an express mail service
not forming part of that service, to pay the postal dues.”
But if the rationale of the nondiscrimination requirement in Article 7(5) of the
Postal Directive is that the USP competes with undertakings providing services
within the scope of the US that implies that the contribution of the USP to the
compensation fund should be calculated by only taking into account its revenues
from the services provided in the markets where there is actual competition, i.e.,
where there are competitors providing “interchangeable services.” It would run
contrary to the equal treatment principle to require that the USP alone contributes to
38
Member States are allowed to designate different undertakings to provide different elements of
USO and/or to cover different parts of the national territory.
39
Case C-340/99, TNT Traco SpA v. Poste Italiane SpA and others, [2001] ECR I-4109, § 58.
162 A. Fratini
the fund with respect to those services within the scope of the US, where there is no
active competitor and no “interchangeability” applies. That would be the case, for
example, of the registered mail service used in the course of judicial or adminis-
trative procedures, if the USP is granted exclusive rights in that respect in accor-
dance with Article 8 of the Postal Directive.40
5.3 The Proportionality Principle
The proportionality requirement under Article 7(5) is to be intended as the propor-

tionality stricto sensu test,41 which refers to the assessment of whether a given
measure, although suitable and necessary in order to achieve the aim proposed to
achieve by using the chosen measure, nevertheless imposes an excessive burden.
In the context of the compensation fund, the proportionality test applies in
particular—above and beyond the amount of the compensation42—to the level of
the financial contributions from services providers. Recital 28 of the Postal Direc-
tive stipulates that, to comply with the principle of proportionality when determin-
ing the level of contribution, Member States shall use transparent and
nondiscriminatory criteria, such as the “the share of these undertakings in the
activities falling within the scope of the universal service.” In the light of the
principle of least market distortion, this requirement is to be intended as imposing
financial contribution’s thresholds which do not hinder the ability of operators
(especially the USP’s smaller competitors—see Oxera 2007) to enter and expand
profitably in market. The proportionality of the level (normally set as a share of the
revenues generated in the relevant services) is to be assessed in relation to the
average profit that any operator can be expected to make in the US market,
including interchangeable services.
This aspect has been covered in the ELTA case, which will be analyzed in the
following sections.
40
Article 8 acknowledges “the Member States’ right to organize the siting of letter boxes on the
public highway, the issue of postage stamps and the registered mail service used in the course of
judicial or administrative procedures in accordance with their national legislation.”
41
According to its conventional understanding, the proportionality principle consists of three tests
applied to a given measure: the suitability, the necessity, and the proportionality stricto sensu test.
The suitability test refers to the relationship between the means and the end (i.e., whether the
chosen measure is suitable in order to achieve the given aim proposed to achieve by using the
chosen measure); the necessity test implies assessing whether the chosen measure is necessary to
achieve the proposed goal; the proportionality stricto sensu aims at checking whether the measure,
while suitable and necessary, imposes an excessive burden.
42
See TNT Traco, cit., § 58: “Article 90(2) of the Treaty [now 106(2) TFEU] does not allow the
total proceeds [. . .] which are paid by economic operators [. . .] to exceed the amount necessary to
offset any losses which may be incurred in the operation of the universal postal service.”
6 The Compensation Fund from a State Aid Law

Perspective
While the second Postal Directive had laid down that the establishment and
implementation of a compensation fund may involve State aid within the meaning
of the Treaty,43 and the third Directive made clear that, insofar as mechanisms for
the financing of the USO involve aid, “this Directive is without prejudice to
Member States’ obligation to respect the Treaty rules on State aid,” it was not
until the 2014 ELTA opening decision that the compensation fund came under State
aid law scrutiny.
6.1 Financing Though the Fund as State Aid
Financing through compensation funds is subject to State aid control as it is likely to

meet the four cumulative conditions under Article 107 TFEU for a measure to
constitute State aid: it is a measure imputable to the Member State and granted
through State resources; it confers a selective economic advantage to an undertak-
ing; it distorts or threatens to distort competition and affects trade between Member
States.
The Court of Justice clarified that Article 107(1) TFEU covers all the financial
means by which the public authorities may actually support undertakings: the fact
that these means remain constantly under public control is sufficient for them to be
categorized as State resources irrespective of whether or not those means are
permanent assets of the public sector.44 In particular, the originally private nature
of the contributions does not prevent them from being regarded as State resources,45
being the relevant criterion that of the degree of intervention of the public authority
in the definition of the measure and its methods of financing.46
In connection with the compensation fund, it can be soundly argued that the
resources of the fund are under public control, as the State decides, through
legislation or regulation, who should contribute to the fund, how their level of
contribution should be calculated, who benefits from these contributions and
administers the fund. Accordingly, financing via the compensation fund involves
a transfer of State resources imputable to the State within the meaning of Article
107(1) TFEU.47
43
Recital 25.
44
Case C-83/98 P, France v Ladbroke Racing and Commission [2000] ECR I-3271, § 50, and Case
C-482/99 France v Commission, § 37.
45
Case T-358/94 Air France v Commission [1996] ECR II-2109, §§ 63–65.
46
Cases T-139-09, T-243-09, T-328-09, France v. Commission Contingency Plans.
47
ELTA opening decision, cit., §§ 122–123.
164 A. Fratini
The second element for the purpose of Article 107(1) TFEU consists in an
economic advantage which the undertaking would not have obtained under normal
market condition. As regards compensation to undertakings entrusted with SGEIs,
the Court laid down in Altmark48 that the granting of an advantage can be excluded
if four cumulative conditions are met. For purposes of the famed fourth condition,49
in the absence of a procurement procedure, there has been this far no precedent for
the successful application of a benchmarking exercise to verify that the USO costs
covered by the State funding are those of a typical efficiently run undertaking, as
there are few or no other undertakings with which to compare the USP. That was the
case with ELTA too: “[t]he fourth condition is clearly not met in the case of ELTA
since the public service was not awarded as a result of an open public procurement
procedure, nor have the Greek authorities argued or otherwise provided the element
demonstrating that ELTA is compensated according to the costs of a typical
undertaking within the sector.”50 The Commission concluded that the compensa-
tion could be qualified as economic advantage within the meaning of Article 107
(1) TFEU. As regards the selectivity nature, under Article 107 TFEU a measure is
selective when it favors “certain undertakings or the production of certain goods.”
Compensation funds are designed to benefit the USP(s) alone and are therefore
selective.
Finally, in order to be caught by Article 107(1), public service compensation
must distort or threaten to distort competition, “in so far as it affects trade between
Member States.” Such an effect generally presupposes the existence of a market
open to competition. Therefore, where markets have been opened up to competition
either by Union or national legislation or de facto by economic development, State
aid rules apply. That is clearly the case in the postal (and financial) sectors, where
there is competition and intra-Union trade: the public support granted to the USP
would strengthen the position of the company in relation to postal and financial
undertakings established in the same or other Member States which, as a conse-
quence, might have more difficulties to enter or to remain in that market. Therefore,
the financing via the compensation fund confers a selective economic advantage to
the USP(s), which is liable to distort competition and affect intra-Union trade
pursuant to Article 107(1) TFEU.51
48
Case C-280/00, Altmark Trans GmbH and Regierungspr€ asidium Magdeburg v Nahverkehrsge-
sellschaft Altmark GmbH [2003] ECR-I-07747.
49
The fourth condition requires that “[w]here the undertaking which is to discharge public service
obligations is not chosen pursuant to a public procurement procedure, the level of compensation
needed must be determined on the basis of an analysis of the costs which a typical undertaking,
well run and adequately provided with means to meet the necessary public service requirements,
would have incurred in discharging those obligations, taking into account the relevant receipts and
a reasonable profit for discharging the obligations.”
50
ELTA opening decision, cit., §§ 109–110.
51
Ibidem, § 115.
6.2 The Compatibility Assessment Under the SGEI

Framework: The ELTA Case
Under certain conditions, Article 106(2) TFEU allows the Commission to declare
State aid in the form of SGEI compensation compatible with the internal market.
The compatibility test is carried out pursuant to the SGEI Decision52 or the SGEI
Framework,53 for compensations respectively below and exceeding €15 million per
year. Interestingly for this analysis, and relied upon in ELTA as it will be seen
below, both the Decision and the Framework commend that the compensation
measure shall also comply with the requirements flowing from the relevant sectorial
Union legislation (i.e., the Postal Directive).
In ELTA, as the envisaged amount of aid could potentially exceed €15 million
per year, the Commission assessed the compensation fund mechanism under the
SGEI Framework and concluded that it met all the conditions thereof which “are
usually sufficient to ensure that aid does not distort competition in a way that is
contrary to the interests of the Union,” pursuant to its paragraph 51.54 However, in
compliance with paragraph 52 of the SGEI Framework, which provides that “[. . .]
in some exceptional circumstances, serious competition distortions in the internal
market could remain unaddressed and the aid could affect trade to such an extent as
would be contrary to the interest of the Union,” the Commission noted that “certain
characteristics of the notified compensation fund mechanism may raise specific
issues concerning the development of EU trade.”55 Those concerns warranted an
in-depth analysis, with a view to preventing potential distortions of competition
contrary to the interests of the EU.
The Commission was in particular concerned of the level of contribution
requested of services providers. Under the notified compensation scheme, in fact,
while ELTA was required to contribute with the 0.5 % of its turnover within the US
area during the respective fiscal year, the contribution from the other operators was
52
Commission Decision of 20 December 2011 on the application of Article 106(2) TFEU to State
aid in the form of public service compensation granted to certain undertakings entrusted with the
operation of services of general economic interest (notified under document C(2011) 9380), OJ L
7, 11.1.2012, p. 3.
53
Communication from the Commission—European Union framework for State aid in the form of
public service compensation (2011), cit.
54
Notably, that the USO financed through the fund was a genuine SGEI; that ELTA had been
entrusted with the operation of the SGEI by way of an official act specifying the public service
obligations and the methods of calculating compensation; that the duration of the period of
entrustment was justified; that ELTA’s accounting and cost allocation system was compliant
with Directive 2006/111/EC; that the direct entrustment of ELTA as the USO provider was
compatible with § 19 of the SGEI Framework and that, being the USO only assigned to ELTA,
there could be not any discrimination in the sense of § 20 thereof; that the methodology used by the
Greek authorities to calculate the compensation amount was in line with the requirements of the
SGEI Framework, with no risk of overcompensation; that Greece had committed to comply with
the transparency requirements under the SGEI Framework.
55
ELTA opening decision, cit., § 191.
166 A. Fratini
calculated as the result of a basic contribution (0.5 % of turnover within the US

area, as for ELTA) plus an “increased contribution” which could range from 0.5 to
10 % of their revenues in the US area (thus potentially 20 times greater) depending
on their fraction of urban distribution.56
In the Commission’s view, on the one hand the differentiation based on urban
distribution, aimed at addressing cherry-picking practices, could discourage com-
petition “as new entrants would naturally start operating in cities where most of the
market is,” and the very high difference between ELTA’s and competitor’s contri-
bution seemed prima facie unjustified.57 On the other hand, the maximum contri-
bution requested from competitors would be higher than ELTA’s own profit in
urban areas [6–7 %] in 2010–2011, when ELTA still benefited from the reserved
area. Therefore, “[i]t should be difficult for new entrants to reach such a profit level
in the future in a market open to competition. The reasonable profit calculated by
the Greek authorities [5.76–6.11 %] [. . .] which corresponds to the normal profit in
the market is also lower than the maximal contribution.”
The Commission noted the likely noncompliance of those aspects with the Postal
Directive as well: “it is also doubtful that the compensation fund mechanism
notified by the Greek authorities complies with the conditions of transparency,
nondiscrimination, and proportionality requested by the third Postal Directive.”58
Likewise, the imposition of “too high” contributions to postal operators competing
with ELTA, to the extent that it could lead to a foreclosure of the postal market or at
least to a re-monopolization of the USO area, “would be completely contrary to the
liberalization objectives of the third Postal Directive.”59
It then resolved to initiate the formal investigation procedure provided for in
Article 108(2) TFEU to assess the potential impact on competition, and in particular
on ELTA’s competitors’ profit level, of the notified measure, with a focus in
particular on the following: the tax base (revenue or profit); the differentiation
mechanism between operators (nondiscrimination); the cap on contributions from
competitors (proportionality); the opportunity of a “délai de grace” for new
entrants in the market; other existing or desirable components of the compensation
fund mechanism.
Based on the financial information received from the various actors in the
relevant postal markets (ELTA and its competitors in the first place), it was
expected that the Commission would complete the assessment of the measure and
take a decision shortly. At the time of writing, the formal investigation procedure
was still ongoing.
56
See Footnote 1: “In practice, the percentage of contribution of a given operator on its turnover
amounts to 0.5 % +0.5 %* [difference in percentage between the urban concentration of ELTA’s
distributed items in the USO perimeter and the urban concentration of the operator]. For example,
if ELTA offers 60 % of its services to urban areas, as soon as an operator distributes 79 % of its
postal items in urban areas, its contribution is equal to 10 % of its turnover in the USO area.”
57
Ibidem, §§ 195–196.
58
Ibidem, § 198.
59
Ibidem, Footnote 50.
7 Conclusion
Until lately, the experience with the compensation fund in the postal sector has
been limited and essentially little effective as a USO financing tool. One of the
most recent schemes aimed at making it “more operational” has been considered
as being potentially incompatible with both the Postal Directive and the SGEI
Framework.
The ELTA opening decision shows that, although the financing of the USO
through a compensation fund is in principle acceptable and foreseen by the Postal
Directive, the concrete design and implementation of the compensation fund
mechanism may be challenged from a State aid law perspective because of con-
cerns of possible distortions of competition “contrary to the interest of the EU.” In
particular, by requesting disproportionate contributions from other postal operators,
the compensation fund could lead to their foreclosure from the market where the
USO applies. In such a case, pursuant to the SGEI Framework,60 the Commission
will not hesitate to “examine whether such distortions can be mitigated by requiring
conditions or requesting commitments from the Member State” or, in the alterna-
tive, prohibit it altogether.
The Commission warned that it will ensure that compensation fund mechanisms
strike a right balance between the objectives of securing the USO and allowing fair
competition in the postal market. In a scenario of increasingly shrinking postal
markets, the application of the principles of nondiscrimination and proportionality
makes the compensation fund little effective for securing the USO, if at all. The
interpretation and application of the notion of “interchangeable services,” whose
providers might be called to contribute to the fund, could provide a solution in
the respect, “for example by spreading contributions as widely as possible” as in the
electronic communications sector. The other, self-explanatory, means to make
the fund effective is to reduce the costs to be compensated, by discontinuing the
elements of the USO that exceeds the needs of users.
References
Oxera (2007) Funding universal service obligations in the postal sector

Wik-Consult (2013) Main developments in the postal sector (2010–2013)
60
§ 52.
The Shape of Postal Inefficiency
Margaret M. Cigno, John P. Klingenberg, and Edward S. Pearsall
1 Introduction
Recent large declines in mail volumes have increased the intensity of many posts’
efforts to control costs by identifying and eliminating inefficiency. However, it is
often difficult to identify the scope of inefficiency in postal operations, let alone
determine the causes. In this chapter, we describe and apply a new method for
estimating the probability density function (PDF) of inefficiency within a large
sample of similar postal operations. This should assist postal operators in identify-
ing what operations are the most inefficient and where efforts to improve efficiency
should be concentrated.
Our method consists of a novel technique for fitting a stochastic frontier to cost
and production data. The method avoids assuming a specific analytic form for the
single-tailed PDF of the errors caused by inefficient operations. Instead, we derive
this PDF from the data by using a combination of econometric and operations
research techniques.
The shape of the fitted PDF can tell us a great deal about inefficiency. First, the
variation of the PDF allows us to distinguish between the effects of inefficiency and
the presence of noise in the data. Second, the weight of the PDF in the region
characterizing efficient operations reveals the strength (or weakness) of systemic
tendencies towards efficient operations. Third, the skewness of the PDF is a sign of
the effectiveness of management controls and incentives at encouraging efficiency
The views expressed in this chapter are those of the authors and do not necessarily represent the
opinions of the Postal Regulatory Commission (PRC).
M.M. Cigno • J.P. Klingenberg
United States Postal Regulatory Commission, Washington, DC, USA
E.S. Pearsall (*)
Independent Consultant, Alexandria, VA, USA
e-mail: espearsall@verizon.net

170 M.M. Cigno et al.
at the operating level. And forth, the PDF can be used to estimate the expected
inefficiency of individual operating units. In general, deriving the shape of the PDF,
rather than assuming it, should be a considerable help in identifying the causes of
postal inefficiency and designing programs to reduce it.
At first glance fitting a frontier function to postal data looks like a made-to-order
application for existing methods. Modern national posts preside over large numbers
of geographically dispersed semi-independent operating units such as delivery
offices and processing plants, all employing labor-intensive technologies and
assumed to be similarly managed. However, on the few occasions when operating
data from these units have been used to fit frontier cost and production functions,
the results have not led to a consensus on the best approach.
An early study of this kind was conducted by Horncastle et al. (2006) using
delivery office cost data collected for a single year by Royal Mail. More recently,
Cazals et al. (2012) have fit stochastic frontiers to panels that comprise observations
of UK delivery offices collected over a series of years. Several of the same
researchers (Cazals et al. 2014) have also fit frontiers to delivery data supplied by
La Poste. Pierleoni and Gori (2013) and Meschi et al. (2015) have applied several
distinct methods to fit frontiers to panel data for national postal operators. How best
to fit stochastic frontiers to postal data has also been studied using simulated data by
Harmon et al. (2010) and by Cazals et al. (2012), with results that are mostly
inconclusive.
The methods used by researchers to fit frontiers to production and cost data fall
into two general categories: parametric and nonparametric. The parametric
methods are variants of stochastic frontier estimation (SFE) for which the standard
references are Aigner et al. (1977), Kumbhakar and Lovell (2000), and Greene
(1990, 1993). SFE is essentially conventional econometrics applied to fit an equa-
tion with a two-part error. The error in observed production or cost is presumed to
be composed of a mix of noise and inefficiency. Specific assumptions are made
about the independence and form of the PDFs of these components, and the
stochastic frontier is then derived by maximum likelihood methods. All of the
extant fits of stochastic frontiers to postal data assume that noise and inefficiency
are independent, that noise is normally distributed with a zero mean, and that
inefficiency has a half-normal distribution.
The nonparametric methods are variants of data envelopment analysis (DEA), an
operations research approach to measuring efficiency that usually is credited to
Charnes et al. (1978). Standard DEA virtually ignores the contribution of noise to
the fitting error. All nonparametric methods fundamentally require that observed
production or cost be nearly noise-free, since the presence of noise favoring
efficiency will seriously distort the fit. This is not an assumption that can easily
be reconciled with stochastic treatments of error. Perhaps the best attempt to do so
is the m-frontier method invented by Cazals et al. (2002). The method has attractive
asymptotic properties, but, in practice, continues to depend upon the noise contri-
bution being small enough to be effectively averaged out when observations are
combined m-at-a-time. This is one of the nonparametric methods that Cazals
et al. (2014) apply to French postal data.
The Shape of Postal Inefficiency 171
Our approach can be described as SFE in reverse. We derive the form of the PDF
of the inefficiency component from the distribution of the residual errors left from a
conventional econometric fit. Our method of finding a PDF for the inefficiency left
in the residual errors is described in detail in the Appendix. In brief, the method is to
solve a Mixed-Integer Linear Program (MILP) to derive the PDF in the form of a
truncated spline function subject to constraints on its shape and smoothness. For
example, the PDF is assumed to be unimodal and without a downside tail. The
MILP is formulated and solved repeatedly with different proportions of the vari-
ance of the error assigned to noise and inefficiency until the best combination is
found. “Best” is the proportion that leads to the closest match to the observed
frequency distribution of the equation errors.
Using this method, we have derived and analyzed PDFs for several large
samples of United States Postal Service (USPS) data. These samples contain out-
puts, inputs, and costs for identical processing and delivery operations performed at
roughly the same time at many different points on the USPS network. A parametric
method is appropriate because the data for the operating units are incomplete and
contain potentially large noise contributions making nonparametric methods unat-
tractive. One of these applications, delivery office city-carrier route operations, is
described in detail to illustrate our method.
In Sect. 2, we lay out the estimation problem that we solve. The first step to
applying the method is a conventional econometric fit of a derived demand function
to a suitable sample. This step is described for the sample of delivery units in
Sect. 3. The PDF derived for the sample is presented in Sect. 4. In Sect. 5, we
demonstrate how the shape of the fitted PDF allows us to make several useful
inferences about the nature and causes of inefficiency at the USPS delivery units.
Section 6 summarizes our findings.
2 Background
A basic assumption in neoclassical microeconomic theory of cost and production

functions is that an enterprise conducts its business efficiently. Production functions
are derived under the assumption that the enterprise will produce its output of goods
and services using no more inputs than are technologically required. Cost functions
are derived under the somewhat stronger assumption that the enterprise will
produce its output at minimum cost. To fit a production or cost function, we
would ordinarily add an error term e to the function to represent external
shocks
and measurement error. The function would take the form y ¼ F xθ þ e, and the
unknown parameter vector θ would be estimated using an appropriate econometric
method.
Virtually all of the standard methods require the assumption that the mean value
of the error e is zero. However, inefficiency contributes to the error asymmetrically.
If we are fitting a production function, inefficiency can only decrease output; for a
cost function, it can only increase cost. Therefore, the error e possesses a non-zero
mean which is the expected contribution of inefficiency to observed production

or cost.
SFE offers a method of dealing with the mean inefficiency embedded in the
error. The error e is broken into separate components for noise u and inefficiency
v which are assumed to be independently distributed. Formally, e ¼ u þ ðv vÞ,
with E½e ¼ 0, E½u ¼ 0, and E½uv ¼ 0. v ¼ E½v denotes the expected value of the
inefficiency component of the error. The PDF of the noise component is assumed to
be a Normal distribution g(u) with a zero mean. The PDF of the inefficiency
component can be any distribution h(v) that is bounded at zero. If we are fitting a
production function, then hðvÞ 0 where v 0, and is zero otherwise; for a cost
function, hðvÞ 0 where v 0, and is zero otherwise. The Exponential distribution,
the truncated Normal distribution (particularly the half-Normal distribution), the
Log-normal distribution, and the Gamma distribution are all bounded PDFs that are
often used in SFE to represent h(v).
An added complication arises when dealing with samples that are panels. Then,
inefficiency is properly represented by more than one component. For example, if
the sample is composed of observations of operating units over multiple time
periods, we would represent v as composed of subcomponents associated with the
operating unit and other dimensions of the panel such as the time period of the
observations.1 Cazals et al. (2012) have explored different formulations of the
inefficiency component for postal data. We have ignored this complication, not
because we think it is unimportant, but to keep the applications of our new method
as basic as possible.
The PDF of the equation error, f(e), is related to the PDFs of u and v by an
integral equation. Under the independence assumption, the joint probability of
u and v is g(u)h(v). Assuming that we are fitting a cost function, we make the
substitution u ¼ e v þ v and integrate over 0 v 1 to obtain the PDF of the
error e2:
ð1
gðu ¼ e v þ vÞhðvÞdv ¼ f ðeÞ ð1Þ
0
The stochastic frontier estimate of θ is found by maximizing the likelihood of the

sample formed using f(e). This can usually be done by applying a suitable combi-
nation of analytical and numerical methods. The principal difficulty with this
approach is that the resulting estimate of θ depends critically on the form that is
assumed for h(v). In particular, the estimate of v is heavily influenced by which of
1
This is the SFE version of the random effects model for panel data.
2
For a production function with 1 v 0, the integral equation is
ð0
gðu ¼ e v þ vÞhðvÞdv ¼ f ðeÞ.
1
the PDFs cited above is chosen.3 Unfortunately, there is not a good conceptual basis
for making this choice.
Our approach is to avoid making any assumption of a specific form and, instead,
to derive the function that best fits Eq. (1) in a segmented form. To find h(v), we
start by estimating f(e) from the sample. This is done by fitting the production or
cost function using standard econometric methods and then using the residuals to
describe f(e). Fitting by standard methods such as ordinary least-squares (OLS) can
be attractive for several reasons. First, OLS is the Best Linear
Unbiased Estimator
(BLUE) of θ under very general conditions when F xθ is linear or can be
transformed into a linear form,4 second, the differences between the OLS estimate
of θ and a maximum likelihood estimate with a non-normal f(e) are almost always
trivial,5 and third, we do not have to make any assumption about the form of h(v) to
obtain an accurate histogram or smoothed kernel estimate of f(e) if we have a large
enough sample. Therefore, OLS gives us an estimate of the vector θ with desirable
properties except for an inhomogeneous additive constant which includes v.
The second step is to find h(v) by fitting Eq. (1) given f(e). Basically, our method
of estimation is to find the segmented representation of h(v) that, in combination
with the Normal distribution g(u), produces the best match for f(e). We do this by
solving an MILP to numerically fit the integral in Eq. (1). To formulate the MILP,
we represent h(v) generally as a truncated spline function subject to restrictions on
its shape and smoothness. Most of the restrictions on h(v) are imposed as linear
constraints within the MILP. However, the variance of h(v) enters the calculations
in a nonlinear way, so the MILP must be formulated and solved with this variance
assumed. Our method of estimation requires the solution of MILPs for different
assumed variances within a simple search plan. The search terminates when the
variance of the h(v) that best fits Eq. (1) is found. The estimate of h(v) in the form of
a tabulation of the distribution is extracted from the solution to the final MILP.
The estimate of h(v) may be used to calculate the expected efficiency of an
individual behavioral unit given the unit’s error e. For an individual unit t, the
conditional distribution of v, given et, is gðu ¼ et v þ vÞhðvÞ= f ðet Þ. The expected
value of the conditional distribution is obtained by numerically evaluating the
integral equation:
ð1

E vt et ¼ ½gðu ¼ et v þ vÞhðvÞ= f ðet Þvdv ð2Þ
0
3
All of the elements of the maximum likelihood estimate of θ are affected by the choice of a PDF;
and small except with respect to the elements of θ representing
however, the effects are secondary
the inhomogeneous part of F xθ .
4
This result is known as the Gauss–Markov theorem.
5
The maximum likelihood estimator is slightly more efficient than OLS when f(e) is non-normal
because it uses a little bit more information, namely the form of the PDF of e which is used to form
the likelihood function. However, the improvement in the efficiency of the maximum likelihood
estimator depends entirely upon having the correct PDF for f(e).
The conditional variance of vt is the integral equation:

h ð1
2 i 2
E vt E vt et ¼ ½gðu ¼ et v þ vÞhðvÞ= f ðet Þ v E vt et dv ð3Þ
0
Equations (2) and (3) are evaluated and used to make interval estimates of the
expected inefficiency of a single operating unit.
3 USPS Delivery Offices
We demonstrate our method of estimating the efficient frontier using USPS data
collected for over 1000 City Carrier Routes within a 2-month period. The routes
comprise all of the routes for 52 representative USPS delivery offices. To assess
efficiency at the level of the delivery offices, we have aggregated and averaged over
the routes the paid hours and workloads for each office and for each day during the
sample period. This produced a set of 2444 data points which were then screened to
eliminate 16 apparent outliers.
The carrier hours were recorded separately for in-office, street, and other time.
The carrier’s daily workload was divided into four categories: mailer sequenced
mail, delivery point sequenced mail, cased letters, and cased flats. The total number
of pieces of mail was also recorded and occasionally exceeded the mail in the four
categories. The difference consists primarily of unsequenced standard mail.
The hours recorded for the routes may understate the value of the hours for
which the carriers were actually paid for two reasons. First, when a carrier works
more than 8 h in a single working day he/she must be paid 1.5 times the regular rate
for the overtime. Second, carriers who have worked less than 8 h but at least 7 h and
one minute may be excused with their supervisor’s permission but are still credited
with a full 8-h day. This noncontractual arrangement is known as the 7:01 rule. We
have estimated and added to the reported hours half the overtime and all of the time
excused under the 7:01 rule each day on each route to obtain a rough estimate of the
carrier’s equivalent paid regular hours. The estimated overtime hours agreed
approximately with weekly overtime totals that were available for all of the city-
carriers at the delivery units in the sample. There was no way to check the 7:01 rule
hours; however, the total hours excused under the rule appear to be small compared
to the added overtime hours.
The sample clearly omits some information about the carriers and the properties
of the routes that should be included. The sample did not include the pay rates of the
carriers; nor do we know the number of delivery points on the routes, the miles that
the carriers travel, whether or not the carriers used a vehicle, or if the mailboxes are
located at curbside. Nevertheless, it remains possible to regard the effects of these
properties and the missing wage rate as part of the noise and fit a regression that
Table 1 Derived demand for city-carrier hours per route

Delivery offices
In equivalent regular hours per route
Reported hours + 1/2 overtime hours + 7:01 rule Working days from 8/3/2013 to 9/27/
hours 2013
Explanatory variable (regressor) Coefficient t-value Sample mean
Intercept 1.260 25.863 1.000
Day-of-the-week dummy variables
Tuesday 0.008 1.999 0.168
Wednesday 0.020 4.446 0.171
Thursday 0.008 1.789 0.171
Friday 0.004 0.979 0.171
Saturday 0.024 5.956 0.170
Volume proportions by mail category
Mailer sequenced mail 0.227 5.417 0.059
Delivery point sequenced (DPS) mail 0.128 4.122 0.665
Carrier-cased letters 0.669 9.573 0.033
Carrier-cased flats 0.122 4.156 0.235
Carriers with mailer sequenced mail
Proportion of days 0.019 2.486 0.268
Delivery unit size
In number of DU routes 0.022 13.374 2.812
Average carrier workload
In average daily volume (000) 0.117 28.898 7.610
Mean inefficiency Hours per day 0.144 R-squared Std. error d.f.
Inefficiency variance Var(v) 0.380 0.999 0.051 2415
relates equivalent regular hours to the carriers’ workload.6 This can be done with
surprising accuracy as shown in Table 1.
The regression equation is a derived demand function for regular carrier hours
per route with the carrier’s relative wage omitted. It has been fit in log-log form,
first, because this form allows us to conveniently interpret several of the estimated
coefficients as fixed elasticities, and, second, to suppress heteroskedasticity that
would otherwise be evident in the errors. In most respects, the signs and magnitudes
of the estimated coefficients are about what we would expect. Carriers have shorter
workdays on Wednesdays and Saturdays; the mail can be delivered more quickly if
it is already sequenced and does not need to be cased in-office; carriers take longer
when they must cope with separate trays of presorted standard mail; and, the returns
to scale at the route level are quite high.
6
The critical assumption that is made about the omitted variables when the equation is fit by least
squares is that the missing variables are uncorrelated in the sample with the explanatory variables
that are included. If this assumption is wrong, then the estimate of the parameter vector θ will be
biased and inconsistent. It is not necessary that the omitted variables have normal distributions.
However, our method does assume that the omitted variables, along with other sources of noise, do
not contribute to v and leave a combined u that is normally distributed.
The coefficient for the average carrier’s workload is the elasticity of equivalent
regular hours with respect to the carrier’s average daily workload. If the daily
volume is doubled, the predicted increase in a carrier’s equivalent regular hours is
only 11.7 %. This relatively low elasticity is primarily explained by our inclusion of
overtime and 7:01 rule time in the definition of equivalent regular hours. These
additional hours severely dampen the response of a carrier’s work hours to the
carrier’s workload.
An unusual structural feature of our derived demand equation is the inclusion of
the number of routes as a proxy for the size of the delivery unit (DU). We found that
a carrier’s work hours are significantly affected by the size of the office that the
carrier works from. The estimate of the coefficient for the number of DU routes
reveals that each doubling of a delivery unit’s size will increase the time taken by its
carriers by 2.2 %. This result has several possible explanations. The number of DU
routes may reflect the size of the geographic area served by the DU (and thus the
travel time needed to reach delivery areas); sorting mail to routes may take longer
when there are more routes; and, large delivery offices may just be harder to
manage efficiently than small ones.
Figure 1 is a scatter plot of the sample. Each point represents the average of the
adjusted hours and workloads for the routes of a single delivery unit on a single day.
Adjusted hours are the delivery unit’s average equivalent regular hours adjusted for
differences between the delivery unit and the sample average, except for the
average carrier workload. The adjustments were made using the estimated coeffi-
cients and sample means displayed in Table 1.
Also represented in Fig. 1 is the fitted derived demand equation. The curved fine
line is the average adjusted hours per route predicted for each delivery unit. The
difference between a point and its corresponding point on the curved fine line is the
error e. This error has three components as described in Sect. 2, e ¼ u þ v v. Our
Adjusted Hours versus Workload

12
Adjusted Hours per Route per Day
11
10
6
0 1,000 2,000 3,000 4,000 5,000 6,000
Average Number of Pieces per Route per Day
Fig. 1 Average city carrier daily hours (adjusted) versus workload

method produces an estimate of v that we deduct from a delivery unit’s predicted

hours to obtain a point on the efficient frontier. For this example, our method yields
a mean inefficiency of 0.144 h per day per route. This translates over all city carrier
routes into approximately $275 million a year. The heavy curved line in Fig. 1
shows how the fitted derived demand equation shifts downwards to become the
efficient frontier when we deduct v from the adjusted hours predicted by the fitted
equation.
4 The Method of Estimation
The PDF of the errors in adjusted hours was derived by conventional fitting
methods from the residuals of the derived demand equation. f(e) was fit to the
residuals in standardized form so it has a zero mean and a variance of one. g(u) is
the PDF of the noise in the sample and is assumed to be a Normal distribution with a
zero mean and variance σ 2u . Nothing is known for certain about the inefficiency PDF
h(v) except that hðvÞ > 0 for v 0 and hðvÞ ¼ 0 for v < 0. Therefore, h(v) has a
non-negative mean and variance σ 2v . However, it is not unreasonable to make
several additional assumptions about h(v). For example, h(v) is assumed to be
uni-modal, single-tailed, and smooth. The assumptions we make about h(v) are
all described precisely in the Appendix. Noise and inefficiency are independent;
therefore, σ 2e ¼ σ 2u þ σ 2v ¼ 1.
Figure 2 shows how well our method of estimation works with the sample of
delivery units. The standardized residuals from the fit of the derived demand
equation have been used to compile a histogram that represents f(e). The observed
Equation Error Probability Densities
-4.00 -3.00 -2.00 -1.00 0.00 1.00 2.00 3.00 4.00 5.00 6.00
Standard Deviation Units
Observed Frequency Smoothed f(e) LP Solution Fitted f(e) Normal Distribution
Fig. 2 Frequency distribution and fitted PDF of the equation error delivery units
frequencies of the residuals in each of 46 equal intervals are plotted as black

diamonds in the figure. Although the distribution of the errors is roughly bell
shaped, it is certainly not a Normal distribution. Clearly, the black diamonds do
not show much of a tendency to fall on the PDF of the standardized Normal
distribution shown as a dashed curve in Fig. 2. The histogram is also noticeably
skewed towards positive values of e, whereas the Normal PDF is symmetric.
Although our method of estimation can be applied using the histogram to
represent f(e), a somewhat better continuous representation of f(e) can be derived
using kernel smoothing methods. The continuous curve in Fig. 2 was derived from
the 2428 screened residuals using a Normal kernel. With this many data points, the
smoothing virtually eliminates the sampling variations that remain evident in the
histogram. The smoothed f(e) was employed as the right-hand side of the integral
equation when we solved the MILP to derive h(v). The left-hand side of the integral
equation is the finely dashed curve in Fig. 2. The continuous and finely dashed
curves in Fig. 2 are almost identical. Therefore, our method of estimation has found
a combination of a noise PDF g(u) and an inefficiency PDF h(v) that explains the
skewed pattern of the residual errors quite accurately.
These PDFs are shown mean centered in Fig. 3. g(u) is the dashed “bell” curve; h
(v) is the sharply peaked finely dashed curve. The variance of h(v) is 0.38 so the
variance of g(u) is 0.62. If we are to associate the shape of h(v) with one of the
theoretical distributions that are often assumed to describe inefficient behavior in
conventional SFE, the best choice would certainly be the exponential distribution.
Also shown in Fig. 3 is the lower bound to h(v). v is the horizontal distance from
the lower bound to the vertical axis located at v ¼ 0. This distance is 0.51 standard
deviation units. This equates to 0.144 h per day. The finely dashed curve in Fig. 3
has been extended to show how h(v) is truncated at the lower bound. h(v) is zero for
Equation Error Component Densities
-4.00 -3.00 -2.00 -1.00 0.00 1.00 2.00 3.00 4.00 5.00 6.00
Standard Deviation Units
Efficiency Distribution h(v) Noise Distribution g(u) Lower Bound h(v) Histogram h(v)
Fig. 3 Fitted PDFs for noise and inefficiency, mean inefficiency delivery units
values of v below v. The black diamonds in Fig. 3 show how the truncation affects
the histogram of h(v). To the right of the bound, the histogram and the finely dashed
curve correspond. However, the height of the histogram for the interval intersected
by the bound is severely reduced by the truncation.
The efficient frontier of the derived demand equation is obtained by subtracting
the average inefficiency v from the equivalent regular hours per route per day
predicted by the demand equation of Table 1. This causes the uniform downwards
shift in the curve representing predicted hours in Fig. 1.
The heavy curved line in Fig. 4 tracks the expected values of the conditional
distributions of vt given et for each observation in the sample, i.e., for each delivery
office for each day of the sample period. The values of the standardized errors
e appear along the horizontal axis; the values of the corresponding expected

efficiency components v are measured vertically. The values of E vt et were
calculated for each point by numerically evaluating the integral in Eq. (2) using
the smoothed distribution for f(e) and the fitted distributions for g(u) and h(v). The
variances of the conditional distributions were similarly calculated using Eq. (3)
and have been employed
to erect two-sigma confidence intervals around the
estimates of E vt eit . These are shown as the two lighter curved lines in Fig. 4.
Roughly, we can be 95 % confident that the true inefficiency lies between the two
lighter curved lines.
Figure 4 exhibits the limitations of using expected inefficiency asa measure
of
individual performance. The wide confidence interval around E vt et does not
statistically preclude the possibility that a delivery unit is efficient unless the error
e is well over two. So, in this instance, it is quite risky to use expected values
calculated using Eq. (2) to identify inefficient delivery units except when the error
e is quite large.
Expected Inefficiency
5.00
Inefficiency (v) Standard Deviation Units
4.00
3.00
2.00
1.00
0.00
-4.00 -3.00 -2.00 -1.00 0.00 1.00 2.00 3.00 4.00 5.00 6.00
-1.00
-2.00
Error (e) Standard Deviation Units
Fig. 4 Individual performance delivery units

5 The Shape of Inefficiency
Ultimately, the reason for isolating the PDFs for noise and inefficiency is to identify
the importance of inefficiency in explaining observed costs or productivity, to learn
what we can about the causes of inefficiency from the shape of h(v), and to identify
steps that postal operators may take to effectively reduce inefficiency at the
operating level. To this end, we have devised three indices with benchmark values
of one that serve to isolate several important characteristics of h(v).
The first and simplest of the indices is the variance ratio σ 2v /σ 2u . This ratio
describes the relative importance of inefficiency to noise in the sample. For the
adjusted hours of delivery units, this ratio is only 0.613. A ratio of one would mean
that inefficiency and noise are equal contributors to the errors. But our method of
estimation reveals that the contribution of inefficiency is small relative to the noise
in the sample of delivery units. This is not too surprising when we recall that the
data failed to include either a carrier’s wage rate or virtually any helpful descriptors
of a carrier’s route. The low variance ratio indicates that the sample should not be
used to evaluate the efficiency of individual delivery units. Analytical methods such
as data envelopment and m-frontier fitting should also be avoided as these methods
depend on the noise component being small.
The second index, which we call the “target” index, is the mean of h(v) within
the range [ v, 0], divided by v=2. This index measures how far h(v) favors
efficiency within roughly the most efficient half of the sample. If the target index
is greater than one, then h(v) has a shape that favors the bound v; if the target index is
less than one, then the shape favors the v ¼ 0 axis. The index tells us whether or not
the more efficient units in the sample are drawn systemically towards more efficient
operation. If the index is greater than one, then improving efficiency is largely just a
matter of obtaining better compliance with existing standards since better compli-
ance will shift h(v) towards its lower bound. On the other hand, an index less than
one shows that better compliance can shift h(v) towards zero. Paradoxically, the
existing standards may be leading the operating units towards less efficient opera-
tion. When this happens, inefficiency is partly systemic. The central tendency of h
(v) tells us that operating units are being drawn by existing incentives and standards
to promote a mode of operation that is inefficient The target index for our sample of
delivery units is 1.137. This value is well within the comfort zone for the target
index. It indicates that the managers of USPS delivery units generally have the
information and institutional incentives that lead correctly towards efficient city
carrier in-office and delivery operations.
The third index is designed to expose issues of control. When a business
organization fails to effectively control its operating units, one of the expected
results is an increase in the proportion of units that are highly inefficient. This will
affect the shape of h(v) by increasing the length and thickness of the PDF’s right-
hand tail, which will be manifested by an increase in a measurable property of the
PDF, its skewness. Our “control” index is Pearson’s moment coefficient of skew-
ness for h(v) divided by the coefficient of skewness for the exponential distribution.
Therefore, a control index less than one means that h(v) is less skewed than a
benchmark exponential distribution with the same mean and variance. Conversely,
a control index greater than one tells us that the right-hand tail of h(v) is longer and
thicker than that of the comparable exponential distribution. The control index for
the sample of delivery units is 1.436. We regard this as evidence that the sample
contains a high proportion of delivery units that are abnormally inefficient. USPS is
apparently not doing particularly well at promoting efficient operations at all of its
delivery units.
6 Conclusion
Altogether, we have applied our new method of estimation to ten samples describ-
ing delivery, sorting and allied operations at USPS facilities. The case exhibited in
this chapter is characteristic of the results that we have obtained for all of these
samples. Our method has also proven reliable extracting the PDF of the inefficiency
component when we simulate samples using the PDFs most commonly used to
represent inefficiency in conventional SFE.
In general, these results demonstrate that we have found a useful way to fit
stochastic frontiers to postal data. We do not need to assume that the effects of
inefficiency take a specific analytic form. The shape of a PDF fitted by our method
is determined by the data subject only to very general restrictions on its shape and
smoothness. This is an important advantage over conventional SFE because it
allows us to draw inferences from the shape of the estimated PDF that cannot be
drawn when the shape is assumed.
We show, with our example, that our method lets us distinguish between noise
and inefficiency as sources of error in postal cost or productivity data. The shape of
the estimated PDF can also tell us if inefficiency is likely to have systemic or
nonsystemic causes, and if the enforcement of performance standards is effective
within the organization. Finally, we can derive interval estimates of the expected
inefficiency of individual operating units. All of this should be helpful in efforts to
control postal costs by reducing inefficiency.
Appendix
In this Appendix, we provide the technical details of our method for estimating h(v),
the PDF of the inefficiency component of the error e. The PDF f(e) of the error e is
assumed to be given either in the form of a histogram or as the result of smoothing
the sample of residual errors using a suitable technique such as kernel smoothing.
The objective of the fitting method is to find the h(v), with several assumed
properties that leads to the best fit of the integral equation (1). This fit is obtained as
a spline function, i.e., as a sequence of continuous parabolic sections with contin-

uous slopes.
The core of our method is a Mixed-Integer Linear Program (MILP) that is solved
to fit Eq. (1) for a given variance σ 2v . A sequence of MILPs are formulated and
solved for different values in the range 0 σ 2v 1. This is done repeatedly within a
simple single-parameter search plan until a value is found that provides the best fit
of Eq. (1). The MILP that we solve repeatedly for different assumed variances was
constructed within an Excel spreadsheet and solved using a commercially available
solver, “What’s Best,” operating as an add-in within Excel.7
“Best fit” can be defined operationally in different ways that lead to similar but-
not-identical estimates of the variance σ 2v . The definitions that we have
experimented with are: 1) the minimum sum of the absolute deviations of the fitted
f(e) (this is the objective quantity of the MILP; 2) the minimum mean square
deviation of the fitted f(e) from the histogram derived from the residuals; 3) the
minimum mean square deviation of the fitted f(e) from the PDF derived from the
sample errors by smoothing; 4) the maximum log likelihood of the fitted f(e) when
applied to the sample of residual errors e. The results in the text were found by
minimizing the sum of the absolute deviations, i.e., by choosing the variance σ 2v that
yielded the best fit of the MILP.
We begin by showing how we have numerically approximated Eq. (1). This is
followed by a detailed description and explanation of the MILP. The next section
describes the numerical approximations made to calculate E vt et from Eq. (2) and
h 2 i
E vt E vt et from Eq. (3). The Appendix ends with a presentation and
explanation of the indices we have invented to describe the properties of h(v).
Equation (1)
ð
1
The integral equation gðu ¼ e v þ vÞhðvÞdv ¼ f ðeÞ is numerically approxi-

0
mated by tabulating the PDFs and, then, approximating the tabulated entries.
f(e) the observed PDF of the standardized error e ¼ u þ v v.
g(u) the normal PDF of the noise component u.
h(v) the PDF with mean v of the inefficiency component v.
7
“What’s Best” is sold in several versions by LINDO Systems Inc., 1415 Dayton St., Chicago, IL,
USA 60642. Their web address is www.lindo.com. We used the “commercial” version of the linear
program solver with integer programming capabilities. What’s Best typically took about 20 s to
solve one MILP.
h(w) the PDF of the mean-centered inefficiency component w ¼ v v,

hðvÞ ¼ hðw þ vÞ and hðwÞ ¼ hðv vÞ.
i, j ¼ 1, . . . , n indices of intervals of width Δ for tabulating the PDFs of e and w.
ei, wj midpoints of the intervals i and j for e and w.
f(ei), i ¼ 1, . . . , n the tabulated representation of f(e). f(ei)Δ is approximately the
probability that e lies within the interval ei Δ=2 e ei þ Δ=2.
σ 2u , σ 2v variances of f(u) and f(v). σ 2u þ σ 2v ¼ 1 for a standardized error e. σ 2v is also
the variance of h(w).
GðÞ the cumulative standard
Normal distribution.

ðei w j þΔ=2Þ ðei w j Δ=2Þ
Ni j ¼ G σu G σu , the probability element of g(u) over the
interval j that includes u ¼ ei w j .

x j ¼ h w j , j ¼ 1, . . . , n the tabulated representation of h(w). xjΔ is approximately
the probability that x lies within the interval x j Δ=2 x x j þ Δ=2.
X n
yi ¼ x j N i j , i 1, . . . , n the numerical approximation of f(e) in tabulated form.
j¼1
This is the fitted f(e) and differs from the tabulated f(e) by an approximation
error.
The variance σ 2v enters the equation for yi nonlinearly because it is used to calculate
σ u which enters nonlinearly in the arithmetic for the probability elements Nij.
The MILP
Minimization of the Approximation Error
The objective of the MILP is to find the values of the variables xj for j ¼ 1, . . . , n
representing the PDF h(w) to minimize the sum of the absolute values of the
approximation errors in the tabulation of f(e).
si 0 and ti 0 the positive and negative components of the approximation error in
interval i, so that f ðei Þ ¼ yi þ si ti . The absolute value of the approximation
error is the sum si þ ti .
X n
ðsi þ ti Þ the objective quantity (to be minimized) of the MILP.
i¼1
The Known Properties of h(w)
Some of the statistical properties of h(w) are known a priori and are imposed
directly on the variables, xj.
x j 0 for j ¼ 1, . . . , n h(w) is non-negative for any w.

X n
x j Δ ¼1 the integral of h(w) is one.
j¼1
X
n
x j Δ w j ¼ 0 the mean of h(w) is zero because w is mean centered.
j¼1
X
n
x j Δ w2j ¼ σ 2v the variance of h(w) is equal to σ 2v which is given for the MILP.
j¼1
X
n
x j Δ w3j ¼ E½e3 the skewness of h(w), as measured by the third central
j¼1
moment, is the same as the third central moment of f(e) because g(u) is
symmetric.
Spline Function Representation of h(w)
The PDF of w is represented by a continuous and smooth second-degree polynomial

spline function. The upper tail is assumed to vanish for w ¼ wn .
αj, βj, γ j parameters of the polynomial representing h(w) within the interval j, hðwÞ
¼ α j w2 þ β j w þ γ j for w j Δ=2 w w j þ Δ=2. These parameters are free
variables within the MILP, i.e., they are not subject to the usual non-negativity
condition.
2 2
Continuity: α j w j þ Δ=2 þ β j w j þ Δ=2 þ γ j ¼ α jþ1 w jþ1 Δ=2 þ β jþ1

w jþ1 Δ=2 þ γ jþ1 for j ¼ 1, . . . , n 1:

Slope continuity: 2α j w j þ Δ=2 þ β j ¼ 2α jþ1ðw jþ1 Δ=2Þ þ β jþ1 for
j ¼ 1, . . . , n 1.
Upper tail: αn ðwn þ Δ=2Þ2 þ βn ðwn þ Δ=2Þ þ γ n ¼ 0 and
2αn ðwn þ Δ=2Þ þ βn ¼ 0.
Lower Bound of w
Inefficiency is non-negative, i.e., v 0. Therefore, w ¼ v v has a lower bound v.

This means that h(w) may be truncated at w ¼ v since hðwÞ ¼ 0 for w < v. The
lower bound, if it is present, is accommodated by disconnecting h(w) from its spline
function representation in intervals that lie entirely below v.
IZj (“I Zero”) a discrete [0, 1] integer variable identifying the intervals for which h
(w) is zero. IZ j ¼ 1 for x j ¼ 0 and IZ j ¼ 0 for x j 0, for j ¼ 1, . . . , n.
zj truncation adjustment, z j 0 in the interval j for which w j Δ=2 v w j

þΔ=2 and z j ¼ 0 otherwise.
k a large positive number, e.g., k ¼ 1000.

z1 ¼ 0, z j k IZ j1 IZ j for j ¼ 2, . . . , n the truncation adjustment is zero
except in the first interval for which xj is permitted to be non-zero.
The estimate of v is retrieved from the solution of the MILP by determining the
first (left-most) interval for which IZ j ¼ 0, and then using the midpoint of the

interval, wj, to calculate v from the formula v ¼ w j Δ=2 þ z j =x j Δ.
h(w) Above the Lower Bound
The spline function representation of h(w) is switched on at the lower bound v by

two sets of matching constraints that have the effect of laying the values of xj at the
midpoints of the spline function polynomial segments.
x j α j w2j þ β j w j þ γ j z j kIZ j for j ¼ 1, . . . , n and

x j α j w2j þ β j w j þ γ j z j kIZ j for j ¼ 1, . . . , n
The [0, 1] variables IZj are used as switches. When IZ j ¼ 1, the pair of con-
straints for interval j are switched off and the variable xj is forced to zero; when
IZ j ¼ 0 the pair of constraints for interval j are switched on and the variable xj is
placed on the polynomial x j ¼ α j w2j þ β j w j þ γ j at the midpoint of the interval
(if zj ¼ 0) or at a distance zj below the midpoint of the polynomial x j ¼ α j w2j þ β j
w j þ γ j z j (if z j > 0).
Smoothing h(w)
h(w) is smoothed by restricting the values of the parameters αj in the polynomial

segments of the spline function above the lower bound. This limits the rate of
hðwÞ 2
change of the slope of h(w) within each segment because d dw 2 ¼ 2α j .
α j 0:199471=σ 2v kIZ j for j ¼ 1, . . . , n. With these constraints h(w) will not
turn downwards in its concave region more rapidly than a Normal distribution
with variance σ 2v turns downwards at its mean.
α j 0:367879=σ 2v þ kIZ j for j ¼ 1, . . . , n. With these constraints h(w) will not
turn upwards in its convex region more rapidly that an exponential distribution
with variance σ 2v turns upwards at its mean.
h(w) Is Unimodal
h(w) is allowed to attain its maximum value only once. This restriction is imposed
by allowing the slope of the PDF to change sign from positive to negative no more
than once above the lower bound.
INj (“I Negative”) a discrete [0, 1] integer variable identifying the intervals for
ðwÞ dhðw j Þ dhðw j Þ
which dhdw is negative. IN j ¼ 1 for dw j 0 and IN j ¼ 0 for dw j 0 for
j ¼ 1, . . . , n.
2α j w j þ β j kIN j for j ¼ 1, . . . , n, and 2α j w j þ β j k 1 IN j for
j ¼ 1, . . . , n.
The [0, 1] variables INj are switches that are set on or off depending upon the
sign of the slope of the polynomial segments of the spline function at the midpoints
of the intervals. When IN j ¼ 1 the slope 2α j w j þ β j cannot be positive; when IN j
¼ 0 the slope 2α j w j þ β j cannot be negative.
IN jþ1 IN j for j ¼ 1, . . . , n 1 The sign of the slope may change from positive to
negative no more than once as j goes from 1 to n.
Upper Tail of h(w)
h(w) is allowed to have no more than one inflexion point above its lower bound at
which the shape of the PDF transitions from concave to convex. This allows h(w) to
have a smoothly diminishing upper tail.
ICj (“I Concave”) a discrete [0, 1] integer variable identifying the intervals for
2
hðwÞ d 2 h ðw j Þ d 2 h ðw j Þ
which d dw 2 is negative. IC j ¼ 1 for dw2 0 and IC j ¼ 0 for dw2 0 for
j j
j ¼ 1, . . . , n.
α j k 1 IC j for j ¼ 1, . . . , n, and α j kIC j for j ¼ 1, . . . , n.
The shape of h(w) within an interval matches the shape of the polynomial
segment of the spline function within the interval. The second derivative of the
polynomial for the interval j is just 2αj. So the concavity or convexity of h(w) within
the interval is determined by the sign of αj.

IC j IC jþ1 k 1 IN j for j ¼ 1, . . . , n 1 and INn ¼ 0. A transition from
concave to convex may occur no more than once in the region, where h(w) may
be negatively sloped. And, at its extreme upper tail, h(w) is weakly convex.
Lower Tail of h(w)
h(w) is not allowed to have a gradually diminishing lower tail.

α j α jþ1 þ kIN j for j ¼ 1, . . . , n 1. The slope of h(w) is nonincreasing in the
region where the slope is non-negative.
Equations (2) and (3)
The integrals in Eqs. (2) and (3) are numerically approximated in the same way that
we approximate the integral in Eq. (1). First, we make the substitution w ¼ v v, to
obtain integral equations that are functions of h(w) rather than h(v). For Eq. (2):
ð1

E vt et ¼ ½gðu ¼ et v þ vÞhðvÞ= f ðet Þvdv
0
ð1
¼ ½gðu ¼ et wÞhðwÞ= f ðet Þwdw þ v
0
and for Eq. (3):

h ð1
2 i 2
E vt E vt et ¼ ½gðu ¼ et v þ vÞhðvÞ= f ðet Þ v E vt et dv
0
ð1
¼ ½gðu ¼ et wÞhðwÞ= f ðet Þw2 dw
0
As with Eq. (1), we denote GðÞ the cumulative standard Normal distribution. For an
individual unit t, we have

ðet w j þΔ=2Þ ðet w j Δ=2Þ
Nt j ¼ G σu G σu , the probability element of g(u) over the
interval j that includes u ¼ et w j .
The expected value of v given et is obtained by inserting the values of xj for j
¼ 1, . . . , n from the solution to the MILP in the equation
X n
E vt et ¼ x j w j N t j þ v. For the variance of vt, we substitute into the equation
j¼1
h 2 i Xn
E vt E vt et ¼ x j w2j N t j .
j¼1
Properties of f(v)
We have devised three indices with the intended purpose of isolating the properties
of h(v) that are relevant for revealing the nature and causes of inefficiency. h(v) is
just h(w) shifted downwards by the distance v, so the important properties of the two
PDFs are mostly the same.
Variance ratio: σ 2v /σ 2u The variance ratio is an index of the relative importance of
inefficiency to noise in explaining the PDF of the errors f(e).
ð0 ð0
Target index: 2 hðwÞwdw= v hðwÞdw This index is computed by numeri-
v v
cally evaluating the integrals as done with Eqs. (1)–(3). The index describes the
location of the mean value of h(w) in the efficiency region v w 0. If the
mean is centered in the region, the index equals one. If the mean favors the lower
bound, then the index exceeds one. If the mean favors the zero axis, then the
index is less than one. The index is used to determine the aim point of the
relatively efficient operating units in a sample. The closer the index is to two, the
better the central tendency of h(w) leads operating units to favor efficient
operation.
Control index: 0.5E[w3]/σ 3v This index compares the third moment of a fitted h(w) to
the third moment of an exponential distribution with the same variance. When
the index exceeds one, h(w) is more skewed than the exponential distribution.
This is an indication that the upper tail of h(w) is comparatively heavy. The
index is used to determine whether inefficiency is the result of highly inefficient
operations by a relatively small number of operating units.
References
Aigner D, Lovell CAK, Schmidt P (1977) Formulation and estimation of stochastic frontier
production function models. J Econom 6:21–37
Cazals C, Florens J-P, Simar L (2002) Nonparametric frontier estimation: a robust approach. J
Econom 106:1–25
Cazals C, Dudley P, Florens J-P, Jones M (2012) A panel data analysis of inefficiency and
heterogeneity in the postal sector. In: Crew MA, Kleindorfer PR (eds) Multi-modal competi-
tion and the future of mail. Edward Elgar, Cheltenham, pp 109–123
Cazals C, Florens JP, Roset A (2014) Non parametric estimation of the efficiency frontier of the
french delivery offices. Paper presented at the 22nd conference on postal and delivery eco-
nomics, CRRI, Frescati, Italy
Charnes A, Cooper WW, Rhodes E (1978) Measuring the efficiency of decision making units. Eur
J Oper Res 2:429–444
Greene WH (1990) A gamma-distributed stochastic frontier model. J Econom 46:141–163
Greene WH (1993) The econometric approach to efficiency analysis. In: Fried H, Lovell CAK,
Schmidt S (eds) The measurement of productive efficiency: techniques and applications.
Oxford University Press, Oxford
Harmon G, Koevoets W, Requejo A, van der Merwe E, Waghe N (2010) The effect of falling
volumes on traditional efficiency analysis. In: Crew MA, Kleindorfer PR (eds) Heightening
competition in the postal and delivery sector. Edward Elgar, Cheltenham, pp 135–148
Horncastle A, Jevons D, Dudley P, Thanassoulis E (2006) Efficiency analysis of delivery offices in
the postal sector using stochastic frontier and data envelopment analysis. In: Crew MA,
Kleindorfer PR (eds) Liberalization of the postal and delivery sector. Edward Elgar, Chelten-
ham, pp 149–164
Kumbhakar S, Lovell K (2000) Stochastic frontier analysis. Cambridge University Press,
Cambridge
Meschi M, Pierleoni MR, Gori S (2015) Advanced semi-parametric and parametric methods to
assess efficiency in the postal sector. In: Crew MA, Brennan TJ (eds) Postal and delivery
innovation in the digital economy. Springer International, Cham, pp 207–217
Pierleoni MR, Gori S (2013) Efficiency analysis postal operators: comparison between the United
States and Europe’. In: Crew MA, Kleindorfer PR (eds) Reforming the postal sector in the face
of electronic competition. Edward Elgar, Cheltenham, pp 261–276
Differentiated Pricing of Delivery Services
in the e-Commerce Sector
Claire Borsenberger, Helmuth Cremer, Philippe De Donder,

and Denis Joram
1 Introduction
Pricing strategies for parcels delivery from e-commerce remain a hot topic for
postal and parcel delivery operators. As shown by Borsenberger (2015), the
e-commerce sector is subject to concentration trends, due to a fierce price compe-
tition between retailers, the existence of increasing returns to scale in e-commerce
activity, and the importance of retailers’ reputation to attract consumers.
This phenomenon is reinforced by the development of “marketplaces,” which
represent virtual intermediate platforms between retailers and e-consumers. Mar-
ketplaces have developed a win-win intermediation model. Consumers enjoy a
greater variety of products (long-tail) and affiliated merchants take advantage of
the marketplace’s ability to generate huge online traffic and to provide a powerful
C. Borsenberger
The Regulation and Studies Department, Direction of Regulation and Institutional and
European Affairs, Groupe La Poste, Paris, France; Laboratoire d’Economie d’Orléans (LEO)
and University François Rabelais, Tours, France
H. Cremer (*)
Department of Economics, Toulouse School of Economics (IDEI, GREMAQ and IuF),
Toulouse, France
e-mail: helmuth.cremer@tse-fr.eu
P. De Donder
Department of Economics, Toulouse School of Economics (IDEI and GREMAQ-CNRS),
Toulouse, France
D. Joram
The Regulation and Studies Department, The Direction of Regulation and Institutional and
European Affairs, Groupe La Poste, Paris, France

192 C. Borsenberger et al.
showcase.1 Affiliation to a marketplace presents also some drawbacks for retailers.

These include fierce price competition leading to reduced margins, the dilution of
their own identity and brand to the benefit of the marketplace, and in a worst case
scenario the creation of a relation of dependence.
The concentration phenomenon raises specific concerns for parcel delivery
operators, as input providers for retailers. Big retailers could have substantial
power to negotiate attractive commercial terms for the provision of parcel delivery
services in a market characterized by fixed costs and returns to scale. Specifically,
this may lead to volume-discount pricing schemes. This trend is exacerbated by the
fact that the major marketplaces provide delivery services to their affiliated mer-
chants.2 They become a sort of parcel aggregators, maximizing the volume of
parcels provided to delivery operators in order to increase their quantity discounts.3
To put further pressure on parcel delivery operators, big retailers like Amazon are
developing their own delivery network in dense areas, threatening to bypass
traditional delivery operators.
In this chapter, we examine the link between the delivery rates charged by postal
and other parcel delivery operators and the e-commerce market structure. What is
the impact of the existence of a marketplace (vs. the presence of independent
retailers) on the delivery rates and on the economic surplus? Could a delivery
operator prevent the development of a marketplace with an appropriate pricing
strategy? Which is the best situation from the point of view of the various economic
agents (retailers, delivery operators, final consumers)?
To answer these questions, we develop a formal model, in Sect. 2, in which we
consider a stylized e-commerce sector with a single parcel delivery operator and
two differentiated retailers, indexed 0 and 1. Retailer 0 is a “big” retailer who also
operates a marketplace platform, which sells retail and delivery services to other
firms.
1
The activity of marketplaces is growing in all countries where e-commerce is well developed. In
2013, two million retailers were affiliated to Amazon’s marketplace around the world, selling more
than one billion items. In France, according to Oxatis (2014), 32 % of e-retailers sold their goods
through marketplaces in 2013. According to the FEVAD (2014), the volume of sales realized in
marketplaces increased by 42 % in the last quarter 2013 and represented 16 % of the global activity
of these e-retailers. The five first most visited e-commerce sites in France were marketplaces.
2
For example, Amazon proposes to its affiliated merchants the service “Fulfilment by Amazon”
(FBA). Merchants pay fees for the various services provided by Amazon: handling the order,
picking and packing products, shipping the order (fees depending on the parcel weight and size and
the value of order). Rakuten.com offer a similar storage and shipping service to merchants
affiliated to its marketplace: Rakuten Super Logistics (RSL).
3
In other words, marketplaces are to some extent similar to consolidators in the letter market. Both
structures undermine the ability of the parcel delivery operator to apply nonlinear pricing
(of which volume discounts are a special case), unless it can differentiate tariffs between market-
places (or consolidators) and “regular” customers. In the parcel market such a differentiation
appears to be rather difficult, in particular because many marketplaces have emerged as extension
of big retailers, a development of which Amazon is a prime example.
Differentiated Pricing of Delivery Services in the e-Commerce Sector 193
We assume that all items are delivered by the parcel delivery operator. The case
where retailer 0 has a bypass technology and can deliver some of its parcels without
using the parcel delivery operator is briefly discussed in the conclusion. Retailer
1 can sell independently or via the other retailer’s marketplace. When it sells
independently, it uses the delivery services provided by the parcel delivery operator
and pays the rate it charges. In case of marketplace affiliation, it pays a fee to the
other retailer who takes care of parcels delivery. Joining the marketplace has other
implications. It reduces the degree of product differentiation (so that competition
intensity increases), and it increases the willingness to pay for retailer 1’s product.
This is because the retailer 1 benefits from the reputation of retailer 0’s marketplace.
We study the case where the delivery operator sets the rates charged to each of
the two retailers to maximize profits and also look at the Ramsey solution (maxi-
mization of welfare subject to a break-even constraint). All variables are evaluated
in the induced subgame perfect equilibrium of the price competition game played
by the retailers. Depending on the fee charged by the marketplace, operator
1 decides to join or not. In the last stage, the retailers then compete in prices either
as independent retailers (Subgame I) or as retailers using the same marketplace
(Subgame M).
In Sect. 3, we determine analytically the solution to these games considering as
reference scenario the case where all the players maximize their profits and the case
where the parcel delivery operator maximizes social welfare. While we adopt the
simplest possible model which is consistent with the main stylized features of the
underlying problem, the analytical solutions are often quite complex. The interpre-
tation is interesting in itself, but to obtain sharper conclusions we resort to numer-
ical simulations in Sect. 4. Section 5 concludes.
2 Model
Consider an electronic retail market consisting of two retailers located at 0 and 1 of

the Hotelling line. Consumers are distributed over this line, with z 2 ½0; 1. Let G(z)
denote the distribution function and g(z) the density. The Hotelling specification is
the simplest way to represent horizontal differentiation. In our setting, z is not
meant to describe a geographical location but rather a parameter characterizing the
individuals’ preferences across retailers.4
The retailers sell a single product that, apart from their specific retail services, is
otherwise homogenous. Its marginal cost, excluding delivery, is constant and
denoted by k.
4
A similar Hotelling specification has been used within the context of parcel delivery by
Borsenberger et al. (2014, 2015). However, in these papers taste differences were across modes
of delivery rather than retailers.
There is a single delivery operator, who charges a rate of r0 to retailer 0. Bypass,

i.e., the ability to deliver parcels by its own means, is ruled out but will be briefly
discussed in the Conclusion. Retailer 1 can either deliver directly via parcel
delivery operator at rate r (r r 0 ); the general rate which also applies for single
piece senders. Alternatively, it can “join” the marketplace and use the delivery
services of retailer 0. This affects utility and also the degree of product differenti-
ation. We will consider these two market configurations separately.
The assumption of a price differential between the delivery rate charged to the
“big” retailer, and the smaller one is a common practice of parcel delivery opera-
tors, justified by the features of the activity’s cost function (that exhibits sunk costs
and economies of scale) and by the fact that retailers provide different volumes of
parcels according to their size. The discount depends on the freight profile of the
senders, i.e., the number of parcels per shipment, average size and weight of the
parcels, etc. Moreover, the place where parcels are collected by the delivery
operator generally differs: a big retailer often drops off its parcels directly in a
sorting center, whereas a small retailer leaves its few parcels in a point of contact,
generating higher collection costs for the delivery operator.
Observe that one can consider our analysis from two different but complementary
perspectives. First, one can think of it as providing a purely positive (albeit stylized)
perspective of the electronic retail market. From that perspective the profit maxi-
mizing solution is probably the most appealing.5 Second, one can take a normative
perspective and ask whether price differentiation across retailers is desirable on
welfare grounds. It is to address this second question that we consider the more
general objective function (which encompasses welfare and profit maximization)
and why we refrain from imposing uniform delivery rates in an ad hoc way.
2.1 Independent Delivery
In this case referred to as subgame I, the utility of consumer z, who buys x units of
the good is given by

αuðxÞ p0 x tz2 if the good is sold by firm 0
ð1Þ
uðxÞ p1 x tð1 zÞ2 if the good is sold by firm 1
where α 1. Retailer 0 is a “big” retailer, whose reputation translates into a higher

quality perceived by the consumer, which is captured by α. Define indirect utility
(consumer surplus) as
vðα; pÞ ¼ maxx αuðxÞ px ð2Þ
5
This is certainly true when the postal operator is privatized. But even when it is public, profits
from parcel delivery may be vital to compensate for the loss of volume in the letter market.
Equation (2) represents the maximum utility consumers can achieve when they
patronize a retailer with quality attribute α who charges a price p. For customers of
retailer 0, we have α ¼ α and p ¼ p0 while for those of retailer 1 α ¼ 1 and p ¼ p1 .
The marginal consumer ^z (α, 1, p0, p1) is defined by that for whom
vðα; p0 Þ t^z 2 ¼ vð1; p1 Þ tð1 ^z Þ2
This consumer is indifferent between buying from retailer 0 or 1. All consumers

with a lower value of z will patronize retailer 0; they represent a share of G[^z (α, 1,
p0, p1)] of the total population. The consumers with z ^z , who represent a share of
ð1 G½^z ðα; 1; p0 ; p1 ÞÞ will buy from retailer 1. Solving for ^z yields
1 vðα; p0 Þ vð1; p1 Þ
^z ðα; 1; p0 ; p1 Þ ¼ þ ð3Þ
2 2t
Aggregate (market) demand for the two products is given by
X0I ðα; 1; p0 ; p1 Þ ¼ xðα; p0 ÞG½^z ðα; 1; p0 ; p1 Þ ð4Þ

X1I ðα; 1; p0 ; p1 Þ ¼ xð1; p1 Þð1 G½^z ðα; 1; p0 ; p1 ÞÞ ð5Þ
Profits of the retailers are given by
π 0I ðα; 1; p0 ; p1 Þ ¼ ð p0 k r 0 ÞX0I ðα; 1; p0 ; p1 Þ ð6Þ

π 1I ðα; 1; p0 ; p1 Þ ¼ ð p1 k r ÞX1I ðα; 1; p0 ; p1 Þ ð7Þ
The retailers simultaneously set their prices and the solution is given by the Nash
equilibrium, denoted by the superscript NI. The equilibrium prices are then ( pNI 0 ,
pNI
1 ), equilibrium demands are

0 ¼ X0 α; 1; p0 ; p1
XNI ð8Þ
I NI NI

1 ¼ X1 α; 1; p0 ; p1
XNI ð9Þ
I NI NI
and equilibrium profits are

π NI
0 ¼ π 0 α; 1; p0 ; p1
I NI NI
ð10Þ

1 ¼ π 1 α; 1; p0 ; p1
π NI I NI NI
ð11Þ
The parcel delivery operator’s profits are given by

Π NI ¼ ðr 0 cÞX0I α; 1; pNI
0 ; p1
NI
þ ðr cÞ X1I α; 1; pNI
0 ; p1
NI
þ Y ðr Þ F ð12Þ
where F denotes the operator’s fixed cost, while Y(r) is the demand for single piece
delivery services (by household and other small firms). Formally, we have
Y ðr Þ ¼ argmax½SðY Þ rY
where S(Y ) is the (aggregate) gross surplus of single-piece customers (other than
retailers).
2.2 Marketplace Delivery
In this case, referred to as subgame M, the utility of consumer I, who buys x units of
the good is represented by

αuðxÞ p0 x δtz2 if the good is sold by firm 0
ð13Þ
γuðxÞ p1 x δtð1 zÞ2 if the good is sold by firm 1
where α 1, 1 γ α, and δ < 1. The parameter δ represents the property that

delivery through the marketplace reduces the degree of horizontal product differ-
entiation. It reduces the utility loss customers experience when patronizing a
retailer whose characteristics differ from their preferred ones. Consequently, the
goods become closer substitutes and price competition will be more intense. When
γ > 1, marketplace delivery also increases the perceived quality of good 1; the
retailer now benefits from the reputation and warranties of the marketplace.
Proceeding as above, the marginal consumer is now determined by
1 vðα; p0 Þ vðγ; p1 Þ
ez ðα; γ; p0 ; p1 Þ ¼ þ
2 2δt
and aggregate (market) demand for the two products is
X0M ðα; γ; p0 ; p1 Þ ¼ xðα; p0 ÞG½ez ðα; γ; p0 ; p1 Þ ð14Þ

X1M ðα; γ; p0 ; p1 Þ ¼ xðγ; p1 Þð1 G½ez ðα; γ; p0 ; p1 ÞÞ ð15Þ
Profits of the retailers are given by
π 0M ðα; γ; p0 ; p1 Þ ¼ ð p0 k r 0 ÞX0M ðα; γ; p0 ; p1 Þ

þ ðs r 0 ÞX1M ðα; γ; p0 ; p1 Þ ð16Þ
π 1M ðα; γ; p0 ; p1 Þ ¼ ð p 1 k sÞX1M ðα; γ; p0 ; p1 Þ ð17Þ
where s is the per unit rate retailer 0 charges to retailer 1 to use its marketplace.6
As in the case of independent delivery we assume that the retailers simulta-
neously set their prices and that the solution is given by the Nash equilibrium,
6
We assume that there is no fixed fee for joining the marketplace.
denoted by the superscript NM. The equilibrium prices are denoted by ( pNM NM
0 , p1 ).
Substituting into expressions (14)–(17) yields the equilibrium demands and profit
levels, i.e., counterparts to expressions (8)–(11).
Finally, the parcel delivery operator’s profits under marketplace delivery are
given by

Π NM ¼ ðr 0 cÞ X0M α; γ; pNM
0 ; p1
NM
þ X1M α; γ; pNM
0 ; p1
NM
þ ðr cÞY ðr Þ F ð18Þ
Comparing Eqs. (12) and (18) shows that the total sales of both retailers are now
delivered at the rate r0. The marketplace thus introduces a secondary market for
delivery services which, even in the absence of bypass, restricts the parcel delivery
operator’s ability to differentiate prices.
2.3 Sequence of Decisions
The timing of the “full game” consisting of delivery and retail pricing is as follows.
In Stage 1, the parcel delivery operator sets r0 and r to maximize welfare subject to
the break-even constraint and anticipating the induced equilibrium. In Stage
2, retailer 0 chooses s, that is the rate at which it is willing to sell its delivery
service to the other retailer. In Stage 3, retailer 1 chooses independent delivery or
marketplace delivery. Finally, in Stage 4 the retailers simultaneously choose their
prices p0 and p1 in subgames I or M, which are described in Sects. 2.1 and 2.2
above.7
As usual, we solve this game by backward induction to characterize the subgame
perfect Nash equilibrium. At each stage, the players (operator or retailers) antici-
pate the impact their choices will have on the equilibrium in the subsequent stages.
Though highly stylized, our model is too complicated to provide a full analytical
solution. However, some analytical results can be obtained and in any event a
thorough examination of the various stages is necessary to properly define the
numerical solutions we will calculate in Sect. 4.
7
The timing we consider is the classical one in monopoly and particularly utility pricing problems.
It is therefore a natural benchmark to consider when taking a first pass at the problem we are
dealing with. However, as suggested by Borsenberger (2015) it may overstate the market power of
the delivery operator. Because large retailers may have a significant monopsony power, the
operator may not be able to commit to a pricing strategy. The threat to develop its own delivery
network and bypass the parcel delivery operator, mentioned in the conclusion, could be a way for
the biggest retailer to enhance its bargaining power. To gain further insight into the “dynamics” of
competition in the e-retail market, our analysis can be extended by considering different specifi-
cations of the timing. In particular, it would be interesting to consider a setting where the shipping
rate of retailer 0 is determined by a bargaining process and study for instance the Nash bargaining
solution. While this goes beyond the scope of the current paper, it is on our agenda for future
research.
3 Equilibrium
We start by studying the last stage of the game. At this point, retailer 1 has already
decided if it delivers independently or via the marketplace. Consequently, the
retailers play subgame I or subgame M. We shall examine them separately.
3.1 Stage 4
3.1.1 Subgame I
At this point, r0 and r are given and s is of no relevance because the retailer has
decided not to join the market place. The equilibrium of the price game yields the
0 (r0, r), p1 (r0, r), and profits, π 0 (r0, r), π 1 (r0, r) as functions
equilibrium prices, pNI NI NI NI
of the variables set in the earlier stages.
3.1.2 Subgame M
Once again r0, r, and s are given. The equilibrium of the price game yields pNM 0 (r0,
r, s), p1 (r0, r, s), and the profit levels π 0 (r0, r, s) and π 1 (r0, r, s). Observe that
NM NM NM
s is now relevant and affects the equilibrium.

Since we study the subgame perfect equilibrium, the comparative statics prop-
erties of these functions, in particular with respect to s are relevant to analyze the
earlier stages of the game. This is not a trivial exercise because we have to study the
equilibrium and not just differentiate the profit functions. However, we can expect
that
∂π NM
1 ðr 0 ; r; sÞ
<0 ð19Þ
∂s
In other words, retailer 1’s profit decreases as the delivery fee by the other retailer
increases. The formal analysis which we skip shows, however, that the results are
not unambiguous but that it would take strong assumptions to obtain a different
result.8
8
The direct effect of s on π 1, even accounting for the induced increase in price is negative by the
envelope theorem. However, as p1 increases (the best-reply function of retailer 1 shifts upwards) p0
will increase (prices are strategic complements) which has a positive effect on retailer 1’s profit.
Similarly, we expect
∂π NM
0 ðr 0 ; r; sÞ
> 0;
∂s
for small levels of s, but for some sˆ(r0, r) we may have
∂π NM
0 ðr 0 ; r; sÞ
¼0
∂s
In other words, sˆ is the profit-maximizing level of s for retailer 0.
3.2 Stage 3
At this stage, retailer 1 will decide whether or not to join the marketplace. To do so,
it will compare π NI 1 (r0, r) and π 1 (r0, r, s). When π 1 ðr 0 ; r Þ > π 1 ðr 0 ; r; sÞ, the
NM NI NM
retailer will choose independent delivery. Otherwise, it will join the marketplace.
Note that for this comparison it is not sufficient to compare r ands; this is because
quality level and degree of product differentiation differ between the two subgames.
Assuming that (19) holds, there exists a critical level of s, sˆ(r0, r) such that
1 ðr 0 ; r Þ ¼ π 1 ðr 0 ; r; ^
π NI sÞ
NM
For this level of s, retailer 1 is indifferent between marketplace and independent

delivery. The equilibrium strategy of retailer 1 in the stage is then to choose I if
s > es ðr 0 ; r Þ and M if s es ðr 0 ; r Þ. Note that we have assumed that in case of
indifference retailer 1 chooses M.9
3.3 Stage 2
Retailer 0, the potential marketplace, sets s, the delivery rate charged to the other
∗ ∗M
0 (r0, r) to π 0 ðr 0 ; r; s Þ, where s
retailer. This is achieved by comparing π NI NM
ðr 0 ; r Þ
is the solution to
max π NM
0 ðr 0 ; r; sÞ
s
s:t: s es ðr 0 ; r Þ
9
This assumption is made for technical reasons. Since one of the possible strategies of retailer
0 may be to set s ¼ es , and we may have existence problems if we adopt the opposite assumption.
Retailer 0 effectively faces two questions. First, what is the profit maximizing level
of s (denoted s* M) which induces retailer 1 to join the marketplace? Second, how
does the profit achieved with s∗ compare to the profit under independent delivery?
For retailer 0, the creation of a marketplace has two conflicting effects on profits.
On the one hand, the fee paid by the other retailer generates additional revenues;
this increases profits as long as the fee exceeds cost. On the other hand, the reduced
degree of differentiation implies a more intense price competition. When the
second effect dominates, retailer 0 sets a sufficiently large fee to discourage the
other retailer from joining the marketplace.
The results obtained for Stages 3 and 4 imply that s∗ is determined as follows.
We have either s∗M ðr 0 ; r Þ ¼ min½es ðr 0 ; r Þ, ^s ðr 0 ; r Þ when there is an interior solution
for the profit maximization in NM, or s∗M ðr 0 ; r Þ ¼ es ðr 0 ; r Þ when there is a corner
solution. Operator 0 then sets the highest s for which operator 1 chooses the M
subgame.
The optimal level of s, s∗ ðr 0 ; r Þ is then given by s∗M ðr 0 ; r Þ when π NM 0 ðr 0 ; r; s
∗M
Þ
∗ e
π NI
0 ð r 0 ; r Þ or by some arbitrary level s ð r 0 ; r Þ > s ð r 0 ; r Þ otherwise. We assume
that when operator 0 is indifferent between the two regimes, I and M, it chooses the
one preferred by the delivery operator.
3.4 Stage 1
We are now in a position to state the problem of the parcel delivery operator who
sets (r0, r). As usual we consider a Ramsey problem where the parcel delivery
operator maximizes welfare subject to a break-even constraint. Recall that this has
profit maximization as a special case.10
The parcel delivery operator’s pricing policy (r0, r) will induce either the inde-
pendent delivery or the marketplace equilibrium in the subsequent stages. Writing
the objective function in a unified way for both regimes would complicate notation
significantly. Consequently, we write two separate problems, one for each regime.
To determine the best policy, one then has to compare the level of the objective
achieved at these solutions. Analytically, this would be difficult but we address this
issue in the numerical section.
We now successively consider the parcel delivery operator’s problem in each of
the two regimes. Recall that maximization is over (r0, r).
10
Formally, the profit maximizing case is obtained by setting the profit target π at the monopoly
level. Observe that in this case the Lagrange multiplier of the break even constraint will tend to
infinity.
3.4.1 Regime I
The Lagrangian expression associated with the parcel delivery operator’s problem
is given by
ð^z ð1h i

LI ¼ v α; pNI
0 tz2 gðzÞdz þ v 1; pNI
1 tð1 zÞ2 gðzÞdz þ SðY Þ
0 ^z
þ ð1 þ λI ÞΠ NI þ π NI 0 ðr 0 ; r Þ þ π 1 ðr 0 ; r Þ λ I π
NI

þ ηI π NI
0 ðr 0 ; r Þ π 0
NM
r 0 ; r; s∗M ð20Þ
∗M
Observe that the arguments of the functions pNI NI
0 , p1 , X0, X1, Y, and s have been
dropped for simplicity. All these expressions are directly or indirectly functions of
(r0, r) and are determined in the subsequent stages, as shown in the previous
subsections and by taking into account expressions (8) and (9).
The condition associated with ηI, namely

π NI
0 ðr 0 ; r Þ π 0
NM
r 0 ; r; s∗M 0
is a Kuhn-Tucker constraint (it may or may not be binding), which ensures that the
vector of rates (r0, r) induces an equilibrium of type I in the subsequent stages.
3.4.2 Regime M
The Lagrangian expression associated with the parcel delivery operator’s problem
is now given by
ð~z ð1 h i

LM ¼ v α; pNM
0 δtz2 gðzÞdz þ v γ; pNM
1 δtð1 zÞ2 gðzÞdz
0 ~z

þ SðY Þ þ ð1 þ λM ÞΠ þ NM
r 0 ; r; s∗M þ π NM
π NM
0 1 r 0 ; r; s∗M λM π

þ ηM π NM
0 r 0 ; r; s∗M π NI
0 ðr 0 ; r Þ ð21Þ
All these expressions are directly or indirectly functions of (r0, r) and are deter-
mined in the subsequent stages, as shown in the previous subsections.
The condition associated with ηM, namely

π NM
0 r 0 ; r; s∗M π NI
0 ðr 0 ; r Þ 0
is a once-again Kuhn-Tucker constraint which may or may not be binding. It

ensures that the vector of rates (r0, r) induces an equilibrium of type M in the
subsequent stages.
4 Numerical Illustrations
In this section, we provide numerical simulations whose aim is to illustrate qual-

itatively the characteristics of the equilibrium obtained with reasonable values of
the parameters. A calibration of our model to a specific postal market would require
a sizeable amount of empirical work to estimate its various constituents. We leave
such a calibration exercise for future research.
We use of the following values: k ¼ 10, α ¼ 1:1, t ¼ 25, γ ¼ 1:05, δ ¼ 0:8, and
c ¼ 0:5. We assume linear individual demand functions (obtained from quadratic
utilities) which are such that (1) their direct price elasticity is 4.2 at a consumer
price of 12, and (2) that xðα; 12Þ ¼ 10. We assume that the distribution of tastes,
G(.), is uniform over [0, 1].
We first describe the equilibrium in the model, before moving to its sensitivity to
various assumptions.
4.1 Benchmark Results
As in the previous sections, we solve the game by backward induction, starting with
Stage 4. We first present the numerical results obtained in subgame I where retailer
1 chooses independent delivery. Table 1 details the first-best allocation, where both
r and r0 are set equal to their marginal cost, c ¼ 0:5. The first row of Table 1
corresponds to the situation where retailer 1 delivers independently. The first-best
retail prices are identical (since they both equal the same marginal cost k þ c ¼
10:5) and result in a (Gð^z Þ ¼ ^z ) 81.65 % market share for retailer 0, thanks to its
higher quality (α ¼ 1:1). We assume away fixed costs for the moment, so that all
three firms exactly break even, and total welfare is composed exclusively of
consumer surplus, denoted by CS. The results presented in Table 1 will help us
ascertain the welfare properties of the market equilibrium, to which we now turn.
Table 2 presents the market equilibrium with independent delivery by retailer
1, as a function of the prices set by the parcel delivery operator, r0 and r. More
precisely, we assume from now on that r is set exogenously at 1. As the reader will
see, the results we obtain are already rich and complex, and would be made
significantly more complex (especially to report) with the parcel delivery operator
optimizing on two price dimensions at the same time. As a first pass, we then set
r exogenously and concentrate on the ratio between r0 and r.
Table 2 reports the equilibrium profit-maximizing price levels chosen simulta-
neously by retailers 0 and 1 ( pNI NI
0 and p1 ) and the ensuing allocation as a function of
the exogenous value of r0 shown in the first column. Comparing with Table 1, we
see that retailer 0 makes use of its higher quality to raise its price above the one
posted by retailer 1, which results in a smaller market share than in Table 1, and in
smaller consumer surplus and total welfare. As its input price r0 increases, retailer
0 increases its retail price and moves further away from the first-best optimal
Table 1 First-best allocation (benchmark)
Subgame r0 p0 p1 ^z x0 x1 Π0 Π1 Π CS CS þ Π 0 þ Π 1 þ Π
I 0.5 10.5 10.5 0.816463 15.25 11.575 0 0 0 25.732 25.732
M 0.5 10.5 10.5 0.709274 15.25 13.5 3.924 s 1.9624 3.924 s + 1.9624 0 28.247 28.247
Differentiated Pricing of Delivery Services in the e-Commerce Sector
203
204
Table 2 Independent delivery equilibrium (benchmark)

NI NI
r0 pNI
0 pNI
1 ^z xNI
0 xNI
1 Π NI
0 Π NI
1 Π NI CSNI CSNI þ Π NI
0 þ Π1 þ Π
0.5 12.0543 11.9741 0.684553 9.81006 5.89984 10.4377 1.81283 0.930544 7.90251 21.0836
0.55 12.0946 11.9807 0.677481 9.66894 5.87428 10.1179 1.85801 1.27481 7.62263 20.8733
0.6 12.1348 11.987 0.670494 9.52815 5.85012 9.80527 1.90255 1.60268 7.3504 20.6609
0.65 12.1749 11.9929 0.663597 9.3877 5.82727 9.49986 1.94643 1.9146 7.08569 20.4466
0.7 12.215 11.9985 0.65679 9.24758 5.80563 9.20155 1.98963 2.21102 6.82832 20.2305
0.75 12.2549 12.0039 0.650078 9.10782 5.78513 8.91025 2.03216 2.49237 6.57815 20.0129
0.8 12.2947 12.0089 0.643461 8.96842 5.7657 8.62587 2.07401 2.7591 6.33503 19.794
0.85 12.3345 12.0137 0.636943 8.82939 5.74726 8.34833 2.11517 3.01163 6.09882 19.5739
0.9 12.3741 12.0182 0.630524 8.69076 5.72977 8.07751 2.15563 3.2504 5.86937 19.3529
0.95 12.4136 12.0226 0.624208 8.55253 5.71316 7.81332 2.19539 3.47583 5.64655 19.1311
1. 12.4529 12.0267 0.617995 8.41472 5.69738 7.55564 2.23444 3.68834 5.43022 18.9086
1.05 12.4922 12.0306 0.611887 8.27735 5.68238 7.30439 2.27278 3.88834 5.22025 18.6858
1.1 12.5313 12.0343 0.605886 8.14044 5.66811 7.05944 2.31041 4.07625 5.0165 18.4626
C. Borsenberger et al.
allocation. We also see that prices p0 and p1 are strategic complements, although p1
is much less sensitive to increases in r0 than p0. An increase in r0 benefits retailer
1 (π NI NI
1 increases) at the expense of retailer 0 (π 0 decreases) and of both consumer
surplus and total welfare. A benevolent social planner would then set a value of r0
below cost. The parcel delivery operator’s profit π NI increases with r0, even when r0
becomes larger than r. The intuition for this result is that a profit-maximizing parcel
delivery operator would like to exploit the larger quality exhibited by retailer 0, and
that its only way to extract profit from retailer 0 is to increase r0.
When the parcel delivery operator is maximizing welfare under a break-even
constraint, the value of r0 it chooses is increasing in its fixed cost F. As the parcel
delivery operator cannot set a value of r0 above r ¼ 1, its profit-maximizing price is
r 0 ¼ 1. We then see that the maximum value of the fixed cost compatible with
breaking-even is 3.688.
We now turn to the equilibrium with marketplace. The second row of Table 1
shows the first-best allocation in that case. First-best prices are not affected, but the
market share of retailer 0 decreases to 70.93 % due to the larger quality of the
service offered by retailer 1 when it uses retailer 0’s marketplace. The optimal
quantity x1, consumer surplus and total welfare are higher than in the first row,
while the value taken by s acts only as a transfer between retailers 0 and 1.
The value of s is chosen at Stage 2 by retailer 0 in order to maximize its profit.
We obtain, as surmised in Sect. 2, that π NM 0 is first increasing and then decreasing in
s. We also obtain that the profit-maximizing value of s is much larger than r0 and
even than r, for all values of r 0 r. The intuition for these results is that retailer
0 anticipates that joining its marketplace will result in retailer 1’s higher quality and
thus higher profit, and that s plays the role of an “access charge” to the marketplace.
Also, increasing s induces retailer 1 to increase its retail price, decreasing the
intensity of competition with retailer 0. These two reasons concur in pushing the
value of s well above the marginal cost of delivery for retailer 0 (r0) and even for
retailer 1 (r). But retailer 0 anticipates that it will realize this profit π NM 0 only if
retailer 1 accepts to join its marketplace, that is if retailer 1’s profit with market-
place delivery is at least as large as with independent delivery. We obtain that π NM 1
is decreasing in s (confirming Eq. (19)) and that this limit pricing constraint is
binding for all values of r0. In other words, the value of s which equalizes π NI 1 and
π NM
1 (denoted as es ð r 0 ; 1 Þ ) is smaller than the value of s which maximizes π NM
0
(denoted as sˆ(r0, 1)).
We then report in Table 3 the equilibrium allocation attained in the marketplace
for values of r0 varying from c ¼ 0:5 to r ¼ 1.11 By comparing Tables 2 and 3, we
see that π NM0 > π NI
0 , for all values of r0. Table 3 then depicts the equilibrium
allocation as a function of the value of r0 chosen by the parcel delivery operator
in the first stage of the game. We obtain that the parcel delivery operator’s profit is
11
We vary s by increments of 0.01 in our computations. The values of s reported in Table 3 are the
highest for which π NM
1 π NI
1 holds.
206
Table 3 Marketplace equilibrium (benchmark)

r0 s pNM
0 pNM
1 ^z xNM
0 xNM
1 Π NM
0 Π NM
1 Π NM CSNM CSNM þ Π NM
0 þ Π NM
1 þ Π NM
0.5 1.6 12.23 12.5308 0.677033 9.19507 6.05357 12.9204 1.81989 0 7.49796 22.2382
0.6 1.59 12.2844 12.535 0.665284 9.00471 6.03844 12.0915 1.90994 0.801185 7.15758 21.9602
0.7 1.58 12.3384 12.5383 0.653749 8.81567 6.02636 11.2786 1.99955 1.56997 6.83342 21.6816
0.8 1.57 12.392 12.5408 0.642444 8.6281 6.01709 10.481 2.08861 2.30836 6.52513 21.4031
0.9 1.57 12.4476 12.5509 0.632095 8.43345 5.98015 9.72386 2.15804 3.01235 6.20087 21.0951
1 1.56 12.5004 12.552 0.621257 8.24863 5.97588 8.95625 2.24529 3.69392 5.92224 20.8177
C. Borsenberger et al.
monotonically increasing in r0. A profit-maximizing parcel delivery operator will

then set r 0 ¼ r ¼ 1; in other words, it will not give any discount to retailer 0. Both
consumer surplus and total welfare are higher with the marketplace than in the
subgame with independent delivery by retailer 1.
We now turn to the sensitivity analysis of these results.
4.2 Sensitivity Analysis
We have studied the sensitivity of our results to variations in three assumptions: a

larger exogenous value of r, a larger impact of joining the platform on the
horizontal differentiation between the products offered by both retailers (i.e., a
lower value of δ) and on the quality of good 1 (i.e., a larger value of γ).
A natural question suggested by the numerical results presented in the previous
section is whether setting a higher exogenous value of r (at 2.5 rather than 1) would
result in an interior profit-maximizing value of r0 for the parcel delivery operator
(i.e., r 0 < r ). The answer is positive in the subgame where retailer 1 delivers
independently, with a profit-maximizing value of r0 of 2.35. At the same time,
the only role played by r in the subgame with marketplace is to degrade the outside
option of retailer 1, namely its profit level under independent delivery. The con-
straint that π NM
1 π NI
1 is then less binding when retailer 0 sets its profit-maximizing
level of s. We even obtain, for low values of r0, that this constraint that π NM
1 π NI
1 is
not binding at the level of s which maximizes π 0 (i.e.,es ðr 0 ; 2:5Þ > ^s ðr 0 ; 2:5Þ). The
NM
value of s chosen by retailer 1 is then much higher, for any value of r0, than in the
benchmark numerical results, and retailer 0 makes much more profit with the
marketplace than with independent distribution by retailer 1. Finally, the parcel
delivery operator’s profit remains monotonically increasing with r0 as long as
r 0 r, and it makes more profit than when r was set at one (so that Ramsey levels
of r0 are lower, for given F, when r is increased to 2.5). Consumer surplus and
welfare remain larger at equilibrium (with the marketplace) than in the subgame
with independent delivery by retailer 1.
To summarize, a higher exogenous value of r induces an interior profit-
maximizing value of r0 when retailer 1 delivers independently, but not in the
equilibrium situation where retailer 1 joins the marketplace, because it degrades
the outside option of retailer 1 and allows retailer 0 to set a much higher level of the
“access charge” s to its marketplace.
We now go back to the setting where r ¼ 1, but assume that joining the
marketplace decreases the degree of differentiation between products more signif-
icantly. Specifically, we set δ ¼ 0:5 rather than 0.8. This change does not impact the
equilibrium with independent delivery by retailer 1, and thus it is outside option.
Less differentiation in the marketplace results in more intense competition, lower
prices, and lower profits for both retailers for all values of r0. We continue to obtain
1 ¼ π 1 ), but with lower values of s than in the
limit pricing by retailer 0 (so that π NI NM
benchmark numerical results. More precisely, as long as r0 is low enough,

π NM
0 > π NI
0 , and retailer 0 posts the highest value of s compatible with retailer
1 joining the marketplace, es ðr 0 ; 1Þ. When r0 is large (although smaller than r), π NM 0
0 when s ¼ e
< π NI s ðr 0 ; 1Þ, and retailer 0 posts a high value of s (larger thanes ðr 0 ; 1Þ) to
deter retailer 1 from joining the marketplace. Turning to the parcel delivery
operator, its profit is increasing in r0 as long as retailer 1 joins the marketplace,
and becomes significantly lower when r0 is large enough that the marketplace does
not form. We then observe some “double limit pricing” when the parcel delivery
operator maximizes its profit: retailer 0 posts the highest value of s compatible with
retailer 1 joining the marketplace, es ðr 0 ; 1Þ, and the parcel delivery operator sets the
highest value of r0 (strictly lower than r ¼ 1)compatiblewith retailer 0 finding it
profitable for retailer 1 to join the marketplace π NM 0 π NI
0 . Both consumer surplus
and aggregate welfare remain higher in the equilibrium with marketplace than in
the subgame with independent delivery by retailer 1.
Finally, we assume that joining the marketplace allows retailer 1 to bridge most
of the quality gap between its product and retailer 0’s (i.e., we set γ ¼ 1:09 while
returning to r ¼ 1 and δ ¼ 0:8). This does not affect the equilibrium with
independent delivery and the outside option of retailer 1. In the marketplace,
retailer 0 increases its price s to very high levels up to the point where
1 ¼ π 1 , and makes much larger profits than with independent delivery by
π NI NM
retailer 1. In other words, retailer 0 can capture all the increase in retailer 1’s profit
generated by a larger γ by increasing the access charge s to the marketplace. The
parcel delivery operator’s profit remains monotonically increasing in r0. Interest-
ingly, although total welfare remains higher with the marketplace than in the
subgame with independent delivery, consumer surplus is lower: in equilibrium,
retailer 0 and the parcel delivery operator make larger profit, at the expense of
consumer surplus, because higher equilibrium prices more than compensate the
larger quality offered by retailer 1 when it joins the marketplace. Intuitively, this
occurs because only consumers who patronize retailer 1 benefit from the higher
quality, and they represent a small share, whereas all consumers support the price
increase of both products.
5 Conclusion
This chapter has examined the link between the delivery rates charged by parcel
delivery operators and the e-commerce market structure. In particular, it has studied
the impact of the parcel operators’ pricing strategies on the retailers’ incentives to
develop a marketplace. It has considered a market where two retailers, a big one,
0, and a smaller one, 1, sell a homogenous good online.
Initially, all parcels are delivered by a parcel delivery operator. However, retailer
0 may or may not offer its competitor the option to join its marketplace through the
payment of an access fee. Affiliation to the marketplace has several consequences
for retailer 1: (1) it reduces the degree of differentiation between the products, (2) it
increases the perceived quality of retailer 1’s product which in turn increases the
consumers’ willingness to pay, and (3) the marketplace consolidates the parcels
send by retailers 0 and 1 and could, in theory obtain better pricing conditions from
the parcel delivery operator.12
Under the assumptions made in the numerical simulations, we obtain that a
marketplace will emerge in equilibrium. This is good news for consumers: their
surplus increase since the perceived quality of retailer 1’s product increases when it
uses retailer 0’s marketplace services. Compared to the case where retailer
1 remains independent, the welfare is also higher: retailer 0 makes higher profits,
retailer 1 is indifferent between both situations, and the parcel delivery operator sets
the same price in both cases and makes higher profits when there is a marketplace
since demand increases.
The fact that the parcel delivery operator does not give any discount to retailer
0 even though it is a bigger customer, with the more elastic demand, could appear
counterintuitive in an industry characterized by scale economies and where volume
discounts are a common practice. A sensitivity analysis shows that the latter result
arises because the delivery rate charged to retailer 1 is set at a rather low level.
When it is fixed at a higher level, the parcel delivery operator concedes a discount to
retailer 0 in equilibrium.
Another sensitivity analysis shows that it is not always in the interest of retailer
0 to offer its smaller competitor the possibility to join its marketplace. This is the
case for instance if the affiliation to the marketplace reduces significantly the degree
of differentiation between retailers products and therefore increases competition
intensity, resulting in lower prices and lower profits for both retailers. The result
also emerges when retailer 0’s competitive advantage on delivery pricing (its
bargaining power towards the parcel delivery operator) is not sufficiently impor-
tant; in other words, when the discount offered by the parcel delivery operator to
retailer 0 is rather small. However, even in this case, it is in the interest of the parcel
delivery operator to see a marketplace emerge, since this will have a positive impact
on the volume of parcels to deliver. Consequently, the parcel delivery operator will
set a price for parcel delivery such that retailer 0 will find it profitable for retailer
1 to join the marketplace.
Finally, when joining the marketplace has a sufficiently large impact on the
perceived quality of retailer 1’s product, the emergence of a marketplace is in the
interest of both retailer 0 and the parcel delivery operator. But, in this case, although
total welfare remains higher with the marketplace than under independent delivery,
consumer surplus is lower. As a matter of fact, only consumers who patronize
retailer 1 benefit from the higher quality, and they represent a small share, whereas
all consumers support the price increase of both products.
12
We have not considered the case where joining the marketplace reduces the perceived quality of
the small retailer. In that situation, retailer 1 might nevertheless find it profitable to join the
marketplace, but only if this one offers sufficiently low shipping rates.
Our analysis can easily be extended to consider the case where retailer 0 is able
to bypass the parcel delivery operator and deliver by itself part or all of its parcels;
see the Appendix. Intuitively, one expects the degree of bypass to depend on the
rate charged by the parcel delivery operator. When it is above a certain threshold,
retailer 0 may even deliver by itself all its parcels. Faced with this threat, the parcel
delivery operator may then give retailer 0 a higher discount compared to the case
where bypass is not available and its profits can be expected to decrease.13
6 Appendix
6.1 Introducing Bypass
So far we have assumed that there is no bypass. Assume now that the retailer 0 can
deliver by itself an exogenously given share, β 2 ½0; 1 of its total volume Q to be
delivered, where Q ¼ X0I in case of independent delivery and Q ¼ X0M þ X1M in case
of marketplace delivery. The delivery cost associated with this bypass technology is
0 00
given by d(QB); we assume d > 0 and d > 0. For any given Q retailer 0 then
chooses QB and QP to minimize delivery costs given by

D0 ¼ d QB þ r 0 QP
s:t:
QB þ QP ¼ Q
QB βQ
We can have two types of solution: (1) maximum bypass with QB ¼ βQ which
0 0
occurs when d ðβQÞ r 0 or (2) partial bypass with QB βQ, when d ðβQÞ > r 0; QB
0 B
is then implicitly defined by d Q ¼ r 0 .
h 0 1 i
We can then define QB ðr 0 Þ ¼ min d ðr 0 Þ, βQ which represents the solution to
this problem.
While some expressions have to be modified in a straightforward way, bypass
does not modify the fundamental structure of the game.
Acknowledgements We thank Anna M€ oller Boivie, Tim Brennan, Michael Crew, and Alberto
Pimenta for their helpful comments.
13
This intuition is illustrated by additional simulations which can be obtained from the authors on
request.
References
Borsenberger C (2015) The concentration phenomenon in e-commerce. In: Crew MA, Brennan TJ
(eds) Postal and delivery innovation in the digital economy. Edward Elgar, Cheltenham, pp
31–42
Borsenberger C, Cremer H, De Donder P, Joram D, Lécou S (2014) Pricing of delivery services in
the e-commerce sector. In: Crew MA, Brennan TJ (eds) The role of the postal and delivery
sector in a digital age. Edward Elgar, Cheltenham, pp 75–92
Borsenberger C, Cremer H, De Donder P, Joram D (2015) Quality and pricing of delivery services
in the E-commerce sector. In: Crew MA, Brennan TJ (eds) Postal and delivery innovation in
the digital economy. Springer International, Cham
Fevad (2014) Bilan du e-commerce en France en 2013, January
Oxatis (2014) Le profil du e-commerçant en 2014, January
Mode of Delivery and Customer Response
to Advertising Mail
Michael D. Bradley, Laraine Balk Hope, and John Pickett
1 Introduction
As the United States Postal Service (USPS) contemplates moving more mail
delivery to neighborhood cluster boxes, it is important to examine and understand
the implications of this potential change. One is whether the type of receptacle in
which postal customers receive their mail influences their interactions with the mail
itself. This chapter examines specifically whether and how cluster box recipients
differ from door or curb recipients in the way they handle and interact with
advertising mail.1 If cluster box recipients report different interactions, are these
due to variations in the mail receptacle or to other factors—i.e., demographic or
other differences among recipients? This chapter investigates these questions by
analyzing survey data, including data from a new survey commissioned by the
USPS Office of the Inspector General (USPS OIG). It also includes statistical
analyses of data from two previous surveys conducted for other purposes.
Section 2 examines the strategic importance of the role of delivery modes,
especially in the case of advertising mail. Section 3 reviews the design of and
findings from the recent USPS OIG-InfoTrends survey of 5000 individuals across
the United States. Section 4 compares these findings with findings based on the
The views presented in this chapter are those of the authors, and do not necessarily represent those
of the Office of the Inspector General, US Postal Service, or any other organization.
1
These are mail recipients, who receive their mail at cluster boxes. We define other mail recipients
who have different receptacle types similarly, e.g., door recipients and curb recipients.
M.D. Bradley
Department of Economics, George Washington University, Washington, DC, USA
L.B. Hope (*) • J. Pickett
USPS Office of Inspector General, Arlington, VA, USA
e-mail: lbhope@uspsoig.gov

214 M.D. Bradley et al.
USPS Household Diary Study and the Mail Moment Study. Section 5 discusses the
results and examines broad findings from the three studies. Finally, Sect. 6 con-
cludes and offers suggestions for future research.
2 Strategic Importance of This Study
Advertising mail is of great strategic importance to USPS. In 2013, USPS products

and services generated $28.2 billion in contribution to institutional cost. The USPS
OIG estimates that 23 % of this total, $6.3 billion, was attributed to advertising mail
sent to households.2 According to the Boston Consulting Group, USPS can expect
the non-advertising component of First-Class Mail volume, its main source of
contribution to institutional costs, to decline significantly by 2020. This change
will force USPS to rely increasingly on revenue from other products, including
advertising mail, to fund its operations.
The USPS has long followed a policy of moving delivery to lower cost delivery
modes.3 New residential communities often have delivery to neighborhood cluster
box units, to single box curbside delivery at the end of the driveway, or even to
multiple box curbside delivery on property lines. Delivery away from the door
reduces the travel time and cost of postal delivery personnel. USPS is operating
under a price cap regime for most of its products, and cost increases can no longer
be passed along to postal customers in the form of price increases.
In the process of moving to curbside and cluster box delivery, USPS’ delivery
costs are being transferred indirectly from USPS to mail recipients. Instead of
opening the front door to the mailbox or receiving mail through a door slot, mail
recipients with curbside or cluster box delivery must travel some distance, perhaps
in inclement weather, to their mailboxes. In addition, customers with disabilities, or
who are temporarily ill, may have physical difficulties in getting to their cluster
boxes. It is reasonable, therefore, to expect that this reduced convenience may
translate into reduced customer engagement with the mail.
Advertising mail was once unique in two important characteristics: (1) advertis-
ing mail efficacy could be tracked; and (2) advertising mail could be used to target
specific markets, neighborhoods, or even individual consumers. With the dawn of
digital direct marketing, advertising mail’s “trackability” and “targetability” are no
2
Calculation using data from three US Postal Service sources: (1) FY2013 Public Cost and
Revenue Analysis, Report No. USPS-FY13-1, December 27, 2013. http://www.prc.gov/dockets/
document/88658; (2) The Household Diary Study, May 2014. http://www.prc.gov/sites/default/
files/uspsreports/USPS_HDS_FY13.pdf; and (3) Market Dominant Billing Determinants, Report
No. USPS-FY13-4, December 27, 2013. http://www.prc.gov/dockets/document/88662
3
See US Government Accountability Office, “USPS Needs to Clearly Communicate How Postal
Services May Be Affected by Its Retail Optimization Plans,” GAO Report No. GAO-04-803, July
2004. http://www.gao.gov/new.items/d04803.pdf, p. 16.
Mode of Delivery and Customer Response to Advertising Mail 215
longer unique.4 At the very time USPS needs to compete with digital communica-
tions, the organization is pursuing a policy that may be reducing the competitive-
ness of advertising mail in order to lower delivery costs.
The United States is not the first or only country to consider a move to more
centralized delivery via cluster box. The experience of its neighbor to the north,
Canada, which announced such a move unilaterally to continuing public outcry and
debate,5 indicates that caution and more study in the area prior to a national
conversion program would be a more pragmatic approach.
Many European countries have cluster boxes only in rural areas or in housing
developments, with the exception of Finland, which is more generally moving to
cluster boxes. However, some studies in European postal markets have suggested
that home delivery is important to residential customers.6 In general, while delivery
to a centralized location (either internal or neighborhood cluster boxes) reduces the
cost of delivery for postal operators (POs), the change makes it less convenient for
customers to retrieve their mail. Observations of POs in Europe suggest that for
delivery cost savings, reduction in the number of delivery days per week may be
preferable for elimination of curbside deliveries.
3 USPS OIG-InfoTrends Survey
The USPS OIG worked with InfoTrends, Inc. to structure a survey designed to
gather insight on the impact of cluster box usage on consumers’ behavior with
respect to advertising mail, including nonprofit solicitation mail. As described in its
white paper, “Advertising Mail and customer engagement by mode of delivery,”
the USPS OIG survey consisted of about 30 questions, with two main purposes.
First, the survey was designed to produce data useful for gaining an understanding
of how postal customers feel about cluster boxes and if/how those receptacles affect
their interactions with advertising mail. In order to do this, the survey investigated
how consumers handled, opened, read, and discarded a wide variety of advertising
4
For a discussion of trends in digital versus direct mail advertising, see US Postal Service Office of
the Inspector General, “State of the Mail,” Report RARC-WP-12-101, April 27, 2012, https://
www.uspsoig.gov/sites/default/files/document-library-files/2013/rarc-wp-12-010.pdf. For a dis-
cussion of the role of targeting in the demand for advertising mail, see Michael D., Jeff Colvin,
and Mary Kay Perkins (2015), “Targeting Versus Saturation: Derived Demand for Direct Mail,” in
Postal and Delivery Innovation in the Digital Economy, Michael Crew and Timothy Brennan
(eds.) Springer International Publishing, pp. 65–76.
5
Canada Post. Five Point Action Plan. December 2013. https://www.canadapost.ca/cpo/mc/
assets/pdf/aboutus/5_en.pdf and, for example, http://www.insidehalton.com/news-story/
5525387-pickering-opposes-canada-post-s-decision-to-end-home-delivery/
6
Charlene Rohr, Urs Trinkner, Alison Lawrence, Chong Woo Kim, Dimitris Potoglou, and Robert
Sheldon (2013), “Measuring consumer preferences for postal services,” in M.A. Crew and
P.R. Kleindorfer (eds.), Reforming the Postal Sector in the Face of Electronic Competition,
Cheltenham, U.K. and Northampton, MA. Edward Elgar, p. 249.
mail. These questions were asked of mail recipients with different types of mail
receptacles, making it possible to investigate different patterns of advertising mail
interactions by mode of delivery. Second, the survey included a series of back-
ground questions about the participants.
The USPS OIG-InfoTrends survey was conducted in August 2014. It included
4500 completed web-based responses and 500 completed phone-based responses
for consumers, who do not have Internet access. The 5000 total responses provided
the primary data used in the analysis of mail recipients and followed standard
sampling protocols, which are designed to ensure reliability of results. The web
survey was sent to a general population sample that was closely monitored for age,
gender, and region. These areas had quotas to match the US census. The computer-
assisted telephone interviewing (CATI) survey went out to a population that did not
have Internet access in their household.7
3.1 Responses by Current Delivery Mode
The following chart shows that 38.6 % of respondents are in households with either
a door slot or a mailbox near the door. Together, these categories comprise the
“door” group and account for the largest group of total respondents by mode of
delivery. To provide some perspective on this distribution, the proportions derived
from survey responses can be compared with the proportions of delivery points, by
type, reported by USPS for FY2013. The next table shows that for USPS, curb
delivery comprises the largest proportion of deliveries. This contrasts with the
USPS OIG-InfoTrends survey sample, which shows more door deliveries than
curb deliveries. The proportion of cluster box and central delivery points is about
the same. (Central delivery boxes are defined here as banks of mailboxes in
apartment buildings, condominiums, etc.)
Distribution of delivery points by mode
Cluster box and central Curb Door
OIG-Info Trends survey 29.0 % 32.3 % 38.6 %
USPS Network FY 2013 30.0 % 41.0 % 28.0 %
Central and cluster box categories from the InfoTrends data are combined into
one category for this comparison because that is the way USPS’ delivery points
data were presented.8 For the purposes of this chapter, the differences in mode
7
While there may always be some sample bias in telephone surveys since some people elect not to
do them, InfoTrends is a nationally recognized and respected polling firm and works to minimize
such bias to the extent possible.
8
See United States Government Accountability Office, “US Postal Service: Delivery Mode
Conversions Could Yield Large Savings, but More Current Data are Needed,” May 2014, p. 6.
proportions are not critical because there were sufficient responses within each
mode. Further, the sample was not intended to replicate the USPS’ network as
currently configured.
3.2 Profile of Participants
The USPS OIG-InfoTrends survey was sent to a general population sample repre-
sentative of the US population and monitored by recipient age and region. Note that
the goal of the survey was to produce data used for comparing response rates across
delivery modes, not to estimate an overall national response rate. We checked the
demographic profiles of recipients to see if they exhibited material differences
across current receptacle types.
Average values for key demographic variables
Full sample Cluster box Curb Door Central
Age (years) 41.1 42.0 43.1 39.8 39.5
College graduates (%) 52.0 53.1 48.7 55.8 52.0
Income ($1000) $60.7 $64.7 $64.8 $61.7 $49.8
Unemployed (%) 13.2 12.5 12.9 11.8 17.2
In all instances, the average values for cluster box recipients were quite close to
the full sample averages. In addition, the average demographics for cluster box
recipients differed little from the average demographics for door and curb recipi-
ents. This observation is important when comparing how these different types of
mail recipients handle their mail. One other area of difference between door
recipients and cluster box recipients relates to their relative Internet access at
home. Door recipients reported a much lower rate of Internet access than the
other types of recipients:
Percentage of Respondents with Internet Access

96.7% 97.0%
100.0% 91.2%
90.0%
76.8%
80.0%
70.0%
60.0%
50.0%
40.0%
30.0%
20.0%
10.0%
0.0%
Central Cluster Box Curb Door
The survey found that among the recipients who did not already have a cluster
box, close to 65 % would be displeased if their mail receptacle were changed to a
locked cluster box. When asked the question, “If your mail delivery was moved to a
locked cluster box, how would you react?” 40.2 % of survey respondents said they
would be “very displeased”; 23.1 % said they would be “somewhat displeased”;
11.1 % said they would be “somewhat pleased”; and 6.1 % said they would be “very
pleased,” with 19.5 % being indifferent.
3.3 Results
The primary focus of the survey was on the attention that different types of mail
recipients gave to the advertising mail they received. The USPS OIG-InfoTrends
survey asked participants to report how they handled different types of advertising
mail. Participants were asked to choose one of the following answers that described
how they handled their advertising:
1. Read and respond to the mail
2. Read and throw away the mail
3. Set the mail aside for later use
4. Throw away the mail without reading it
The USPS OIG-InfoTrends survey examined many different types of advertising
mail to see how interactions with that mail varied by type. Participants were asked
which action they took for the following seven types of advertising mail:
1. Advertising mail from local companies I do business with
2. Advertising mail from national companies I do business with
3. Advertising mail from companies I do not do business with
4. Donation solicitations
5. Credit card solicitations
6. Catalogs
7. Mail that includes a coupon.
The next table provides a summary of the survey results rounded to whole
numbers.
Summary of OIG-Info Trends survey results
Read and response rates (%) Toss without reading rates (%)
Cluster box Curb Door Cluster box Curb Door
Known local business 10 13 24 13 12 12
Known national business 9 9 19 15 14 11
Unknown business 3 4 8 46 46 32
Donation solicitation 4 5 12 40 42 31
Credit card solicitation 3 4 11 60 61 40
Catalogs 6 6 12 17 14 12
Coupons 14 17 20 6 5 6
All types 7 9 15 30 30 22
This table contains two interesting perspectives on advertising mail. First,

reading down each column, one can see how certain types of advertising mail
have inherently higher levels of recipient engagement. For example, credit card
solicitations have low read and respond rates and high discard rates relative to other
kinds of advertising mail, regardless of receptacle type. The survey results make
intuitive sense: for example, response rates to advertising from known businesses
are higher than response rates to advertisements from unknown businesses.
The second perspective, and the one of paramount interest in this chapter, comes
from looking across the rows. For each kind of advertising mail, cluster box
recipients have a lower level of engagement with advertising than do door recipi-
ents. Cluster box recipients indicated that they will read and respond at lower rates,
and discard advertising mail at higher rates, than will door recipients. One can also
clearly see that engagement with advertising mail is quite similar for cluster box
and curb recipients.
3.4 Hypothesis Testing
Although these results suggest that cluster box recipients were less responsive than
door recipients were to the ad mail they received, statistical evidence on these
differences is helpful in assessing the validity of this comparison. This requires
formal testing as to whether there is a statistically significant difference between the
responsiveness of cluster box recipients and the responsiveness of door recipients.
The statistical testing starts with the assumption that the groups of door and curb
recipients are independent samples for basis of calculating the proportions of mail
read and responded to.9
The table below presents the hypothesis tests for the seven types of advertising
mail included in the survey. It shows that for all seven types of mail, the read and
respond rate for cluster box recipients was lower than for door recipients. For
“advertising mail from national companies with which the recipient does business,”
the read and respond rate for cluster box recipients was half of what it was for door
recipients and for “donation solicitations” the cluster box recipient rate was about
one-third of the door recipient response rate. Moreover, all of the test statistics are
well above 1.96, indicating that there is little chance that these results are simply
due to statistical fluctuation:
9
This is typically the case in surveys, unless the subset respondents are matched. For respondents
to be matched, they either have to be the same people in both groups, or people who have been
individually matched to an important characteristic that might potentially influence the outcome.
For example, matching would occur if both cluster box and door recipients were restricted to
women between the ages of 45 and 60. That is not the case in the InfoTrends survey. Door
recipients and cluster box recipients are two completely different groups of individuals that we
want to compare to determine if they are significantly different from one another with regard to
their interaction with advertising mail.
Tests of the hypotheses that the read and respond rate for cluster box recipients is the same as for
door recipients
Mode Sample proportion (%) Test statistic Result
Known local business
Cluster box 10.0
Door 23.5 7.04 Reject at 95 %
Known national business
Cluster box 9.0
Unknown business
Cluster box 3.1
Donation solicitations
Cluster box 4.0
Credit card solicitations
Cluster box 3.1
Catalogs
Cluster box 5.5
Mail that includes a coupon
Cluster box 13.8
The previous results indicate that door recipients were more responsive to
advertising mailings than cluster box recipients, in terms of their positive responses
to those mailings. However, it is also possible to test for differences in mail
recipients’ negative responses to ad mailings. As discussed above, the USPS
OIG-InfoTrends survey also measured the proportion of participants reporting
they threw out their advertising mail without reading it. The statistical tests used
to examine read and respond rates can also be used to see if cluster box recipients’
discard rates were the same as those for door recipients.
As before, the null hypothesis is that discard rates are the same across receptacle
types and the absolute critical values are the same. The one difference is that
negative values for the calculated test statistics mean that the cluster box discard
rate is greater than door discard rate. This means that the associated critical values
will be negative at 1.65 and 1.96.
The next table presents the results of testing this hypothesis for the seven mail
categories. The results are not as uniform in rejecting the null hypothesis as for the
read and respond tests, but they do indicate that cluster box recipients generally had
higher discard rates than door recipients. In all cases, the discard rate was higher for
cluster box recipients than for door recipients, and the null hypothesis of an equal
discard rate for door and cluster box recipients is rejected for five of the seven mail
types. Only two categories, “advertising mail from local companies with which the
recipient does business” and “mail that includes a coupon,” had statistically similar
discard rates. In some cases, the difference in the discard rates was large. Door
recipients discarded 40.2 % of their credit card solicitations without looking at them
while cluster box recipients discarded 59.9 %. Similarly, door recipients immedi-
ately discarded about 31.6 % of “advertising mail from companies with which they
do not do business,” whereas cluster box recipients immediately discarded 46.4 %
of that mail.
Tests of the hypotheses that the do not read and discard rate for cluster box recipients is the same
as for door recipients
Cluster box 13.2
Door 12.3 0.55 Do not reject
Cluster box 14.5
Unknown business
Cluster box 46.4
Cluster box 40.0
Cluster box 59.5
Catalogs
Cluster box 17.1
Cluster box 6.1
Door 6.0 0.10 Do not reject
Comparison of cluster box recipients with door recipients produced consistent

evidence that door recipients were more engaged with advertising mail. The same
methods can be used to examine possible differences between cluster box recipients
and curb recipients. While the read and respond rates were slightly larger for curb
recipients than cluster box recipients, the differences were small. Similarly, discard
rates were very similar across people receiving mail in these two types of
receptacles.
The similarity of cluster box recipients and curb is highlighted by results of
statistical tests. The next set of tables contains the results of testing the hypotheses
that the read and respond rates were the same for cluster box and curb recipients for
the seven types of advertising mail. In no instance can this hypothesis be rejected at
the 95 % level, and only twice can it be rejected at the 90 % level.10 This suggests a
strong similarity between the read and respond rates for the two types of recipients.
Tests of the hypotheses that the read and respond rate for cluster box recipients is the same as for
curb recipients
Cluster box 10.0
Curb 12.9 1.82 Reject at 90 %
Cluster box 9.0
Curb 9.1 0.08 Do not reject
Unknown business
Cluster box 3.1
Cluster box 4.0
Cluster box 3.1
Catalogs
Cluster box 5.5
Cluster box 13.80
Testing discard rates also reinforces the similarity in the way that cluster box and
curb recipients interact with advertising mail. In only one case—catalogs—was the
hypothesis of equal discard rates rejected:
Tests of the hypotheses that the do not read and discard rate for cluster box recipients is the same
as for curb recipients
Cluster box 13.2
Cluster box 14.5
Unknown business
Cluster box 46.4
(continued)
10
If a one-tailed test were performed, two of the hypotheses could be rejected at the 95 % level.

Cluster box 40.0
Cluster box 59.5
Catalogs
Cluster box 17.1
Cluster box 6.1
4 Other Evidence
In order to evaluate the results from the USPS OIG-InfoTrends survey, the USPS
OIG obtained relevant data from two recent USPS surveys: the Mail Moment
survey and the Recruitment Questionnaire of the Household Diary Study. Although
these two postal surveys were not specifically designed for cluster box research and
asked different questions from the USPS OIG-InfoTrends survey, they both gener-
ated data that can be used to potentially corroborate or contradict the USPS
OIG-InfoTrends results. (Such a mode-specific analysis had not previously been
done by the Postal Service.)
InnoMedia conducted the Mail Moment survey in 2012. It had 1078 participants
and was designed to foster understanding of the value of the consumer’s daily “mail
moment.” That is, the Mail Moment survey was designed to investigate the value
that households receive from different kinds of commercial and advertising mail.
The Recruitment Questionnaire of the Household Diary Study is part of an ongoing
study of mailing behavior by USPS. The 2013–2014 version of the Recruitment
Questionnaire was used in this analysis and had 4549 participants. While much of
the Household Diary Study analysis focuses on the quantities of mail sent and
received by households, the Recruitment Questionnaire elicits a variety of attitudi-
nal and demographic information from diary study participants.
The Mail Moment survey data support the inference that cluster box recipients
have lower retention rates and higher discard rates than either door or curb
recipients. The Mail Moment survey asked several questions that can contribute
to an investigation of different interactions with advertising mail for cluster box
recipients than for door or curb recipients. For example, one question in the Mail
Moment survey elicited participants’ interest in picking up their mail at the first
opportunity. As the next table shows, cluster box recipients were far less likely to
pick up their mail at the first opportunity than were either door or curb recipients.
This suggests that cluster box recipients were less engaged with their mail than
other types of recipients.
Picked up mail at first opportunity
Receptacle type Proportion (%)
Cluster box 62.5
Curb 87.7
Door 93.9
Another question in the Mail Moment survey asked participants to identify what
they did with the advertising mail, catalogs, and flyers they received. The choices
included keeping the mail, discarding the mail without reading it, or reading the
mail and then discarding it. The next table presents both the retention rate and the
discard-without-reading rate for door, curb, and cluster box recipients. The table
shows that cluster box recipients had the lowest retention rate and the highest
discard-without-reading rate:
Handling of advertising mail, including catalogs and flyers
Receptacle type Proportion kept (%) Proportion discarded without reading (%)
Cluster box 20.9 38.1
Curb 25.7 34.9
Door 30.9 31.8
Using the statistical procedure described in the previous section, it is possible to

test whether the retention rate for cluster box recipients was significantly less than
for door recipients or curb recipients. It is also possible to test if the cluster box
recipient discard rate was significantly greater than the discard rate for door or curb
recipients. The next table presents the calculated test statistics for those tests (n.b. a
positive calculated test statistic implies the cluster box rate is smaller and a negative
calculated test statistic implies that the cluster box rate is larger).
Testing the hypotheses of different retention and discard rates for cluster box recipients
Hypothesis Retention rate test statistic Discard rate test statistic
Curb ¼ cluster box 4.89 2.91
Door ¼ cluster box 8.61 5.03
If the calculated test statistic is larger (in absolute value) than the critical value of
1.96, then the hypothesis of equal retention or discard rates can be rejected at the
95 % level.11 In all cases, the null hypothesis is rejected. Thus, the Mail Moment
survey data support the inference that cluster box recipients have lower retention
rates and higher discard rates that either door or curb recipients. This is consistent
11
This level of significance is for a two-tailed test. A critical value of just 1.65 needs to be
exceeded to reject the null under a one-tailed test.
with the USPS OIG-InfoTrends result for door and cluster box recipients, but
different from the USPS OIG-InfoTrends result for curb and cluster box recipients.
The USPS OIG-InfoTrends survey found that curb and cluster box recipients had
similar retention and discard rates.
One important question in the Household Diary Study Recruitment Question-
naire was geared to how recipients responded to advertising mailings. It asked
participants to indicate how much attention members of the household pay to
advertising material received through the mail. The options are: (1) usually read
it; (2) usually scan it; (3) read some but do not read others; and (4) usually do not
read it. The following table lists the proportions of recipients, by receptacle type,
that reported usually reading their advertising mail.
Proportion of recipients that reported they usually read their advertising mail
Receptacle type Proportion (%)
Cluster box 13.3
Curb 16.1
Door 18.3
In other words, the Household Diary data indicated that cluster box recipients
had a significantly lower advertising mail read rate than door and curb recipients.
This matches the USPS OIG-InfoTrends results showing similar differences
between cluster box and door recipients but shows more differences for cluster
box and curb recipients than the OIG-InfoTrends results.
The data from the two USPS surveys of mail recipients corroborates the main
result found in the USPS OIG-InfoTrends survey: cluster box recipients are less
likely to read and retain their advertising mail than door recipients. The USPS
OIG-InfoTrends data showed little difference between cluster box and curb read
and respond rates, but data from the two USPS surveys indicated that cluster box
recipients had lower read and respond rates than did curb recipients. Finally, the
postal surveys produced evidence that cluster box recipients were more likely to
discard their advertising mail without reading it than were either door or curb
recipients.
5 Interpreting Alternative Hypotheses
The evidence from the USPS OIG-InfoTrends survey indicates a different pattern of
interactions with advertising mail for those who receive advertising mail in cluster
boxes rather than in other types of mail receptacles. However, further investigation
is warranted to confirm that these observed differences are due to the difference in
mail receptacle rather than underlying variables, such as attitude toward advertising
mail, household age, household income, population density, type of residence, or

Internet use that might be correlated with cluster box use.12
One possible reason for the differences in advertising mail responsiveness is that
cluster box recipients, apart from the way in which they receive their mail, simply
do not like advertising mail. To investigate this hypothesis, an analysis was
performed on an USPS OIG-InfoTrends survey question that asked the participants
to indicate how they feel about the advertising mail they receive. Participants were
asked to rate their view of advertising mail on a scale from strongly liking it to
strongly disliking it.
The following chart presents the results for door, curb, and cluster box recipi-
ents. Very few people strongly like their advertising mail, and relatively few people
strongly dislike it. More importantly, the general pattern is the same for all three
receptacle types:
Attitudes Toward Advertising Mail Received

45
40 Door
% Naming Each Answer
35 Curb
30 Cluster Box
25
20
15
10
5
0
Strongly like Somewhat Indifferent Somewhat Strongly
like dislike dislike
Possible differences among delivery modes can be quantified by forming

an index of advertising mail appeal by assigning the categories a numerical value,
ranging from a “1” for strongly liking to a “5” for strongly disliking. The calculated
index values for the different receptacle types are Cluster Box and Curb: 3.3,
Door: 3.1.
As noted in Sect. 2, door recipients have the lowest value, indicating the
strongest preference for advertising mail, but the differences are small and would
not appear to be sufficient to explain the significantly lower read and respond rates
12
It is also important to note that this investigation is limited to survey data examining postal
customers’ attitudes and self-reported actions. These analyses are not intended to investigate
causality. As detailed below, the survey design and analyses controlled for other factors, but we
have not constructed a causal model of cluster box recipients’ interactions with advertising mail
and do not claim to have done so.
for curb recipients. (In fact, the inconvenience cluster box recipients face when
disposing of any unwanted advertising mail could cause these differences.) This
factor could also explain why the index value for curb recipient is higher than for
door recipients.
The average values for key demographics including age, income, and education
are similar for door and cluster box recipients, ruling out these variables as potential
causes for different response rates. However, nearly all cluster box respondents
indicated they had Internet access (96.7 %), while just over three-quarters of door
respondents had Internet access (76.8 %). This raises the possibility that the
observed difference in response rates between door and cluster box recipients
may have to do with Internet access and not mail receptacle type. To test this, the
analysis that was performed on the full set of data is repeated for just the 4403
respondents that said they had Internet access. In other words, all of the door
recipients and cluster box recipients included in this test reported having Internet
access, so differential Internet access can be ruled out as a source of mail respon-
siveness differences.
The next table provides the results of repeating the statistical tests for Internet
users. As with the complete sample, the test for read and respond rates showed that
among Internet users, the door response rate was greater than the cluster box
response rate for all mail categories. This indicates that the differences between
door and cluster box read and respond rates are not due to differential Internet
access.
Testing the response and discard rates between door recipients and cluster box recipients for
internet users
Mail category Testing read and respond Testing do not read and respond
Known local business Reject at 95 % Reject at 90 %
Known national business Reject at 95 % Reject at 95 %
Unknown business Reject at 95 % Reject at 95 %
Donation solicitation Reject at 95 % Reject at 95 %
Credit card solicitation Reject at 95 % Reject at 95 %
Catalogs Reject at 95 % Reject at 95 %
Mail that includes a coupon Reject at 95 % Reject at 90 %
The test for discard rates provides even stronger evidence of differences between
cluster box recipients and door recipients. When all responses were used, there were
two categories, “advertising mail from companies with which I do business” and
“mail that includes a coupon” for which we could not reject the null hypothesis of
equal discard rates. Using the Internet access only subsample, the hypothesis for
both of those categories is rejected at the 90 % confidence level. In sum, there is no
evidence that difference in Internet access is the basis for the observed differences
in advertising mail response between door recipients and cluster box recipients.
Data from the two USPS surveys were also investigated to see if there were any
evidence of variables other than the mail receptacle causing the observed differ-
ences in interaction with advertising mail. The Household Diary Study survey
indicated that the type of residence in which the recipient lives did not appear to
affect how cluster box recipients interact with advertising mail. A smaller propor-
tion of cluster box recipients than door recipients read their advertising mail,
whether both groups lived in single detached houses, apartments and condomin-
iums, or townhouses and duplexes.
The two USPS surveys did not produce data that permitted testing the hypothesis
as to whether a greater affinity of Internet use among cluster box recipients was the
source of difference in responsiveness of cluster box recipients as compared to door
recipients. However, the USPS OIG-InfoTrends survey data did support such an
analysis; results of testing indicate that the differences remained, even when just
comparing Internet users.
6 Conclusions
The USPS OIG-InfoTrends survey produced statistically significant evidence indi-

cating that cluster box recipients report a weaker connection to advertising mail
than that of door recipients. There were far fewer differences between curb and
cluster box recipients than there were for door and cluster box recipients.
Results from analyses of the survey data indicate that cluster box recipients
reported significantly lower read and respond rates than door recipients for all types
of advertising mail. In a number of instances, the cluster box read and respond rates
were half of those for door recipients. Cluster box recipients also reported higher
discard without reading rates than door recipients. Additional statistical testing was
performed to verify that the lower read and respond rates for cluster box recipients
did not result from another variable (i.e., recipient age, income, type of residence, or
Internet access). In all cases, the evidence indicated that the reported differences in
read and respond rates did not result from alternative demographic or attitudinal
variables.
The results of this study should be of concern to USPS, mailers, and even mail
recipients, who want to receive their mail at home. While our study does not
diminish the importance of centralized delivery to USPS cost savings efforts, it
does suggest that some of the revenue from advertising mail could be at risk if door
delivery is changed to cluster box delivery and, to a lesser extent, to the curbside. It
would be wise for USPS to step up its customer research in this area to understand
more clearly the potential implications of substituting alternatives to door delivery.
Part of that effort should be engagement with advertising mailers. Additional
information on the impact of mail delivery modes on response rates would help
mailers make more informed decisions about whom to target in their direct mail
advertising campaigns.
Finally, it should be reiterated that we examined mode of delivery’s potential

impact(s) on advertising mail in broad terms. Certain types or locations of cluster
boxes might mitigate some of the adverse behaviors we discovered. Both USPS and
advertising mailers would benefit from having additional data on modes of delivery
as soon as possible, for advertising mail and other mail products. We recommend
further work examining modes of delivery in the United States and elsewhere.
References
Bradley MD, Colvin J, Perkins MK (2015) Targeting versus saturation: derived demand for direct
mail. In: Crew MA, Brennan T (eds) Postal and delivery innovation in the digital economy.
Springer International, Cham, pp 65–76
Canada Post (2013) Five point action plan. https://www.canadapost.ca/cpo/mc/assets/pdf/aboutus/
5_en.pdf
Rohr C, Trinkner U, Lawrence A, Kim CW, Potoglou D, Sheldon R (2013) Measuring consumer
preferences for postal services. In: Crew MA, Kleindorfer PR (eds) Reforming the postal sector
in the face of electronic competition. Edward Elgar, Cheltenham, pp 241–260
The Boston Consulting Group (2010) Projecting US mail volumes to 2020, final report—detail.
https://about.usps.com/future-postal-service/bcg-detailedpresentation.pdf. Accessed 2 Mar
2010
US Government Accountability Office (2004) USPS needs to clearly communicate how postal
services may be affected by its retail optimization plans. GAO report, GAO-04-803. http://
www.gao.gov/new.items/d04803.pdf
US Postal Service Office of the Inspector General (2012) State of the mail. Report no. RARC-WP-
12-101. https://www.uspsoig.gov/sites/default/files/document-library-files/2013/rarc-wp-12-
010.pdf
US Postal Service Office of the Inspector General (2015) Advertising mail and customer engage-
ment by mode of delivery. Report no. RARC-WP-15-009. https://www.uspsoig.gov/sites/
default/files/document-library-files/2015/rarc-wp-15-009.pdf
Commercial and Regulatory Challenges
for Postal e-Services in Switzerland
Christian Jaag, Martin Maegli, and Denis Morel
1 Introduction
Swiss Post has been innovating for several years, strengthening its expertise as an
actor in the digital world and exploiting the unique selling propositions it owns in
the physical world. The emergence of the Internet in combination with gradual
liberalization has given rise to new customer needs, increased and changing com-
petition as well as new business models in the postal sector. The rationale for Swiss
Post’s investment in postal e-services is twofold: First, new services may enhance
the value of traditional services by adding complementary services; second, they
may compensate losses due to the progressing substitution of physical letter mail
which progresses at a rate of about 2 % per year.
This chapter documents digital postal services in Switzerland in the context of
the postal regulatory and business environment. The chapter describes the devel-
opment of postal e-services in Switzerland, puts it in a commercial and regulatory
perspective and provides an outlook to future developments. The remainder of the
chapter is structured as follows: Sect. 2 discusses the challenges and opportunities
for Swiss Post’s core business due to electronic communication, Sect. 3 presents
Swiss Post’s approach while Sect. 4 highlights regulatory aspects. Section 5
concludes.
The views expressed here are those of the authors and do not necessarily reflect those of the
organizations with which they are affiliated.
C. Jaag (*)
Swiss Economics, Zurich, Switzerland
e-mail: christian.jaag@swiss-economics.ch
M. Maegli • D. Morel
Swiss Post, Bern, Switzerland

232 C. Jaag et al.
2 Challenges and Opportunities for Swiss Post’s Core

Business Due to Electronic Communication
With 280 addressed letters annually per capita, Swiss mail volumes are still high.
However, they are continuously declining at a rate of approximately 2 % per year.
The loss of economics of scale as a result of the decline in letter mail puts Swiss
Post’s traditional business model at risk. Additionally, electronic and hybrid ser-
vices are increasingly targeted at recipients which may undermine the senders’
interests. However, parcel volumes are increasing as a result of growth in
e-commerce.
2.1 Regulatory Background
In the course of the most recent revision of the postal law in the years 2010–2013,
Swiss Post was converted into a corporation under special law, as foreseen in the
new Postal Organization Act. At the same time, Swiss Post’s division for financial
services, PostFinance, was divested and placed under the control of the Swiss
Financial Market Supervisory Authority as a limited company under private law.
Former PostReg was replaced by PostCom, which has a firm legal foundation and
clear responsibilities for the supervision of universal postal services.
Today, Swiss Post is an autonomous public corporation, owned entirely by the
Swiss Confederation. It operates within the institutional limits laid down by the
federal legislation. The Federal Council not only determines the scope of the postal
universal service obligation as defined in the Postal Act, but also defines the
strategic objectives of Swiss Post every 4 years (Swiss Federal Council 2012).
Postal legislation is directly derived from Article 92 of the Swiss Constitution,
which states that “The Confederation is responsible for postal and telecommunica-
tions services. The Confederation shall ensure the adequate, universal, and reason-
ably priced provision of postal and telecommunications services in all regions of the
country. The rates shall be fixed according to standard principles.”
Postal law is relevant for digital postal services in mainly two respects. First, the
Postal Organization Act, together with the strategic objective of the Federal Coun-
cil for Swiss Post, defines the scope of Swiss Post’s business activities. The act
states that Swiss Post is allowed to conduct any business related to the conveyance
of letter and parcel mail (in addition to certain financial services and activities in
public passenger transport). The Federal Council expects Swiss Post to develop its
services, especially new services, in accordance with changing consumer needs (see
Swiss Federal Council 2012).
The message of the Federal Council related to the Postal Organization Act states
that, based on its core business and within the boundaries of its legal mandate,
Swiss Post faces three guidelines. One is to diversify along its traditional value
chain by integrating up- and downstream activities geographically in Switzerland
Commercial and Regulatory Challenges for Postal e-Services in Switzerland 233
and abroad by utilizing economies of scope. Another is to position itself in

converging postal communications and logistics markets. A third is to reduce its
dependence on traditional and potentially regressive services. These guidelines
define the scope of digital postal services which Swiss Post is allowed to offer.
Second, the Postal Act defines the scope of the universal service obligation with
respect to services offered and daily nationwide delivery. The Postal Act entitles the
Federal Council to stipulate alternative forms of mail delivery. However, this
possibility was not concretized in the Ordinance on the Postal Act. Hence, postal
regulation asks for a very high-quality provision of physical postal services to the
Swiss population and businesses. Thereby, it strongly affects the commercial
viability of electronic delivery which relies on potential cost savings (through a
reduction in physical delivery) and consumer’s willingness to pay for such services.
Both of these depend negatively on the recipients’ entitlement to physical delivery
services which is very strong in Switzerland (daily delivery to the doorstep for most
of the Swiss population), see Jaag and Dietl (2011).
In sum, the narrow scope of Swiss Post’s legally defined activities in connection
with a high-quality level of traditional postal services—and therefore a low demand
for alternative services from customers—represent quite a harsh environment for
the successful development and marketing of digital postal services.
2.2 e-Law
The scope of legislation related to digital topics and services concerns a broad range
of topics and covers subjects as data protection, digital identities or certificates,
secure communication, and sharing information under sound legal protection. The
field of e-legislation is rather young compared to other traditional and historically
grown legal frameworks. In Switzerland, legal frameworks have started to emerge
at the federal level. However, it is the cantons that define the implementation of the
legal framework in detail. Therefore, the final framework and the status of legal
implementations can and do differ from canton to canton, hindering nationwide
harmonization. This county-specific peculiarity harms the indispensable interoper-
ability between different systems in the field of electronic solutions.
e-Health is one example for digital services in a specific sector. Some cantons
already have a specific legislation for e-health (e.g., Geneva). A framework legis-
lation on the federal level—with data protection concerns addressed as the main
topic—is expected soon.
Other examples include e-justice and other e-government. The legal rules
governing (postal) activities in these fields are diverse and fragmentary. In Swit-
zerland, formal requirements on the transfer of legal documents (among others
related to digital signatures) in a legal procedure are defined in the specific context
of various codes of procedure at different legal hierarchy levels. In addition to these
diverse and very scattered specifications, the Swiss Code of Obligations states in
Art 14 para 2bis that a qualified electronic signature based on a qualified certificate
234 C. Jaag et al.
Table 1 Specific projects and legal frameworks in e-Law

Topic Legal framework Status
Data protection Federal Act on Data Protection Enacted; main framework will be
(FADP) revised
Certificates and Federal law on certification Under revision
legal binding services and electronic signa-
signatures tures (ZertES)
Health data Federal Act on the electronic Draft law under discussion in
patient data (EPDG) parliament
Digital identity Bill on digital identities Prospective draft law
e-Commerce No specific framework and Harmonization with EU standards in
covered by established consumer protection discussed in fed-
legislation eral administration
Secure exchange of No specific framework Parliamentary letter of enquiry; exami-
legal data and nation by federal council
Information
e-Voting Ordinance an political rights Enacted; will be revised soon
and ordinance on electronic
voting
issued by an accredited certification service provider is deemed equivalent to a

handwritten signature. The prerequisites and the procedure for the official recog-
nition of these certification service providers are regulated in specific acts (ZertES
and corresponding amendments).
Table 1 provides an overview of current projects and frameworks in Swiss
e-Law.
The electronic equivalent to the physical signature is an important prerequisite
for every kind of electronic communication in the public and the private sector. To
that end, the combination of the legal sources mentioned above would have to
represent a consistent and conclusive framework. However, many legal uncer-
tainties and risks related to the acceptance of electronically signed documents
remain, especially concerning the legal effects of an electronic message and
international recognition. Moreover, the standardization of electronic infrastruc-
tures, signatures, and identities is still in the early stages and is heavily depended on
the regulatory evolution on the international and especially European level. This
constitutes an effective barrier for the development of digital postal services which
necessitate clear rules on the legal status and consequences of sending, conveying,
and receiving digital messages. This barrier impedes Swiss Post’s development, but
also its competitors’ (e.g., those potentially providing authentication of documents
by email).
2.3 Economic and Geographic Background
Swiss Post’s business strongly relies on the Swiss economy in which it is an import
enabler for trade and one of the biggest employers. Until the end of the 1990s, mail
and parcel volumes used to grow in parallel with GDP. Even though this direct
relationship no longer holds, general economic activity remains an important
driving force for mail and parcel volumes as well as other postal businesses.
Switzerland mostly extends across the north side of the Alps. It contains the
mountainous area in the south (Alps), the Swiss plateau (middleland), and the Jura
Mountains in the north of the country. The average population density is around
190 people per square kilometer, with a considerably lower population density in
the mountainous regions of the Alps which comprise about 60 % of the country’s
total area.
The heterogeneous population and settlement structure causes also the cost of
postal service provision to be quite different across regions. This facilitates direct
competition by firms that are able to selectively enter the market with a focus on
business customers and the most densely populated areas.
2.4 Postal and Telecommunications Infrastructure

in Switzerland
The distribution of the Swiss population across the country is closely mirrored by
Swiss Post’s network of physical access points which is very dense in the middle
and less so in the rather sparsely populated areas (e.g., in the Alps, see Fig. 1).
Compared to other European countries, the network of access points is very high,
Post offices
Agencies
PostFinance branches
Home delivery service
PickPost points
Fig. 1 Swiss Post’s network of access points. Source: Swiss Post

236 C. Jaag et al.
Fig. 2 Broadband coverage (download bandwidth >10 Mbit/s). Source: Bakom (2014)
measured in relation to the population and the area of the country. The high
proximity of postal access points to the population is due to regulation in the postal
sector, and it is an important basis for Swiss Post’s strong position and its reputation
in the Swiss population.
Switzerland’s geography and the distribution of the population are also reflected
in the Swiss broadband coverage. Figure 2 shows that accessibility of broadband
services is distributed very much the same way as the postal outlet network. High
bandwidths are driven by consumer needs rather than regulation as the telecom-
munications universal service only prescribes bandwidths up to 1 Mbit/s.
Broadband penetration has been increasing continuously during the past years
and is almost ubiquitous by now. Hence, electronic means of communication are
available to the entire population. They enable electronic commerce and allow for
services which represent a close substitute for physical mail. In an empirical study
of mail demand in Switzerland, Buser et al. (2008) find that the emergence of
electronic substitutes reduced mail demand. Therefore, the telecommunications
infrastructure, with the services provided on it, is important determining factor
for postal services.
The overview of economic and geographic conditions along with the postal and
communications infrastructures shows that demand for digital postal services is
driven by the accessibility of electronic communications means and customers’
needs. The supply of digital postal services relies on the available technology and
the regulatory framework within which Swiss Post operates.
3 Universal Service Regulation and Electronic Delivery
Business cases for digital postal services strongly depend on the regulatory and
legal environment in Switzerland. Whereas the traditional postal business in letter
mail and parcels is highly regulated and subject to sector-specific law, the legal
framework for electronic services is still quite loose. The narrowly defined scope of
Swiss Post’s business in combination with a strong universal service obligation for
physical services make it hard to successfully develop and market digital services as
complements or substitutes. Weak regulation of standards for electronic exchange
imposes considerable legal uncertainty and risk on digital postal services.
The challenges to postal strategy lead to the question of how postal regulation
affects such a modification and transformation of business models. Regulators
determine the incentives and possibilities of business model transformation in the
postal sector through a variety of regulatory instruments.
Consider the following example from Switzerland: Swiss Post offers its cus-
tomers the “Swiss Post Box” service. Whenever Swiss Post collects and sorts a
letter addressed to the receiving customer, the envelope is scanned and emailed to
the customer’s cell phone. The customer then has the options to have the letter
opened and scanned, recycled, archived, or delivered to the physical address. Thus,
not all letter mail needs to be delivered physically. As a result, there are some cost
savings, especially in rural areas, on days with only one letter per household if it is
delivered electronically. Then, individual buildings do not have to be served daily.
However, with a strong physical USO in place, demand for electronic delivery is
low since physical delivery to the doorstep is free for receiving customers. Hence,
cost savings are minimal and the viability of hybrid services would greatly benefit
from a technologically neutral formulation of the USO. One possible implementa-
tion could be a joint definition of universal postal and telecommunications services:
Every household must be served daily, either with postal or broadband telecom-
munications services and a Swiss Post Box. If the latter is offered, the frequency of
physical delivery may be reduced to three times per week. Then, offering hybrid
services would allow Swiss Post to save a larger part of the cost of physical
delivery.
A technologically neutral USO means that the focus is on the satisfaction of
consumer needs, not on the technology used to achieve it (see Jaag and Trinkner
2012). For example, the main needs of recipients concerning postal services are
physical and timely delivery. The technology used by the operator is of little
interest. Reverse Hybrid Mail services, such as Swiss Post Box, improve physical
delivery—they are the secure electronic complement to the physical mailbox. As a
prerequisite for such a service, broadband and mobile penetration rates have to
reach a critical mass. Then, universal services can become a technologically neutral
multichannel concept which is built on the original idea of the USO: to safeguard
the public’s access to a defined range of basic services.
With the concept of a communications USO, no matter how quickly communi-
cation technologies change, the right to a minimum level of service quality is of
238 C. Jaag et al.
high importance for the economic development of a society. Therefore, a unified

definition of the universal service consists of the basic principle of having the
possibility to communicate from senders to recipients irrespective of whether it is
physical or electronic. Only a technologically neutral USO allows the postal USP to
save the cost of physical delivery if it offers electronic delivery instead. This
strengthens the business case for digital postal services that may not be commer-
cially viable otherwise.
4 Swiss Post’s Approach
As shown in the previous section, the regulatory and legal environment in Swit-
zerland is not ready for many viable business cases in universal digital delivery (see
Maegli et al. 2011). For this reason, Swiss Post’s diversification focuses on those
digital postal services which profit from the high trust of Swiss Post (see Morel
2011). However, the digital postal services’ revenue and profit contribution is still
very low compared to Swiss Post’s total range of business. Digital services can
directly linked to a traditional service and enhance its value or add a distribution
channel. Alternatively, independent services may provide new platforms that rep-
resent real diversification away from traditional postal services.
4.1 Digital Services Linked to Traditional Services
In the last few years, Swiss Post developed a lot of digital services directly linked to
traditional services. This permits Swiss Post to enhance their value. The typical
example for such an example is electronic information exchange between courts
and Swiss Post for the physical delivery of the courts documents. This
e-government service ensures the quality of a physical process by exchanging
transaction information electronically. The delivery of court documents is a special
physical service that is defined by the Postal Act. It requires very high traceability
and quality of delivery, since the service must ensure that the recipient receives the
document in person.
Swiss Post has implemented a process that facilitates the electronic exchange of
delivery information directly with court systems. This electronic integration
ensures high traceability of delivery for the court. As Swiss Post generates the
receipt, the unique identifier and the barcode, it can ensure the delivery standard
throughout Switzerland. Delivery quality is thus increased, and can better match
legal requirements.
The interest of Swiss Post for the digital service is not to increase the price of the
service, but due to this automatic exchange it is to reduce the operative costs and to
increase the quality and the traceability.
A second example is given by the new service ePostOffice. Many companies and
organizations, particularly public administrations, rely on solutions that enable
them to transmit information in a reliable manner (e.g., notifications about tax
matters or documents related to the next vote or election). To date, this could only
be accomplished through physical delivery performed by Swiss Post. The delivery
could not be made to an e-mail address.
With ePostOffice, the issue could be solved by Swiss Post. The recipient can
decide if he or she wants to receive letters from an originator physically or
electronically. The choice can be different for each originator. The originator
sends the mailing batch electronically to Swiss Post, without knowing the delivery
choice of the recipient. Swiss Post assumes the role of an intermediary between the
originator and the recipient of a mail item, and takes responsibility to filter the
mailing batch according to the recipient’s chosen delivery option and to send to
document physically if the recipient cannot be addressed electronically anymore or
if the recipient did not access the document electronically.
With ePostOffice, the originator can then send mail at the same service quality as
that of a physical delivery. The delivery options for the recipient are neither
disclosed to the originator, nor open to the originator’s influence. Only the origi-
nator knows the physical address of the recipient. Swiss Post provides a safe and
undisclosed link between the physical address and the electronic delivery.
The business case depends on the service offered and the conditions related to
it. In the first example, the use of the digital component permits Swiss Post to
increase the quality of delivery. Hence, process optimization permits to save costs
for Swiss Post and its customers.
In the second example, the sender is paying a price which is independent of the
delivery channel chosen by the recipient. The price is similar to a letter’s price, but
processes are much simpler for senders: they do not incur the additional cost to
manage the various channels for communicating with their customers.
4.2 New Digital Services
Swiss Post started also the implementation of new digital service in order to extend
his portfolio and to find new markets extended the traditional markets of letters and
parcels. In this section, such examples are provided.
4.2.1 SuisseID and Yellow Identification: The Link Between a Person

and Their Digital Identity
In the digital world, it is important to ensure that a digital identity is correctly

matched one to one to each individual person in order to avoid identity theft and
related fraud. This is particularly relevant for e-government services, which permit
240 C. Jaag et al.
a digital identity to get access to and change personal information (this applies, for
example, to tax accounts or electronic votes and elections).
The Yellow Identification service enables individuals to acquire an official proof
of identity. Individuals can have their Yellow Identification performed at a postal
office using an official photo-based identification document. Swiss Post carefully
checks the identification document in the presence of the applicant. By means of the
“Original document seen by” stamp, Swiss Post confirms that a copy of the
identification document is based on the original. For example, this service is used
for the opening of bank accounts by PostFinance, and complies with the Swiss
Money Laundering Act. Because the service can be performed at every postal
office, it is a simple procedure for any individual.
The SuisseID constitutes the next step after the Yellow Identification. The
SuisseID is a high-quality digital identity, and a qualified digital signature1 that
can be used to electronically sign electronic documents or forms. Swiss Post pro-
duces the SuisseID according to the standards defined by Swiss Federal Law on the
electronic signature. To acquire a SuisseID, an individual must visit a post office in
order to perform the Yellow Identification and to link his or her physical person to
the SuisseID (i.e., to the related digital identity and digital signature).
The SuisseID is used to access e-government services such as online adminis-
tration portals. Individuals do not need to apply to an administrative office in order
to officially identify themselves, as this was already taken care of by Swiss Post
when the SuisseID was acquired. This service provided by Swiss Post enables
public administrations to outsource the identification process, which would other-
wise be costly for both the applicant and the administrative office.
4.2.2 Swiss Post’s Electronic Health Record Platform
e-Health is a new market that Swiss Post has addressed with its services for a
number of years. Based on its innovative solutions, Swiss Post has implemented an
electronic health record platform that is already in use in various cantons. This
section will explain Swiss Post’s role and USP in the e-health market.
A central element of the definition of the electronic health record in Switzerland
is a virtual electronic health record that allows the electronic management of health
information of a patient by health actors. The electronic health record ensures a safe
and efficient exchange of medical information relating to individual patients, and
thereby creates value for both healthcare providers and patients. Its implementation
should comply with the following main principles:
Security: Patient security is the core requirement. Thus, the security, quality, and
integrity of the data must be ensured.
Trust: Each exchange happens in a trusted environment, where the protection of the
patient’s privacy is safeguarded.
Data protection: The electronic health record contains all the mechanisms that
ensure the system’s compliance with data protection laws.
Fig. 3 Swiss Post’s electronic health record platform “vivates.” Source: Swiss Post
Local: Data should not be centralized. The system ensures central access to local
medical data that pertains to specific health actors.
Switzerland is a federal state, and accordingly the health sector is under the
responsibility of the cantons. The Swiss Government defines rules and frame
conditions, while the cantons are responsible for the governance and implementa-
tion of solutions in the health sector within their territories. In order to satisfy the
above principles and to comply with the federal state this market requires one or
many Swiss-wide players to have the necessary trust.
Swiss Post is an important actor in the health sector today, thanks to its vast
expertise regarding the exchange of medical documents through physical postal
services. The electronic health record has begun to complement and replace this
physical exchange process with secure electronic exchange processes. Swiss Post
decided early on that it would play a role in this emerging market by providing
cantons with an electronic health record. Together with the Canton of Geneva,
Swiss Post implemented an electronic health record (under the name of “vivates”).
When the system of Swiss Post was introduced in 2011, it was the first Swiss
electronic health record pilot project. Today, Swiss Post manages three cantonal
pilot projects. Figure 3 shows the model developed by Swiss Post.
One question arises regularly: Why is Swiss Post active in the e-health sector,
and what is its role? Geneva’s electronic health record, the first Swiss pilot project,
serves as an adequate answer to this question, and confirms the legitimacy and
success of Swiss Post in the hitherto nontraditional e-health market.
One of the main success factors is Swiss Post’s neutrality. The health sector is
very complex, comprising very different and diverging interest groups. Swiss Post
is a neutral provider and has no interests related to the health sector other than to
242 C. Jaag et al.
provide a system for the safe and efficient exchange of information. For that reason,
health actors are willing to enter and manage medical data within Swiss Post’s
electronic health record platform. If health records were to be provided by one of
the actual actors in the health sector, it would meet with resistance from the other
actors (for instance, if a hospital set up such a platform, not all doctors would agree
to participate).
Another important success factor is the trust that health professionals, the
government, and the population put in Swiss Post. The health actors are confident
that Swiss Post supplies a system that complies with legal provisions, and with all
security and data-privacy requirements.
Swiss Post always provides a clear and transparent explanation of the scope of its
business and expectations. It will not interfere with medical processes, but will
provide all the necessary means by which to optimize and support these processes
through a safe and efficient electronic exchange of documents. Swiss Post verifies
the identification of users, along with their authorization to access the medical data,
as well as the secure transport of this data. The responsibility of publishing the
medical data and its use remains clearly in the hands of the health professionals.
All these components ensure that the integration of Swiss Post’s electronic
health record as noninvasive as possible. The processes within the organization,
and the management of medical data and documents, are not interfered with or
changed. By complying with international standards, the implementation of tech-
nical interfaces bears only minimal consequences.
4.2.3 New Digital Services and Business Models
The services described above have not yet proven to have a business case that can
compensate the reduction of the revenue of the classical physical business. Each
service has a defined pricing and business model, but the evidence of the necessity
of such products on the market has to be still proven.
In the case of the e-health, it is necessary to bring the evidence to the actors and
the deciders that such a system brings an optimization of the processes by the actors
or to the health system. Only with this evidence the actors will get the willingness to
pay for such a service. This process requires an investment of Swiss Post.
5 Trade-Offs in Developing e-Postal Services
Since digital postal services are very diverse and many of them have been intro-
duced only recently, there is not yet a clear business case or even a recipe for
success. For Swiss Post, there are three fundamental trade-offs in developing digital
postal services:
Trade-off in content: Those services which are closely related to traditional
services are currently most successful. However, they often do not represent real
diversification and may become obsolete together with their physical counterparts,
especially in the communications market. More radical innovations rely on adapted
consumer expectations and may not fit well with Swiss Post’s core competencies.
There is not yet a proven and viable business case.
Trade-off in timing: To some degree, many digital services substitute traditional
postal services. Hence, pushing digital services may accelerate the transformation
and destruction of the Post’s traditionally profitable business. However, Swiss Post,
doing nothing, could not stop or retard the transformation. Additionally, in the mid
to long term, this substitution through the emergence of new communication
channels and business processes is inevitable. There is a considerable risk of joining
too late.
Organizational trade-off: Proximity to consumers and knowledge of their needs
is key to successful innovation. This is a challenge especially for a postal operator
that only transports letters, but does not need to have the knowledge about the
customers’ processes associated with their mail stream. This is a knowledge that
Swiss Post has acquired through close cooperation and process integration with
clients. Hence, it makes sense to develop digital postal services in a decentralized
manner in the various business units of Swiss Post. However, the development of
digital postal services in business units will be biased in favor of add-on services
which relate to existing services. For independent services, which may cannibalize
existing services, a structure besides the business units may be more appropriate. At
any rate, a certain degree of coordination is necessary to benefit from synergies and
to reach critical mass in new markets. Moreover, the transformation of Swiss Post
into a new role in the digital world necessitates an extensive and coordinated effort
beyond product innovation.
6 Conclusion
Swiss Post offers a broad range of digital postal services. The development and
marketing of these services is decentralized across all business units. Thanks to its
financial strength, Swiss Post is able to finance its capital expenditure (e.g., for new
business) with its operational cash flow. Most of the digital postal services are
directly linked to a traditional service and enhance its value or add a distribution
channel. This allows to profit from time-tested consumer habits and expectations to
pay by increasing the willingness to pay for established products. Few digital
services represent real diversification away from traditional postal services except
for e-health, these have not yet proven to have a viable business case, as the
willingness to pay for stand-alone services is low and the development of compe-
tition is uncertain.
The business case for digital postal services strongly depends on the regulatory
and legal environment. The narrowly defined scope of Swiss Post’s business in
combination with a strong universal service obligation for physical services makes
it hard to successfully develop and market digital services as complements or
244 C. Jaag et al.
substitutes. Weak regulation of standards for electronic exchange imposes consid-

erable legal uncertainty on digital postal services.
The analysis of digital postal services in Switzerland reveals that digital services
already permeate traditional services and have become a matter of course. Most
services have been developed in close relationship to physical services to which
they add convenience and value. In addition, there are hybrid services and purely
digital services which are an expression of entire Swiss Post’s transformation away
from its physical core services. This transformation is still in its infancy and its
direction is still vague. As a result to a fast changing environment with new
technology, consumer needs and regulation, also Swiss Post’s organization of
digital postal services is subject to ongoing modification and adjustment.
From the outset, Swiss Post seems to possess the right assets to diversify into
electronic services that complement its traditional business or leverage its reputa-
tion and proximity to consumers. Physical mail has become more attractive due to
complementary electronic services. This may have contributed to the low pace of
the volume decline in Switzerland. However, the commercial success of postal
e-services has been somewhat impeded by a strong USO (which reduces demand
for new services), legal uncertainty with respect to the allowed scope of business
and generally an uncertain legal framework for electronic services.
What regulation is necessary to promote postal e-services? An important step
would certainly be a technologically neutral formulation of the postal USO. Cur-
rently, there are two communications infrastructures (postal and telecommunica-
tions) which fundamentally serve very similar purposes. A relaxation of postal
obligations related to services for customers who are served well with other means
of communication would remove the attractiveness of postal USO products and
hence stimulate demand for e-postal services. Additionally, a clarification of the
rules governing electronic services, e.g., in the health sector would be beneficial for
new services in these industries.
References
Bakom (2014) Breitbandatlas. http://map.geo.admin.ch/?topic¼nga&lang¼de&X¼190000.00&

Y¼660000.00&zoom¼1&bgLayer¼ch.swisstopo.pixelkarte-grau&
catalogNodes¼317,320,327,334,338
Buser M, Jaag C, Trinkner U (2008) Economics of post office networks: strategic issues and the
impact on mail demand. In: Crew MA, Kleindorfer PR, Campbell JI Jr (eds) Handbook of
worldwide postal reform. Edward Elgar, Cheltenham, pp 80–97
Jaag C (2013) Digital postal services in Switzerland, mimeo
Jaag C, Dietl H (2011) Postal and regulatory reform in intermodal competition. In: Sund K, Osborn
D (eds) The future is in the post vol 2: perspectives on transformation in the postal industry.
Libri, Faringdon, pp 49–56
Jaag C, Trinkner U (2012) Defining and financing an intermodal USO. Swiss Economics working
paper 35
Maegli M, Jaag C, Koller M, Trinkner U (2011) Postal markets and electronic substitution:
implications for regulatory practices and institutions in Europe. In: Crew M, Kleindorfer P
(eds) Reinventing the postal sector in an electronic age. Edward Elgar, Cheltenham
Morel D (2014) Swiss post and public administration: from physical to digital. In: Finger M,
Bukovc B, Burhan M (eds) Postal services in the digital age. IOS Press, Amsterdam
Swiss Federal Council (2012) Strategische Ziele des Bundesrates für die Schweizerische Post AG
2013–2016
Demand and Regulation for e-Commerce
in Goods
Heikki Nikali, Juri Mattila, Ilona Rintanen, and Ville Huuhtanen
1 Introduction
As paper communication volumes decline, postal operators (POs) are developing

parcel delivery-based growth in e-commerce to compensate for revenue lost due to
lower letter demand. Optimistic forecasts suggest that e-commerce will grow at an
exponential rate. Such projections, however, often overlook the fact that goods only
account for 40 % of the total value of e-commerce (Statistics Finland 2014a).
Furthermore, the high growth figures for e-commerce are also attributable to higher
value purchases than to higher quantity of purchases. Since 2000, the volume of
B2C e-commerce in goods has grown by 30 % in Finland, while the monetary value
has tripled. There is also a growing trend of online shopping that involves the
consumer picking up the goods themselves from a physical store. It would appear
that the revenue accrued from these services will only compensate for a small
proportion of the revenue lost as a result of the decline in paper communication.
Our chapter discusses the development of B2C e-commerce in goods in Finland
and the main factors driving its growth. The evolution in B2C e-commerce is
analyzed based on an econometric model and consumer interviews. The aim of
the interviews is to explore consumers’ reasons for shopping online rather than in
physical stores, and also to understand their experiences with online shopping
(Nikali 2014a; Elkelä 2015a; Elkelä et al. 2015). Using the new estimated
The views expressed in this chapter are those of authors and do not necessarily reflect the views of
Posti.
H. Nikali (*) • I. Rintanen • V. Huuhtanen
Posti Ltd., Helsinki, Finland
e-mail: heikki.nikali@posti.com; ilona.rintanen@posti.com; ville.huuhtanen@posti.com
J. Mattila
The Research Institute of the Finnish Economy (ETLA), Helsinki, Finland
e-mail: juri.mattila@etla.fi

248 H. Nikali et al.
econometric model, forecasts are provided to understand prospects in e-commerce

for the post and other goods transportation operators.
In addition, e-commerce growth has also generated more competition among
other logistic companies. However, postal transportation of goods is regulated
differently from other transportation of goods in international as well as national
regulations. Transportation laws are usually binding, and they may put POs and
other logistics service providers in an unequal position. For example, the liability of
an operator may be set extremely high in the domestic postal regulation compared
to the liability described in the road transportation regulation.
The content of this chapter is as follows: In Sect. 2, we look at the structure of
e-commerce, introduce different viewpoints for defining it, and explain why size
estimates considerably deviating from each other may be presented of e-commerce.
In Sect. 3, we analyze the evolution of e-commerce in goods and the factors driving
sales online, and present future prospects. Sect. 4 examines whether e-commerce
can compensate the loss of revenue from traditional paper communication. We
examine whether changes in regulation could promote fair competition in
e-commerce between postal services companies and other logistics firms. In
Sect. 5, we summarize the results and present conclusions.
2 Characterization of e-Commerce
e-Commerce is a highly multidimensioned phenomenon. Therefore, in order to give

a meaningful definition for it and to distinguish between e-commerce and tradi-
tional brick-and-mortar trade, several considerations have to be made. In each
online shopping event, it is possible to distinguish a number of different phases
that may or may not be included in a definition of e-commerce, including (1) search
for information, (2) order, (3) payment, and (4) delivery.
The digital characteristics of trade can manifest themselves in various different
combinations of these phases, all of which could still justifiably be categorized as
e-commerce. A person may, for example, compare household appliances online,
and discuss them with different salespersons through websites. After making a deci-
sion to purchase on the basis of the comparison, the purchaser may go to the selected
store to order, pay, and pick up the product they selected (1). Conversely, a person may
go to electronics stores to try out and compare different software and, after making a
decision to purchase, order, and pay the product online and download it electronically
for themselves at home (2–4). A customer of a pizzeria, in turn, may look for
restaurants close to them using a mobile phone and order a pizza from the company’s
website, but pay for it in cash when picking it up (1 and 2). A holiday traveler may
notice an experience service in a hotel’s holiday brochure, book it on the phone, pay by
credit transfer by way of an online bank, and enjoy the service on site (3).
In the delivery phases of online purchases, there is a large difference between the
delivery of goods and services. Naturally, postal and logistics operators specialize
in the delivery of goods. They transmit a physical object, i.e., an artifact, which may
Demand and Regulation for e-Commerce in Goods 249
in this connection also be termed a consignment in a letter, a parcel, or groupage

(Elkelä 2015b). In a service delivery, digital content is transmitted, e.g., a digital
audio file or a printable voucher to use the purchased service, such as in the case of a
travel ticket.
In addition to the phases of the process, the transacting parties must also be
defined. The most natural and also the most common area of trade to be included in
the definition of e-commerce is business-to-consumer (B2C), where the seller is a
business and the purchaser is a consumer. Other areas of trade that are sometimes,
but not always, included in the definition are business-to-business (B2B)—either
inclusive or exclusive of trade based on electronic data interchange (EDI)—cus-
tomer-to-customer (C2C) and business-to-government (B2G).
It is also necessary to recognize geographical aspects, including not only domes-
tic e-commerce, but also both exports and imports. In doing so, one has to choose
the criteria for defining the location. Possible criteria for the vendor include where
the company faces tax liability, its ownership base, or the location of its head office.
For the consumer, the definition may, for example, be based on the person’s
citizenship or physical location, or the source of payment. Finally, in addition to
legal e-commerce, one’s definition may also include rapidly growing illegal
e-commerce and the gray economy. One must also determine whether to include
taxes, delivery costs, and product returns in their figures on e-commerce.
Taking into account the multidimensional nature of e-commerce, it is thus not at
all surprising that very different results and estimates of various sizes concerning
the scope of the phenomenon abound. Trade is becoming digitized in different ways
in different industries, depending on the special characteristics of each industry, and
formulating one universal and consistent definition is in no way easy. In worst
cases, parties discussing e-commerce may completely talk past each other if their
views of the definition of e-commerce are very different from one another. It is,
therefore, always important, when examining e-commerce, to be aware of how the
phenomenon is defined in the very context at hand.
In our study, we examine e-commerce of goods in the B2C segment in Finland,
measuring it by the number of deliveries and taking into account domestic trade and
imports. Our examination focuses on legal e-commerce, although if the subject of
the trade is illegal as such but is successfully conveyed through legal
transportations, we are unable to separate it from legal e-commerce.
3 Factors Driving B2C e-Commerce in Goods
Finland is technologically one of the world’s most developed countries. In 2014, its
penetration of Internet access was 93 % (Eurostat 2014), only surpassed in Europe
by the other Nordic countries, the Netherlands, and Luxembourg. The number of
people buying goods online has rapidly grown in Finland (Fig. 1). By 2014, already
74 % of Finns had made purchases online, and the number of online stores’
customers has tripled over the past 10 years. Until 2008, the growth in the number
Fig. 1 Shares of those who made orders or purchases on the Internet in Finland 2004–2014
(Statistics Finland 2014a)
of purchasers was very swift. Subsequently, the growth has waned, which is natural,
as the majority of the population already is e-commerce customers.
The decline in economic conditions has also slowed down the growth. In the past
years, ordering and purchasing on the Internet were new phenomena, and they grew
regardless of the economic trends. The most popular products purchased on the
Internet are not necessities; this means that they are susceptible to fluctuations in
economic conditions.
As stated in Sect. 2, e-commerce can be defined in a number of ways, and
therefore it is difficult to find consistent statistics. However, by combining different
sources of information, it is possible to estimate volumes of goods purchased
online. Figure 2 presents the number of shipments since 1990. e-Commerce ship-
ments in parcels are transported by all logistics operators operating in Finland. Posti
delivers letters and goods which have been purchased online. The volume shown in
Fig. 2 includes shipments transported in both parcels and letters.
A number of highly different statistics have been presented in Finland
concerning the development of the monetary value of e-commerce in goods (TNS
Gallup 2015; Statistics Finland 2014a; PostNord 2015). Figure 3 presents an
estimate of the development of e-commerce in goods in Finland in 1998–2014,
based on the number of online purchases referred to above and the average value of
one purchase (Elkelä 2014).
The value of e-commerce in 2014 was EUR 2.4 billion. The value of
e-commerce in goods has doubled since 2009. Because the number of online
purchases has simultaneously grown by a quarter, this means that the highest
growth in e-commerce in goods is related to the value of the average purchase.
As previously stated, good experiences in online shopping strengthen consumers’
Fig. 2 B2C goods e-commerce volume in Finland 1990–2014 (Virta et al. 2011; Nikali 2015;
Nikali and Hautakoski 2015)
Fig. 3 Value of B2C e-commerce in goods in 1998–2014
courage to buy increasingly expensive products. Measured in monetary value, B2C

e-commerce has grown an average of 10 % per year in 1998–2014, while from the
viewpoint of transportation the annual average growth has been only 5 %.
The number of online purchases strongly depends on the state of the economy.
The correlation between online purchases and GDP is very high (coefficient 0.97),
just as the correlation between online purchases and consumers’ purchasing power
(coefficient 0.90). In 1998–2008, the GDP grew strongly and so did the total
number of e-commerce shipments. The international financial crisis of 2008 caused

GDP to contract by 8.5 % which led the number of online purchases to decline by
15 %. The collapse of 2009 flattened out rapidly, but with the prolonged financial
crisis growth has been muted. Volumes remained fairly unchanged until 2014, but
last year growth was 10 %.
e-Commerce is important for the parcel transportation business in general. At
present, approximately 30 % of all parcels delivered in Finland arise from
e-commerce, compared to a share of 16 % in 2010. Throughout the 2010s, the
total number of parcels, including those not related to e-commerce, has remained
unchanged. Without e-commerce, however, the total number would have fallen.
Compared with the situation before the financial crisis, the number of e-commerce
parcels in 2014 was 17 % higher than in 2008, whereas the total number of parcels
is still close to 10 % lower than in 2008 (Nikali 2015).
The Internet has become an important channel for shopping across the popula-
tion and all age groups in Finland. The products most commonly purchased vary
depending on gender with women buying mainly clothes and men purchasing
entertainment electronics (Elkelä 2014). Finns prefer to buy goods from domestic
online stores, but Western European stores serving in Finnish are also popular.
Purchasing online is becoming more diversified, covering new range of products
and types of stores. As purchasing experience increases and has mainly been
positive, the average purchase has become more expensive and the popularity of
foreign online stores has grown. e-Commerce is expanding to cover new products
and ways of purchasing driven by young people aged under 30 and veterans of
online shopping, who have purchased from online stores for years (Elkelä 2014). It
is thus possible to describe e-commerce as a process where good experiences
encourage people to buy more expensive goods in a more diversified manner, and
increasingly from abroad. Recently, Chinese products have become popular among
consumers from Scandinavian countries and Russia (Elkelä 2015a).
The most common reasons for using e-commerce are the ease of use, range of
products, and opportunities to make extensive price comparisons (Statistics Finland
2014a; Elkelä 2014). Price alone is not the most important criterion for the decision
to purchase. Consumers also consider the quality of the products, so the price-
quality ratio determines which online store and product is selected (Virta and
Marttila 2009). Consumers expect free or at least very cheap fees charged for the
delivery of the goods purchased online, so transportation costs are a key factor in
the evaluation of e-commerce.
It represents a challenge for Finnish online retailers that shoppers order approxi-
mately 30 % of their online purchases from abroad, but the share of foreign sales in
Finnish online stores only represents a minute percentage (Elkelä et al. 2015). Finnish
consumers seek broader product ranges and more opportunities for price comparisons
than what they find in domestic online stores. On the other hand, Finnish online stores
have not had the same willingness to grow and become global as those in Germany and
Sweden, where online retailers consider expansion to nearby markets as a natural and
self-evident part of e-commerce (Elkelä et al. 2015). They have a systematic interna-
tionalization strategy, which includes thorough familiarization with the culture and
shopping behavior in the target market and preparation for the possibility of making
losses for a few years in new markets.
Econometric modeling of the demand for e-commerce can help explain the total
number of shipments of B2C e-commerce in goods presented in Fig. 2. When we
focus on volume instead of the value of e-commerce, it is not necessary to take into
account changes in the average purchase. It is also more reliable to make a forecast
on the basis of volume, as it is very difficult to anticipate future changes in the value
of the average purchase. The model describing the development of volume is
(Nikali 2014a):
lnðqt Þ ¼ 5:87 þ 3:71lnðT t Þ 0:53lnðSt Þ 1:40lnðK t Þ þ εt

ð1Þ
ð4:4Þ ð11:3Þ ð4:4Þ ð2:9Þ
t ¼ 1990–2014
R2 ¼ 0.98
s ¼ 0.07
DW ¼ 1.84
Where qt ¼ Number of B2C e-commerce shipments (in parcels and letters) in
Finland
Tt ¼ GDP
St ¼ Business cycle indicator
Kt ¼ Cost index for road transport of goods
εt ¼ Residual error
the t-values of coefficients are presented in brackets below the formula (1).
The model is quite simple, but reliable as to its statistical characteristics. It
explains 98 % of the changes in the volume of e-commerce. The model is presented
in Fig. 4. It does not follow the actual development very precisely only in the first
half of the 2000s; the model seems to take a short cut. But it seems to explain the
development from 2008 onwards exceptionally well.
The GDP is clearly the strongest factor explaining the volume of e-commerce.
Its coefficient is fairly high, 3.71. This means that GDP growth of approximately
3 % drives the volume of online purchases of goods to a growth of approximately
11 %. This well explains the fast growth in e-commerce in 1994–2004 (e-commerce
on average +11 % per year with GDP +3.6 % per year), the slowing of growth
afterwards, as well as the waning during the financial crisis and its aftermath
(e-commerce on average +0.4 % per year and GDP 1.0 % per year since 2008).
The business cycle variable—much faster than the GDP in its variation—reflects
the expectations of the businesses and the consumers regarding the economic
development. The negative sign of the variable is noteworthy. From the viewpoint
of the development of the national economy, it functions as a balancing factor in the
competition between e-commerce and brick-and-mortar stores: When the national
economy develops well, the cyclical indicator usually also grows, indicating that
the increase in wealth is sufficient for both e-commerce and brick-and-mortar
stores. Conversely, when the national economy develops poorly, the cyclical
indicator is usually falling, implying that the structural changes between
Fig. 4 Total volume of e-commerce in 1990–2014 and the model describing it (Nikali 2014a)
brick-and-mortar stores and e-commerce mostly benefit e-commerce due to the fact
that people especially go online to seek for cheaper goods.
The price index of the transportation charges for online purchases was not
available for the study, as there are no statistics on it. The general transportation
price index for truck transport in Finland was used as proxy for fees charged to
deliver goods (Statistics Finland 2015). It describes the development of transpor-
tation costs in general, irrespective of whether such costs are included in the price of
a given product, or charged as additional fees on top of the actual price. Therefore,
the index reflects one of the key cost factors related to e-commerce in goods. The
coefficient of the price index is surprising, at 1.40. This means that the volume of
online purchases reacts strongly to the development of transportation costs. As
transportation costs in some way or another are charged from the maker of online
purchases, the purchasers also react to this. The result obtained is supported by the
fact that consumers state in interviews that they expect free or at least very cheap
transportation charges for the delivery of their online purchases (Elkelä 2014).
The model indicates that the national economy and transportation charges are
key determinants of e-commerce in goods in Finland. There is a high elasticity of
economic activity, with 1 % increase in the economic activity will lead to a more
than 3 % increase in volumes of online purchases. The prolonged recession of the
past few years and the poorly developed consumers’ purchasing power have slowed
the growth. The increase in e-commerce has been more a result of the structural
change in trade rather than increases in total consumption. On the other hand,
compared with the total demand for consumer goods, e-commerce has grown
rapidly. While in 1998 the share of e-commerce of the total demand for consumer
goods was 7 %, by 2014 it had grown to 20 % (Statistics Finland 2014b).
Figure 5 shows forecasts for volumes of online purchases until 2020. These
projections are derived using the econometric model, if annual GDP growth and
Fig. 5 Forecast for the volume of B2C e-commerce in goods in Finland until 2020 (Ministry of
Finance in Finland 2015; Nikali 2014a)
real transportation charges remained unchanged. If an increase in transportation

costs forces online stores to collect transportation charges that increase faster than
the general inflation rate, this may decrease the forecast presented in the figure. On
the other hand, if the drop in fuel prices leads to a significant decrease in transpor-
tation costs, e-commerce may pick up and exceed the forecast.
According to the forecast, the volume of e-commerce in goods is expected to
grow in Finland by approximately 60 % in 2014–2020. This means an average
annual growth of +8 %, estimated to accelerate at the end of the forecast period. The
forecast is in harmony with consumers’ own views of the future of online shopping,
as long as e-commerce can meet consumers’ increasing expectations: an easy and
well-functioning process, inexpensive prices and high-quality products, safe pay-
ment, and good product information (Elkelä 2014). If these conditions are met,
online shopping will become more common and be diversified through good
shopping experiences to new buyers, product groups, and stores.
If e-commerce is able to provide a total price benefit compared with the
traditional brick-and-mortar stores, taking the prices of the products and the trans-
portation costs into account, the annual growth figures of e-commerce may even
exceed the forecast in Fig. 5. However, the development of e-commerce in goods in
Finland will, for a number of years, primarily be based on structural changes in
brick-and-mortar stores and e-commerce—not growth in total demand because
consumers’ purchasing power will in Finland at its best grow very slowly for the
entire remainder of the decade (Ministry of Finance in Finland 2015). Genuine
two-digit e-commerce growth figures may only be achieved when GDP growth
picks up considerably from what it is now, whereby consumer’s purchasing power
will also increase.
4 e-Commerce from the Viewpoint of Logistics Operators
The importance of e-commerce for the parcel transportation business is well

reflected by the fact that in Finland the total number of parcels in 2014 was at the
same level as in 2005, even though the number of parcels generated by e-commerce
had simultaneously grown by more than one-third (Nikali 2015). The share of
parcels based on e-commerce of all parcels has grown fast, and this development
is forecasted to continue (Nikali 2014a). At the same time, traditional paper-based
communication is undergoing a crisis (International Post Corporation 2014). In
Finland, the number of addressed letters has fallen by 35 % in 2000–2014 and is
forecast to be halved in 2014–2020. The total demand for newspapers and maga-
zines is predicted to decrease by 40 % in 2014–2020 (Nikali 2014c, d).
The biggest individual reason for the fall in letter volumes is the transfer of
consumer invoices from letters to electronic networks (Nikali 2014b). Of the paper
communication markets, unaddressed direct marketing will manage best; its vol-
ume will remain at approximately the current level (Nikali 2014e). Overall, this
means that the delivery market of traditional paper communication will fall by
nearly 40 % from the level of 2014 in Finland by 2020, which at the price level of
last year means a decrease of approximately EUR 320 million in net sales for
transport and delivery operators. The goods transport market, which includes the
transport of both goods purchased from online stores and other parcels, will,
however, grow by up to EUR 100 million.
The result is that e-commerce cannot in any way replace the loss in net sales
caused by the substantial changes in traditional paper communication (Fig. 6). The
problem becomes more complicated by the fact that competition in the traditional
paper communication market is considerably weaker than in the transport of
parcels, where each transport agreement becomes subject to fierce competition.
This is the situation not only in Finland, but everywhere in the developed countries.
Because POs have not discovered a realistic way of resolving this matter, they place
excessive expectations on e-commerce.
The total market of goods transport will not grow nearly as fast as e-commerce in
goods presented in Fig. 5, because for the present the majority of parcels stem from
fields other than e-commerce, and their growth will be slow (Nikali 2014a).
The EU is attempting to support the growth of e-commerce through regulation to
lower barriers to trade through the harmonizing consumer protection directive
2011/83/EU. With the directive, the rules concerning distance selling will be
harmonized in the Member States. There may no longer be national regulation
that sets consumer rights on a higher level than what is covered by the directive. In a
Green Paper “An integrated parcel delivery market for the growth of e-commerce in
the EU” (European Commission 2012), the European Commission stated that the
structurally fragmented parcel market and the high logistics costs, particularly in
cross-border transport, are significant barriers for the development of distance
selling within the EU. Because there are all kinds of players in the logistics market,
with different procedures and IT systems, the use of their services is too
Fig. 6 Paper communication and goods transport market sizes in Finland 1990–2020 (Nikali
2014a, b, c, d, e)
complicated. In March 2015, the Commission communicated that one of the key
tasks of the Digital Single Market Strategy will be to facilitate cross-border
e-commerce, especially for SMEs, with harmonized consumer and contract rules
and with more efficient and affordable parcel delivery (European Commission
2015).
The Green Paper particularly emphasized that postal regulation does not take
into account the characteristics related to parcel transport that are important to
distance selling customers, namely delivery time and traceability (track & trace).
Regulations concerning other (non-postal) goods transport also does not cover these
areas. Regardless, solutions taking the needs of distance selling customers into
account have also arisen in the industry. The multitude of procedures has led to a
situation where the current transport services do not sufficiently meet the needs of
distance selling. From the point of view of the retailer and consumer, it would be
important to be able to anticipate the delivery time and type, trace the shipment, and
return products easily. Due to the Green Paper, IPC Member States are preparing a
global shipping platform with full track and trace capabilities offering end-to-end
visibility to retailers and consumers alike. The attempt is to find a market-based
solution for the needs of distance selling without any rigid regulation.
Even though the fragmentation of the parcel market is considered a barrier to the
distance selling, little attention has been paid to the fact that the different legal
treatment of postal and other parcel operators leads to different service solutions,
which are difficult for online retailers and consumers to compare. The Universal
Service Obligation (USO) is strongly linked to postal operations, which secures that
users in sparsely populated areas as well have reasonably priced services available.
However, when all postal operations are not within the scope of USO and shipments
of distance selling goods transmitted through posts may be transported in parcels or
letters, shipments transported by POs are consequently regulated in a number of

different ways.
The key characteristics of transportation law are internationality and restricted
liability of the carrier for compensation. The liability of a logistics operator is
restricted by international treaties, and the provisions are very consistent in differ-
ent countries. Transportation laws consist mainly of rules relating to liability
whereas postal regulation tends to govern also operations. This sets operational
and administrational demands for POs.
Postal transportation is regulated through UPU rules applicable to international
postal deliveries and national legislation. Laws applicable to national postal oper-
ations differ considerably from each other and may include complicated structures
concerning the liability for compensation of the PO in different situations. In
addition, the liability of the operator for national postal shipments often exceeds
the maximum compensation determined by international treaties. For example, in
Finland, a PO may not forbid sending goods in a letter, and its liability for damages
or loss of a registered letter item may be as much as EUR 340, whereas liability for
a non-postal parcel is EUR 20 per kg at most. e-Parcel Group members (31 POs)
have agreed to compensate for damage or loss of a postal priority parcel up to SDR
450 compared to UPU’s general rule of SDR 40/item plus SDR 4.50/kg and SDR
8.33 per kg in international road transportation.
Liability for compensation of a transport operator directly impacts the price of
the transport service. National provisions may set the liability for compensation of a
PO at a significantly higher level than that of other operators. Because comparing
the compensation rules for different carriers is difficult, both online retailers and
consumers most often select the cheapest transport service without realizing the
related liability for compensation. Many POs also act as logistics companies, which
means that they offer parcel services outside postal regulation, whereby the variety
of regulation applicable to them makes it more difficult for them to compete. In
addition, postal regulation often sets requirements for the operator that increase
administrative and process costs compared with other transport operations.
5 Conclusions
POs in all developed countries continue to struggle with problems of the demand
for traditional paper communication. Letter communication, in particular, has been
at the core of the postal service business and suffered from substantial difficulties.
Transportation of parcels is among the few areas of the business operations that is
growing. As a result, POs expect good prospects in the parcel business. It is even
expected to outweigh the losses in net sales due to the fall in traditional
communication.
The key driver for growth in the number of parcels is e-commerce and is
expected to continue. Two-digit growth and forecast figures are often presented
for it. However, two fallacies are related to the development of e-commerce from
the viewpoints of transport operators. First of all, e-commerce in services is also

calculated to the total e-commerce, and the share of services is more than half.
Secondly, growth in e-commerce is being seen as increase in revenues from this
activity. When we separate the trade in goods from all of e-commerce and the
development of the volume of shipments needing transportation from the value of
these purchases, at least in Finland the growth has been only one-third of that of
e-commerce. The value of an average purchase has risen much more than the
number of purchases. Good experiences in e-commerce lead to the acquisition of
even more expensive goods online.
When comparing the growth in net sales expected from the transport of goods
purchased from online stores with the decrease in paper communication, it is
impossible, at least in Finland, for e-commerce to replace the losses in the next
5 years. Growth in e-commerce and the entire parcel business during the next
5 years is predicted to be up to EUR 100 million, but the decrease in traditional
communication is more than triple. In order to manage this difficult situation, the
posts should be considerably more innovative than at present and boldly expand to
quite new industries outside goods transport and message transmission.
Key factors explaining the demand for e-commerce in goods are the national
economy and the development of transportation costs. The importance of the
national economy is natural, as it reflects consumers’ purchasing power and the
development of trade generally. e-Commerce in goods is often advertised by stating
that no separate delivery fees are charged from the customer. In reality, however,
transportation costs strongly explain the development of e-commerce in goods.
Although the transportation costs are not necessarily observable to the customer as
separate fees or explicit expenses, they are included in the price of the product in
one form or another. Consequently, such embedded costs have a latent impact on
the development of e-commerce, as there is constant pressure to outprice traditional
brick-and-mortar stores, even when the cost of transport is factored in.
The EU attempts to promote the development of e-commerce in many ways.
However, regulation related to goods transport is still very varied and based on
different national legislation. This concerns, in particular, cross-border goods
transport, which makes the rapid internationalization of e-commerce more difficult.
From the point of view of the posts, the problems are increased by the fact that they
transport e-commerce goods in both parcels and letters, and these services are
regulated in very different ways. In addition, part of the parcel traffic of POs is
subject to USO, which has its own regulation. This variety has led to a situation
where the services of operators managing goods transport are very different,
whereby it is difficult for consumers and online retailers alike to compare them
and understand their content. As necessary, a PO might have USO in areas where
the market does not offer sufficient alternatives for the shipment of goods. But for
POs to be able to compete equally with other logistics operators, both international
and national regulation has to be harmonized to apply to all operators. Even though
the growth in e-commerce will not alone be sufficient to save posts from economic
losses caused by the decrease in traditional paper communication, it will play a key
role for the business offerings of POs.
References
Elkelä K (2014) Verkko-ostamisen vetovoima [Attraction of online buying]. Itella Corporation,

Research series 25/2014
Elkelä K (2015a) Russians’ shopping in foreign web stores of goods. Posti BI, Research series
1/2015
Elkelä K (2015b) Tavaraverkkokauppa Suomessa [E-commerce of goods in Finland]. Posti BI,
White Paper 26 Feb 2015
Elkelä K, Huuhtanen V, Nikali H, T€ orr€onen P, Hemmi H (2015) Future and challenges for
e-commerce of goods. Posti BI, Research series 5/2015
European Commission (2012) Green Paper, an integrated parcel delivery market for the growth of
e-commerce in the EU, 29 Nov 2012, COM (2012) 698 final (“Green Paper”)
European Commission (2015) Communication from the Commission, Digital single market
strategy. European Commission http://appsso.eurostat.ec.europa.eu/nui/show.do agrees areas
for action, 25 March 2015, http://europa.eu/rapid/press-release_IP-15-4653_en.htm
Eurostat (2014) Individuals—Internet use. http://appsso.eurostat.ec.europa.eu/nui/show.do
TNS Gallup (2015) E-commerce statistics, quarterly
International Post Corporation (2014) Global postal industry report
Ministry of Finance in Finland (2015) Economic survey. Ministry of Finance publications
17b/2015
Nikali H (2014a) Parcel market in Finland. Demand models for parcel services and e-commerce.
Itella Corporation, Research series 6/2014
Nikali H (2014b) Character of substitution and its significance for letter demand, case Finland. In:
Finger M, Bukovc B, Burhan M (eds) Postal services in the digital age. Ios Press, Amsterdam,
pp 15–29
Nikali H (2014c) Sanomalehtien kysyntä Suomessa [Demand for newspapers in Finland]. Itella
Corporation, Research series 7/2014
Nikali H (2014d) Aikakauslehtien kysyntä Suomessa [Demand for magazines in Finland]. Itella
Corporation, Research series 22/2014
Nikali H (2014e) Suoramainonnan kysyntä Suomessa [Demand for unaddressed advertising mail
in Finland]. Itella Corporation, Research series 33/2014
Nikali H (2015) Posti’s CEP market share, total market 1990–2014 and forecast 2020. Posti BI,
quarterly
Nikali H, Hautakoski K (2015) Content survey for Letter Communication. Posti BI, Research
series 3/2015
PostNord (2015) E-handeln i Norden [E-commerce in the Nordics]
Statistics Finland (2014a) Väest€
on tieto- ja viestintätekniikan käytt€
o-tutkimus 2014 [Use of ICT in
Finland 2014]
Statistics Finland (2014b) Kansantalouden tilinpito, kotitalouksien käytettävissä olevat tulot ja
kulutus [National accounts]
Statistics Finland (2015) Cost index for Road Transport of Goods, monthly
Virta A-M, Marttila J (2009) Kuluttaja ja verkko-ostaminen [Consumers and online buying]. Itella
BI, Research series 14/2009
Virta A-M, Nikali H, Elkelä K, Marttila J, Koivisto P (2011) Kuljetuspalvelujen asiakastarpeet
organisaatioissa ja kotitalouksissa 2011 [Customer’s needs of transport services in organiza-
tions and households 2011]. Itella BI, Research series 12/2011
Personal Data and Privacy Issues
and Postal Operators Stand
Claire Borsenberger, Denis Joram, Olaf Klargaard, and Philippe Regnard
1 Introduction
Data has become the raw material of production, a new source of considerable
economic and social value. Advances in data mining and analytics, the massive
increase in computing power and data storage capacity coupled with decreasing
cost (falling by a factor of 6 since 2005 according to Podesta et al. 2014) and the
increasing number of people, devices, and sensors that are now connected by digital
networks and able to communicate with each other (“Internet of Things”) have
revolutionized the ability to generate, communicate, share, and access data.
According to the World Economic Forum (2011), 15 billion devices will be
connected to the Internet by 2015 and 50 billion by 2020. The amount of data
stored on the Internet is predicted to grow exponentially and looks set to be 44 times
larger in 2020 than it was in 2009 (World Economic Forum (2011), figure 3, pp. 14).
The opinions expressed here do not necessarily reflect the position of La Poste. We thank Yassine
Lefouili, Professor of Economics at Toulouse School of Economics; Michael Crew, Director and
CRRI Distinguished Professor of Regulatory Economics at Rutgers University; Tim Brennan,
Professor of Public Policy and Economics, University of Maryland, Baltimore County (UMBC);
and discussants at the Rutgers conference, Margaret Cigno and Edward Pearsall, for their helpful
comments and suggestions.
C. Borsenberger (*)
Department of Doctrine et Modélisation, Direction of Regulation and Institutional and
European Affairs, Groupe La Poste, Paris, France; Laboratoire d’Economie d’Orléans
(LEO); and University François Rabelais, Tours, France
e-mail: claire.borsenberger@laposte.fr
D. Joram
Department of Regulation and Studies, Direction of Regulation and Institutional and European
Affairs, Groupe La Poste, Paris, France
O. Klargaard • P. Regnard
Direction of Digital Affairs, Groupe La Poste, Paris, France

This ocean of data and shared information should increase welfare, lower search
costs, boost economic productivity, reduce economic inefficiencies, and improve
our own experiences, thanks to data analysis and the development of predictive
algorithms. It could also be a source of losses, economic inequalities, and power
imbalances between those who generate (more or less consciously) the data and
those who control and exploit these data. The “big data” new world exacerbates
privacy concerns and could threaten the right “to be let alone” (Brandeis and
Warren 1890). This threat on cherished liberties could stir a regulatory backlash,
dampening the data economy and stifling innovation. Indeed, the tasks of ensuring
data security and protecting privacy become harder as information is multiplied and
shared ever more widely around the world.
This raises the issue of determining the right balance between the additional
societal values generated by information disclosure and uses of data on the one
hand, such as public health, national security and law enforcement, environmental
protection, and economic efficiency, and the potential risks to individual autonomy
and privacy (discrimination, exclusion, loss of control on data disclosure) on the
other hand. We will focus on solutions that give consumers (the data subjects) a
better control on their personal information.1 These solutions require further trans-
parency in data collection practices from businesses, on their privacy and security
policies. It also requires allocating property rights over personal information to
consumers and better education on the risks and benefits of data disclosure. To help
consumers, intermediaries could emerge (certified third parties). Section 2 sets the
scene and summarizes lessons that could be drawn from the economic literature on
privacy. Section 3 discusses the possible role by National Postal Operators (POs).
Section 4 concludes.
2 Should Personal Data Be Protected? Learnings from

Economic Theory on Privacy
Data protection and disclosure generate both positive and negative externalities
(Shapiro and Varian 1997) for those who create these data (the “data subjects”—
individuals, citizens, consumers, and so on) and those, who exploit them (the “data
holders”—firms, public and private organizations, government, and so on).
Firms can benefit significantly from the ability to learn detailed information
about their current or potential customers. By observing individual behavior, they
can identify individuals’ preferences, improve their customer relationships, and at
the end of the day, increase their revenues (see for example, Varian 1985; Linden
1
Other solutions could be self-regulatory measures implemented by firms to deal with their
customers’ data (assuming that intense media scrutiny on personal data misuse creates a financial
pressure on the business community roughly analogous to Pigouvian taxes) or a proactive policy
enforced by public authorities.
Personal Data and Privacy Issues and Postal Operators Stand 263
et al. 2003; Acquisti and Varian 2005; Ball et al. 2006; Bennett and Lanning 2007;
Richards and Jones 2008). From the point of view of data subjects (consumers),
disclosure of personal data could also be beneficial, since they can benefit from
immediate monetary compensation for revealing her personal data (Laudon 1996)
and from improved services, customized or targeted offers, or even new products
(Blattberg and Deighton 1991). On the other hand, excessive use of personal
information may harm consumers through improper use of personal information,
credit card fraud, or identity theft (Culnan 1993; Smith et al. 1996). These privacy
concerns hurt firms directly by forcing them to invest in data protection (Hoofnagle
2007; Acemoglu et al. 2013). Self-interest may also provide powerful motivation
since consumers’ propensity to purchase may negatively be affected after a mer-
chant has violated a consumer’s data. So, what is the right level of data protection?
2.1 Lessons from Economic Theory on the Need of Data

Protection
According to Chicago school scholars, rational consumers make a “privacy calculus”

and assign an economic value of data disclosure based on a cost–benefit calculation
(Bennett 1995; Ackerman 2004). This explains why, as noted by Varian (1996), a
consumer may naturally be interested in disclosing certain personal traits to some
firms and at the same time in keeping other types of information private. In this
context, according to Posner (1978, 1981) and Stigler (1980), the protection of privacy
creates inefficiencies in the marketplace by concealing potentially relevant informa-
tion from other economic agents, or at best remains ineffective, since individuals who
decide to protect their personal information are de facto signaling a negative trait.
These views rely on the strong assumption that people are rational and fully
informed about the way their data are exploited. This is not always the case (Tene
and Polonetsky 2012). Behavioral economics studies have evidenced the “privacy
paradox”, the fact that despite self-reported privacy concerns, some consumers still
share their personal information (Acquisti and Grossklags 2005) and few are
willing to pay to protect their personal data (Hann et al. 2007; Rose 2005; Varian
et al. 2005; Png 2007; Tsai et al. 2011).
The existence of asymmetries of information and uncertainties hamper individ-
uals’ to act “rationally” when facing privacy trade-offs making the assessment of
costs and benefits of information disclosure particularly complex. According to
Pavlou et al. (2007), individuals do not know ex ante which entities have appropriate
information protection practices (adverse selection) and ex post whether their per-
sonal information will be appropriately used or not (moral hazard). As illustrated by
Acquisti (2010), an intuitive way of describing the state of uncertainty associated
with privacy costs is the “blank check” metaphor: as an individual reveals private
information to other parties, she is signing a blank check that may never come back to
her, or may come back for an indeterminably small or large price to pay.
According to Taylor (2004), the welfare enhancing or diminishing nature of the

presence of privacy regulatory protection depends on consumers’ level of sophis-
tication. If consumers are naı̈ve, they do not anticipate the seller’s ability to use past
consumer information for price discrimination; therefore, in equilibrium all their
surplus is taken away by the firms, unless privacy protection is enforced through
regulation. In this context, some privacy regulation could be welfare improving.
Shy and Stenbacka (2014) for example showed that from the consumers’ point of
view, an increase in the degree of privacy protection is always beneficial, limiting
the ability of firms to extract consumers’ surplus with more precise discriminatory
pricing scheme. But even from the firms’ point of view, a small and positive degree
of privacy protection is a good thing: it relaxes the competitive intensity, which is
maximal under perfect information and firms make higher profits than under no
privacy protection.2
2.2 Lessons on the Ways to Protect Personal Data
In sum, economic theory seems to plead for a better control given to data subjects
on their personal information. In particular, academics, privacy groups and busi-
nesses gathered in a project entitled “Rethinking personal data” and hosted by the
World Economic Forum, call for the establishment of a “balanced personal data
eco-system,” based on “user-centricity,” which “integrates diverse types of per-
sonal data while putting end users at the center of data collection and use” (World
Economic Forum 2012, 2014).
Assigning ownership or property rights on personal data could be considered as
the first step toward data subjects’ empowerment, but clarifying the property rights
over personal data is not enough to solve the privacy issues and to lead to an
efficient Coasian bargaining outcome for two reasons.
Firstly, as already said, Internet is characterized by the existence of huge
asymmetries of information. These asymmetries result in high monitoring costs
faced by consumers when it comes to control data circulating on Internet and deal
with privacy issues and in difficulties (and costs) to identify potential breaches and
to effectively enforce any agreement. According to Martin (2013), “for the online
user, ex ante costs [are] the time and cost required to identify all Web sites and
tracking parties involved, their approaches to handling information, and the benefit
(if at all) of handing over information. In other words, the time and money
necessary to identify relevant parties and privacy options online, to incorporate
2
In fact, the relationship between the degree of privacy protection and equilibrium profits is not
monotonic: weak privacy protection generates higher profits than no privacy protection but strong
privacy protection generates lower profits than weak or no privacy protection (having access to
customer-specific knowledge on individual preferences makes it possible for each firm to differ-
entiate its price among its own customers which promotes the firm’s profits compared with the
prices the firm sets under strong privacy protection).
possible contingencies in an online environment, and to write findings into an

explicit contract increases the costs of bargaining and decisions online. Ex post
costs occur after the signing of the contract or after users give consent by engaging
with the privacy policy online. This includes the ability to (a) identify a problem;
(b) enforce and penalize the offending party; and (c) switch to an alternative
provider if the explicit contract is broken” (pp. 3). Even when firms communicate
to consumers the way they use personal data through privacy notices, consumers
incur high transaction costs to read and understand long written contracts.3
In consequence, the so-called Coase theorem (Stigler 1966) that relies on the
assumption of low transaction costs associated with bargaining does not apply.
Coase himself, in his 1960 paper acknowledged that real-world transaction costs are
rarely low enough to allow for efficient bargaining. Secondly, many economists
have demonstrated the importance of the perfect information assumption in this
theorem and shown that inefficient outcomes are to be expected when this assump-
tion is not met—as in the personal data market.
Taking account of these limitations of a “simple” property scheme, Schwartz
(2003) proposes a “model for propertization of personal data that will fully safe-
guard information privacy” (pp. 2058). This model suggests to associate property
rights with limitations on an individual’s right to alienate personal information,
namely a “restriction on the use of personal data combined with a limitation on their
transferability” (pp. 2095). In other words, it would permit the transfer for an initial
category of use of personal data, but only if the customer is granted an opportunity
to block further transfer or use by unaffiliated entities. Any further use or transfer
would require the customer to opt in—that is, it would be prohibited unless the
customer affirmatively agrees to it. This element would reduce information asym-
metry problems by forcing firms to disclose hidden information about their data
processing practices. The personal data-specific property rights should also include,
according to Schwartz, an expanded form of consent materialized by a right of exit
for participants in the market.
Moreover, to reduce asymmetries of information over the data disclosed on
Internet and to overcome the difficulty in specifying contract terms and in moni-
toring contract compliance, as said by Prins (2004), “we need (. . .) instruments to
enhance the visibility of and to increase our knowledge about how personal data are
used and combined, on the basis of what data individuals are typified, by whom and
for what purposes. (. . .) Privacy protection in our present-day society presumes that
we have the capability to know about typifying people and to control this process.
It requires the availability of instruments that enable awareness of the context in
which personal data are used and to monitor the data impression that individuals are
exhibiting to others.” (pp. 6–7).
3
Nissenbaum (2011) speaks about the “transparency paradox”: the more that information is shared
through notice statements, the less understandable those notice statements are for authentic
consent by the user.
The existence of transaction costs related to the control of data circulating on

Internet pleads also for the emergence of new intermediaries which manage
customer-specific data contracts. Do Posts have a role to play on this market?
3 Consumer Empowerment: Do POs Have a Role to Play?
Over the past, POs have provided a secure, universally accessible platform for
physical commerce and communications. Today, the opportunity exists to extend
their trusted intermediary role into the digital age and act as a “bridge” to facilitate
the advancement of access to the digital world. In fact, POs benefit from several
assets to invest the market of data protection tools and be a major actor to give
consumers empowerment on their personal data.
Firstly, protection of personal information has been in the “DNA” of POs for
centuries. Since always, POs have been trusted intermediaries providing secured
communications services. The confidentiality of correspondence is a fundamental
legal principle enshrined in the constitution of many democratic countries.
It guarantees that the content of sealed letters is never revealed and letters in transit
are not opened by any third party. It is thus the main legal basis of the privacy of
correspondence right. Consequently, respect of confidentiality, data protection, and
privacy are logically associated to postal brands.
Secondly, the public has generally a high degree of trust in the national PO. POs
are generally considered as safe, trusted, and reliable institutions. Often, they
provide services of general economic interest (SGEI) on behalf of the state; some
are still public administrations or state-owned companies, increasing the feeling of
trust and security of these institutions.
Thirdly, many POs benefit from a large physical network of outlets which could
become a bridge between the physical and digital worlds. Post offices could become
the place where low digital skilled people learn to use, search, and communicate
with digital tools. As an example, the French PO has launched an initiative in some
postal outlets, providing free access to Internet and educational support to the
population on digital tools linked to data, document, and identity protection. This
is in line with the idea that access to the Internet is not an end in itself but a first step
that needs to be accompanied with educational support, so that citizens and
consumers are empowered.
Many POs around the world have already investing in the digital economy. They
are developing relevant and promising digital services and solutions for consumers
like hybrid mail, e-letters, e-registered mail, digital mailboxes, online payment
solutions, and authentication and secured archiving services. In more recent
years, POs have started to position their services as Personal Data Store (PDS).
PDSs provide both a secure tool both in terms of conservation of personal data,
but also in access to the data (identity management with secure authentication
services) as well as a new class of user-driven services so that consumers can give a
controlled access to their data, and contract with third parties. As Rubinstein (2012)
noted, with PDSs, individuals manage themselves collection and use of their
personal data. They can share some of their data selectively, without disclosing
more personal data than they wish to and control the purpose and the duration of
primary and secondary use of their data (through contractual and technical means).
In other words, individuals are able to “tag” every unit of their personal data with
metadata describing privacy-related requirements and preferences. Associated with
a minimum regulatory background to ensure a fair and protective use of personal
data by individuals, PDSs constitute a tool that allows consumers, companies, and
the economy as a whole to benefit from the value of data.
As an empowerment tool for consumers, PDSs allow rational consumers to
control value and share their data as they would manage any good or service on a
market. PDSs constitute a marketplace where consumers and producers can meet to
realize transactions on personal data. It would provide consumers the opportunity to
measure and materialize their “taste for privacy” and assign an economic value of
data disclosure based on a cost–benefit calculation. It would allow producers to
collect relevant information on consumers’ behaviors, needs, and intents. With all
economic agents benefiting from better information, inefficiencies on the market
would be reduced and economic welfare improved. For example, PDSs, in provid-
ing tools for the consumer to control and manage her personal data allow her to
react to well-informed firms engaging into price discrimination. Without any
management tool for her personal data, consumer would have taken her entire
surplus by well-informed firms, with a negative impact on total welfare. With PDS,
consumers are aware of how firms will exploit their data, and able to adapt their
behavior accordingly, generating equilibrium more balanced situation between
consumer surplus and producer profits.
PDSs functionalities are also in line with privacy protection rules, as defined by
Schwartz as part of property rights on personal data. PDSs materialize restrictions
on use and transferability of data, through online contracts and consent, and allow
users to opt in or out of various services at any time. By reducing asymmetries of
information and uncertainties for both sides (consumers and firms), PDSs not only
protect individuals from unwanted use of their personal data but internalize positive
and negative externalities associated with personal data, ensuring efficient alloca-
tion of resources and improving total welfare.
PDSs have been provided experimentally, and even made available to the market
by various public and private stakeholders. MIT researchers have developed a
prototype system, named OpenPDS, which stores data from different digital
devices in a single location specified by the end user (cloud, personal computer,
external storage). Any cell phone application or online service that wants to use the
user’s data has to query the data store, which returns only as much information as is
required (De Montjoye et al. 2014). The key benefit of OpenPDS is that it requires
external applications to specify what information they need and how it will be used.
OpenPDS is currently being tested with several telecommunications companies in
Italy and Denmark, with a storage option limited to the cloud at this stage (http://
openpds.media.mit.edu).
Another example is given by Personal.com, a start-up company that enables

individuals to own, control access to, and benefit from their personal information. It
does so by providing individuals with an online “data vault” divided into compart-
ments called “gems,” where they can store and share information about their
shopping habits, travel, log-in credentials on various sites, location information,
and more. There are currently more than 100 gems with more than 3000 fields of
data. The food preferences gem, for example, includes allergies, religious and
dietary restrictions, and whether a user likes spicy food. Users can share gems
with family, friends, employees or colleagues, and more importantly, monetize
their own data by selling access to gems to commercial entities. (Personal.com
collects a 10 % fee on such sales.)
In the postal sector, services like “My Post” (Australia Post), “Connect”
(NZ Post), “Digipost” (Norway Post), “E-boks” (Post Denmark), or “Digiposte”
(French Post) are now largely used by citizens and consumers. Top of the list is
obviously E-boks with four million users thanks to a supportive regulation (man-
datory for citizens and companies for communications with public authorities), but
also in Norway with 300,000 users and in France, where Digiposte gathers 1.5
million users. Initially conceived as digital mailboxes, simple counterparts of
physical mailboxes, these services tend to be transformed in user-driven tools to
collect and share personal data.
When compared to Rubinstein (2012) definition, these postal solutions clearly
constitute the first steps toward PDSs with key features already implemented:
secured collection and conservation of data and documents, under the control of
individuals, based on universality (not tied specifically with postal services).
Dynamic use of personal data is also a key asset of these services with the
possibility to disclose data by giving a controlled access to third parties for a chosen
duration.
However, some features are missing for these services to be fully considered as
PDSs: mostly focused on administrative use and file sharing, tools for monetized
transactions are still lacking. At this stage, contractual and technical features do not
allow users to sell and to control the purpose and duration of primary and secondary
use of their data. However, evolutions of functionalities on these postal platforms,
as well the positioning of POs on privacy and consumer empowerment issues, let us
think that full PDSs with VRM tools may soon be launched.
Some POs have even anticipated big data issues and taken into account the rapid
development of connected objects: the French Post took the opportunity of the 2015
Consumer Electronic Show (CES) to announce and launch a “platform of
connected objects.” This new service is aimed for both companies, which can use
the platform to develop services around their connected objects (data storage and
analysis, billing, link with physical services provided by postmen at the door, etc.)
and for consumers, which can retrieve and control the data from their personal
connected objects. In line with its digital strategy, La Poste is providing a service
that encourages a dynamic use of data (crossing and analyzing data from connected
objects to bring value) while protecting privacy.
4 Conclusion
This chapter shows how the development of big data, and the explosion of personal
data disclosure through connected objects, is threatening privacy as never before.
Analyzing the economic impact of personal data for consumers, firms, and the
economy as a whole, it is clear that a general ban on the sharing of personal data
would work for the detriment of all parties. On the other hand, allowing consumers
and firms to contract freely on data would clearly harm many consumers, consid-
ering notably asymmetry of information between market players and the transaction
costs incurred, and have an overall negative impact on total welfare.
Allowing consumers’ empowerment through the creation of specifically
designed property rights would both accelerate the (monetized) disclosure of data
and protection of consumers through limited transferability of data, right to exit a
contract and informed consent mechanisms. Consumers need information and tools,
in order to be fully “empowered.” This is clearly the place where POs have a role to
play, and some pioneers in France, Denmark, Australia, or New Zealand have
started to provide solutions, e.g., PDSs allowing consumers to centralize and
share their data securely. With a historical role based on the anonymous transport
of personal data, POs can reinvent themselves in the world of Big Data.
References
Acemoglu D, Malekian A, Ozdaglar A (2013) Network security and contagion. Working Paper,
MIT
Ackerman M (2004) Privacy in pervasive environments: next generation labeling protocols. Pers
Ubiquit Comput 8(6):430–439
Acquisti A (2010) The economics of personal data and the economics of privacy. Background
Paper no. 3, Joint WPISP-WPIE Roundtable “The Economics of Personal Data and Privacy:
30 Years after the OECD Privacy Guidelines” (December)
Acquisti A, Grossklags J (2005) Privacy and rationality in individual decision making. IEEE Secur
Priv 3(1):24–30
Acquisti A, Varian HR (2005) Conditioning prices on purchase history. Market Sci 24(3):367–381
Ball D, Coelho P, Vilares M (2006) Service personalization and loyalty. J Serv Mark 20
(6):391–403
Bennett CJ (1995) The political economy of privacy: a review of the literature. Center for Social
and Legal Research, Hackensack
Bennett J, Lanning S (2007) The Netflix prize. In: Proceedings of KDD Cup and Workshop
Blattberg RC, Deighton J (1991) Interactive marketing: exploiting the age of addressability. Sloan
Manage Rev 33(1):5–14
Brandeis DL, Warren SD (1890) The right to privacy. Harv Law Rev 4(5):193–220
Coase RH (1960) The problem of social cost. J Law Econ 3:1–44
Culnan MJ (1993) How did they get my name? An exploratory investigation of consumer attitudes
toward secondary information use. MIS Q 17(3):341–363
De Montjoye Y-A, Shmueli E, Wang SS, Pentland AS (2014) OpenPDS: protecting the privacy of
metadata through SafeAnswers. PLoS One 9(7), e98790
Hann I-H, Hui K-L, Lee TS, Png IP (2007) Overcoming online information privacy concerns: an
information processing theory approach. J Manag Inf Syst 42(2):13–42
Hoofnagle CJ (2007) Security breach notification laws: views from Chief Security Officers.
Technical report, University of California, Berkeley
Laudon KC (1996) Markets and privacy. Commun ACM 39(9):92–104
Linden G, Smith B, York J (2003) Amazon.com recommendations: Item-to-item collaborative
filtering. IEEE Internet Comput 7(1):76–80
Martin K (2013) Transaction costs, privacy, and trust: the laudable goals and ultimate failure of
notice and choice to respect privacy online. First Monday 18(12):2. http://firstmonday.org/ojs/
index.php/fm/rt/printerFriendly/4838/3802
Nissenbaum H (2011) A contextual approach to privacy online. Daedalus 140(4):32–48. http://
www.amacad.org/publications/daedalus/11_fall_nissenbaum.pdf
Pavlou PA, Liang H, Xue Y (2007) Understanding and mitigating uncertainty in online exchange
relationships: a principal–agent perspective. MIS Q 31(1):105–136
Png I (2007) On the value of privacy from telemarketing: evidence from the ’Do Not Call’ registry
Podesta J, Pritzker P, Moniz EJ, Holdren J, Zients J (2014) Big data: seizing opportunities,
preserving values. Report of the Executive Office of the President, The White House
Posner RA (1978) An economic theory of privacy. Regulation (May/June)
Posner RA (1981) The economics of privacy. Am Econ Rev 71(2):405–409
Prins JEJ (2004) The propertization of personal data and identities. Electron J Comparative Law 8
(3). http://www.ejcl.org/
Richards K, Jones E (2008) Customer relationship management: finding value drivers. Ind Mark
Manag 37(2):120–130
Rose E (2005) Data users versus data subjects: are consumers willing to pay for property rights to
personal information? In: Proceedings of the 38th Hawaii International Conference on System
Sciences
Rubinstein Ira S (2012) Big Data: The End of Privacy or a New Beginning. Public Law and Legal
Theory Research Paper Series, Working Paper no 12–56 (October)
Schwartz PM (2003) Property, privacy and personal data. Harv Law Rev 117:2056–2128
Carl S, Varian HR (1997) US government information policy. University of California, Berkeley
Shy Oz, Rune Stenbacka (2014) Customer privacy and competition. Working Paper
Smith HJ, Milberg JS, Burke JS (1996) Information privacy: measuring individuals’ concerns
about organizational practices. MIS Q 20(2):167–196
Stigler GJ (1966) The theory of price, 3rd edn. Macmillan, New York
Stigler GJ (1980) An introduction to privacy in economics and politics. J Leg Stud 9(4):623–644
Taylor CR (2004) Consumer privacy and the market for customer information. RAND J Econ 35
(4):631–651
Tene O, Polonetsky J (2012) Privacy in the age of big data: a time for big decisions. Stan Law Rev
Online 64:63–69
Tsai JY, Edelman S, Cramer L, Acquisti A (2011) The effect of online privacy information on
purchasing behavior: an experimental study. Inf Syst Res 22(2):254–268
Varian HR (1985) Price discrimination and social welfare. Am Econ Rev 75(4):870–875
Varian HR (1996) Economic aspects of personal privacy. In: Privacy and self-regulation in the
information age. National Telecommunications and Information Administration, Washington,
DC
Varian HR, Wallenberg F, Woroch G (2005) The demographics of the do-not-call list. IEEE Secur
Priv 3(1):34–39
World Economic Forum (2011) Personal data: the emergence of a new asset class. WEF, Geneva
World Economic Forum (2012) Rethinking personal data: strengthening trust. WEF, Geneva
World Economic Forum (2014) Rethinking personal data: trust and context in user-centered data
ecosystems. WEF, Geneva
Using Market Research to Analyze
Consumer Preferences
Stefano Gori, Bianca M. Martinelli, and Gennaro Scarfiglieri
1 Introduction
Analyzing the welfare effects of pricing practices is not simple in the postal sector,
where the sender pays all and the preferences of the recipients are not incorporated
directly in the price.1 The paradox that has always puzzled postal economists is that
the highest cost component of the service (delivery) is supplied free of charge
(to the recipient), and such has been the case since the 1837 Rowland Hill reform.
This paradox of postal delivery, termed by Panzar (1991) “the most unpriced
service,” has stimulated proposals for a delivery charge to the addressee (Owen
and Willig 1983; Schwarz-Schilling 2001; Felisberto et al. 2006).
The objective of this chapter is to understand the preferences of recipients
concerning the frequency of delivery. To do so, at the end of 2014 we carried out
a market survey on the perceived and actual frequency of delivery in Italy. We also
Any views or opinions presented in this chapter are solely those of the authors and do not
necessarily represent those of Poste Italiane or CRA or any of the affiliated institutions. We
would like to thank the discussant Frank Rodriguez, Associate, Oxera, and the Chairman of the
session Michael Crew, CRRI Rutgers University for their constructive comments.
1
Intuitively, business senders take into account the value of their message to the recipient in
deciding whether and when to send a message. For example, advertisers mail when a special
campaign needs to reach the recipient on Friday or Saturday morning for the weekend shopping.
However, this does not generally apply to all recipient of the campaign and the campaign turns out
only for a minority of recipient. If the recipient does not cover the cost directly or pay the sender
indirectly, then probably the recipient is neutral or negative to a campaign (environmental
concerns).
S. Gori (*)
CESPI, BEA and Poste Italiane, Rome, Italy
e-mail: Goriste2@posteitaliane.it
B.M. Martinelli • G. Scarfiglieri
Poste Italiane, Rome, Italy

272 S. Gori et al.
examined the impact on consumers’ preferences of a possible reduction of quality

levels (i.e., frequency). Our literature review in the postal sector indicates that
extensive research has not been carried out on consumer preferences linked to the
frequency of delivery.
In Sect. 2, we will explain why our approach is different from previous research.
Section 3 contains the description of the method used in this market research and
presents our findings. Section 4 discusses the impact of the results and a possible
way forward for future research on welfare economics applied to the postal sector.
2 Our Approach
In the past few years, several research projects have been carried out on consumer’s
preferences on delivery frequency. Some have covered multiple countries (Copen-
hagen Economics 2010; RAND Europe 2011) while others focused on individual
countries, mainly the UK (Accent 2008; TNS 2010a, b) and in France (Cecchet
et al. 2011). Some of them used SPDCEs (stated preference discrete choice
experiments) to quantify consumers’ willingness to pay for different levels of
frequency of home delivery. For example, the RAND Europe (2011) study devel-
oped a method for measuring consumers’ preferences using information collected
from consumers regarding choices among hypothetical postal services at different
prices to quantify consumers’ willingness to pay for specific aspects of postal
services. Others used (used choice-based conjoint analysis TNS 2010a, b) or
Contingent Valuation (CV) (Cecchet et al. 2011). With CV, consumers are asked
their willingness to pay for improvement of a service or their willingness to receive
compensation for a deteriorated service (RAND Europe 2011, p. 3). These studies
focused mainly on the reduction in days of delivery from 6 to 5 and on Saturday
delivery.
We used a different approach because the CV method may present cognitive
problems to respondents (RAND Europe 2011, pp. 3–5), and SPDCEs provide
values for attributes which are merit or public goods (RAND Europe 2011,
pp. xix–xx).2 In addition, these methods presented a high risk of irrelevant
responses because in 2014 Italian families spent on average an estimated 2.05
euros per month on postal and courier services (70 % less than in 1998, versus
more than 57 euros per month on telecommunications services) and on average
each citizen receives 1.7 letters per week. In countries like Italy with low volumes
2
Choice-based conjoint analysis also has disadvantages “. . .. . .due to the fact that having respon-
dents make choices is an inefficient way to elicit preferences. Each concept is typically described
on all the attributes being considered in the study, and each choice set contains several concepts.
Therefore, the respondent has to process a lot of information before giving a single answer for each
choice set. Although this does mimic what happens in the marketplace, the analyst ends up with far
less information than would have been available had the task been to rate each alternative in the
set.” (Sawtooth 2013, p. 3)
Using Market Research to Analyze Consumer Preferences 273
per capita, data may not be sufficient to estimate the willingness to pay or to be
subsidized for different delivery frequency.
Furthermore, the responses we obtained justified changing our focus from the
willingness to pay for higher frequency to addressing the consumer attitude toward
lower frequency per week. To do so, we carried out a consumer survey in Italy (to
attempt to assess if there are differences between perceived current frequency of
delivery and how they would regard reductions in delivery frequency). Finally, we
attempted to set up a method to identify the optimal service level by maximizing
total welfare, taking into account consumer preferences and the cost of the delivery
network.
3 The Method Used and Findings from the Market Survey
The research was carried out by CRA srl (Customized Research and Analysis)
using 1000 Computer Assisted Telephone Interviews during 2 months September–
October 2014 interviewing randomly selected adults (at least 18 years old) from the
Italian population.3 The four relevant questions were:
1. What is your perception of the current delivery quality levels (days of delivery)
in your area (from 1 to 5 days)?
2. How many times per week do you receive mail (from 1 to 5 days)?
3. How would you perceive a reduction of the current quality of service from
5 days per week to 2/3 days per week (four possible responses, completely/very
dissatisfied, dissatisfied, satisfied, very satisfied).
4. Of those responding satisfied or very satisfied, asking the main reason for their
response.
The first question, “What is your perception of the current delivery quality levels
(days of delivery) in your area (from 1 to 5 days)?,” is useful to understand how
consumers in the sample perceive the frequency of delivery independently from the
mail they receive. The perception of the frequency is generally linked to perception
of how many times per week the postman passes by in the neighborhood even if
mail is not delivered at a particular address. From Table 1 it emerges that the
perception of the current frequency of delivery is, not surprisingly, higher for those
living in metropolitan areas than for those in smaller towns. More than 75 % of the
population perceives that delivery frequency is at least 4 days per week; the average
frequency is for all the subgroups at least 4 days per week.4
3
The sample was created using a casual sample from the phone directory taking into account the
gender, age, geographical distribution of the population and the type of city of residence (e.g.,
metropolitan, small city).
4
Results are statistically significant, 95 % confidence interval.
274
Table 1 What is your perception of the current delivery quality levels (days of delivery) in your area (from 1 to 5 days)?
Perceived 12 Outskirts of
frequency Total Metropolitan metropolitan Towns <2000 Towns 2001< and Cities 10,001< and Cities >50,001
(per week) Pop. areas areas inhabitants <10,000 inhabitants <50,000 inhabitants inhabitants
Sample 1000 148 116 58 239 268 171
5 days 71.4 % 78.5 % 63.2 % 61.6 % 77.3 % 71.5 % 65.9 %
4 days 4.2 % 4.1 % 8.3 % 7.3 % 3.8 % 3.2 % 2.7 %
3 days 10.3 % 5.3 % 8.1 % 22.0 % 8.0 % 12.8 % 11.6 %
2 days 6.3 % 5.9 % 10.4 % 5.1 % 5.6 % 5.2 % 6.9 %
1 day 2.5 % 2.8 % 5.3 % 2.2 % 2.2 % 1.0 % 3.2 %
Less than one 5.2 % 3.4 % 4.7 % 1.8 % 3.1 % 6.2 % 9.8 %
delivery
Average 4.2 4.4 4.0 4.2 4.4 4.2 4.0
frequency
S. Gori et al.
Table 2 How many times per week do you receive mail (from 1 to 5 days)? (By age group)
Age group
Real frequency (per week) Population 18–34 35–44 45–55 56–64 Over 65
Sample 1000 229 189 197 135 150
5 days 10.3 % 9.9 % 9.5 % 10.5 % 9.8 % 11.6 %
4 days 3.9 % 2.3 % 1.4 % 5.3 % 4.5 % 5.6 %
3 days 11.4 % 12.8 % 11.2 % 13.1 % 11.6 % 9.0 %
2 days 16.6 % 10.5 % 21.8 % 17.4 % 16.5 % 17.6 %
1 day 19.9 % 17.9 % 23.7 % 21.65 24.2 % 15.2 %
Less than once a week 37.9 % 46.6 % 32.5 % 32.1 % 33.5 % 41.0 %
Average frequency 1.7 1.6 1.7 1.9 1.8 1.8
The second question, “How many times per week do you receive mail (from 1 to
5 days)?,” identifies the real needs of the consumers analyzed by age group.
An interesting element emerges from Table 2. The results confirm the findings
from Jimenez et al. (2006) where the mail received is linked to the age of the
recipient and the peak of the average frequency is the 45–55 age group mainly due
to the relation between the employment status and the amount of mail received.
However, the differences between age groups are less relevant than they found.
The responses presented in Table 3 are based on the location where the recipient
lives are interesting.
Residents in metropolitan areas, large cities, and very small towns seem to
receive more mail. The first two categories probably receive more mail because
they are more connected to economic activities, and the last category receives more
mail because mail is one of the few possible networks to be able to interact.
Two important results emerge from Tables 2 and 3. The first result is that nearly
75 % of the population receives mail twice or less per week; the average frequency
is 1.7. It is interesting to compare these results with those of a study carried out by
BCG 4 years ago (BCG 2011) when volumes across Europe were much higher.
They carried out a consumer survey encompassing 14 countries (Australia, Austria,
Canada, France, Germany, Italy, the Netherlands, Norway, Russia, Spain, Sweden,
Switzerland, the UK, and the USA). For each country, they surveyed 400 mail
recipients aged 18–65. They sampled the recipients’ requirements also taking into
account the willingness to pay the rough equivalent of the associated costs com-
pared with a basic no-charge service level (BCG 2011, p. 3). BCG found that the
average actual frequency required in the main European markets ranged from 2.3 to
3.1 per week per customer.
In case of Italy, the difference between 1.7 of our study and the 2.7 from the
BCG study can be explained by two factors, first of all volumes, like in other
European countries, dropped significantly in the three period, with double digit
drops year to year. Secondly, in 2011 delivery in Italy took place also on Saturday
while this was not the case for 2014 (6 % of respondents responded 6 days per
week). Figure 1 below is interesting because it shows the degree of service over-
specification varies among countries (BCG 2011, p. 4).
276
Table 3 How many times per week do you receive mail (from 1 to 5 days)? Response by type of residence
12
Real frequency Total Metropolitan Outskirts of Towns <2000 Towns 2001< and Cities 10,001< and Cities >50,001
(per week) Pop. areas metropolitan areas inhabitants <10,000 inhabitants <50,000 inhabitants inhabitants
Sample 1000 148 116 58 239 268 171
5 days 10.3 % 15.7 % 7.3 % 11.1 % 7.8 % 10.1 % 11.4 %
4 days 3.9 % 4.6 % 1.5 % 4.7 % 4.9 % 3.2 % 4.2 %
3 days 11.4 % 11.3 % 10.3 % 10.4 % 14.7 % 10.3 % 10.0 %
2 days 16.6 % 13.0 % 22.0 % 18.5 % 14.3 % 15.4 % 20.3 %
1 day 19.9 % 22.1 % 21.7 % 22.6 % 22.9 % 20.1 % 11.3 %
Less than one 37.9 % 33.4 % 37.2 % 32.7 % 35.4 % 40.9 % 42.9 %
delivery
Average 1.7 2.0 1.6 1.8 1.7 1.7 1.8
frequency
S. Gori et al.
2.3 2.5 2.6 2.6 2.7 2.7 2.7 2.8 2.9 3.0 3.0 3.0 3.1 3.1 Average days/wk
Percentage 100
of respondents
80
60% 61% 58% 58%

64% 63%
68% 71% 71% 69% 68%
74%
75%
60 85%
40
14% 13%
15% 14%
12% 13%
12%
19% 14% 11% 13% 2 days/wk
12% 17% 16%
20 8% 14% 16%
11% 19% 23% 3 days/wk
6% 10% 16% 11%
11% 12% 9% 5 days/wk
10% 10% 11% 13%
9% 6% 7% 9% 6% 6 days/wk
3% 2% 3% 4% 3% 4%
0
ES CA RU FR IT AU NL DE SE AT USA UK CH NO
Country
Highest overspecification
Lowest frequency needs including fee
Fig. 1 Benchmark on frequency in delivery in main European countries and the USA. Source:
The Boston Consulting Group—The postman always brings twice, 2011
Table 4 presents results from the third survey question: How would you perceive
a reduction of the current quality of service from 5 days per week to 2/3 days per
week? (Completely/very dissatisfied, dissatisfied, satisfied, very satisfied).
From Table 4 emerges a very strong finding. Nearly 75 % of the population
would be satisfied or very satisfied if the frequency would be reduced from 5 days
per week to 2/3 days per week. This finding does not depend on the size of the city
of residence of the respondent.
To better understand this result, it is important to consider the reasons they
would be satisfied or very satisfied with a lower frequency. Hence, we asked the
fourth question, the main reason for their response. Table 5 summarizes the stated
reasons.
For nearly 50 % of the 744 persons who responded that they would be satisfied or
very satisfied with a reduction in delivery frequency, their reasons were linked to
low volumes or to current actual needs. This finding matches the low propensity to
spend for postal services and the low volumes per capita, and it reinforces the link
between national attitudes toward mail versus other media and volumes per capita.
4 Conclusion: Implications of the Results

and Way Forward
Not much emphasis has been placed in the past in assessing mail recipients’
preferences for frequency of mail delivery. This chapter attempted to address this
gap by analyzing the result of a consumer survey, where the perceived quality in
278
Table 4 How would you perceive a reduction of the current quality of service from 5 days per week to 2/3 days per week?
12 Outskirts of Cities
Metropolitan metropolitan Towns <2000 Towns 2001< and Cities 10,001< and >50,001
Response Population areas areas inhabitants <10,000 inhabitants <50,000 inhabitants inhabitants
Very satisfied 15.3 % 15.4 % 17.7 % 19.4 % 15.0 % 11.4 % 18.4 %
Satisfied 59.1 % 56.9 % 58.0 % 54.7 % 58.6 % 63.3 % 57.3 %
Dissatisfied 17.9 % 18.5 % 17.2 % 21.5 % 19.6 % 15.6 % 18.1 %
Very dissatisfied 7.7 % 9.3 % 7.2 % 4.5 % 6.7 % 9.7 % 6.2 %
Very satisfied + 74.4 % 72.3 % 75.7 % 74.1 % 73.6 % 74.7 % 75.7 %
satisfied
S. Gori et al.
Table 5 Main reasons for judging satisfied or very satisfied with a reduction in the frequency
Population
Sample
744
The new frequency is adequate to my personal needs 26.4 %
I receive very low volumes of mail 21.0 %
It is indifferent for me 17.5 %
I never had problems with delivery of mail and I do not expect them in the future 7.1 %
It is ok for me, the important thing is that it does not have to create delays in the 5.2 %
delivery of my mail
I do not receive urgent mail 4.9 %
Currently, the level of service in my area is already 2/3 days per week 2.3 %
To avoid further delays in the delivery of mail 2.3 %
It could be a way to rationalize and to improve the service 2.1 %
delivery (measured by the frequency of delivery) was compared to the actual

delivery of at least one piece of mail taking place (4.2 perceived vs. 1.7 real quality
of service).
We further asked consumers how they would perceive a reduction in delivery
frequency. The results from the consumer survey are simple and strong, despite
appearing counterintuitive at first sight. One would expect a strong negative
response to a reduction in the frequency of delivery. Instead it counterintuitively
emerged quite clearly that Italian people perceive that the current frequency of
delivery is by far superior to their needs. Furthermore, due to the difference
between perceived quality of service and needed quality of service, 75 % of the
population would not object to reducing the frequency. This is even more aston-
ishing because no reference was made to the consumers during the interview to the
fact that a higher frequency has a higher cost and thus a negative impact on the
general welfare. This finding brings good news, i.e., the possibility of optimizing
the network with a minor impact on consumers, but also bad news—the diminishing
importance of postal services for customers. It has important implications for the
future of mail and the long-term sustainability of postal operators: Are these the first
signs of the graveyard spiral of these services?
The way forward is to continue this approach and build up a method to identify
the optimal service level by maximizing total welfare taking into account consumer
preferences (both senders and recipients) and the cost of the delivery network,
considering also the eventual contribution of the State to finance the Universal
Service Obligation (the cost of the USO would increase with the increase in
frequency), as represented in Fig. 2. The Operator and State welfare (OSW) is the
sum of the Operator welfare, which falls with cost it bears, while State welfare falls
the greater is the welfare cost of taxes used to raise funds for the State contribution.
For future research, it would be useful to complement this analysis on con-
sumers’ preferences with an update on the analysis on the savings in different
countries where they change their current delivery behavior to three or fewer days a
280 S. Gori et al.
Welfare:
TW= CW+ OSW
Total (TW),
consumer (CW)
operator and
state(OSW)
OSW
CW
Frequency of delivery/days of delivery
Fig. 2 Total Welfare and frequency of delivery
week. It could be interesting to combine our current research with an update on the
analysis carried out by Boldron, Joram, Martin, and Roy (2006) that tried to
estimate savings in different countries if they changed their current delivery
frequency to 3 days a week. The model they developed was based on unit cost
with three variables: population density, average number of people per household
and postal traffic per person. Through their model, it may be possible to determine
the link between frequency of delivery, volumes, and economic sustainability of the
postal operator.
References
Accent (2008) Postal universal service obligation: value to the citizen. Report Prepared for
Postwatch
Boldon F, Joram D, Martin Felisberto C, Friedli B, Finger M, Krähenbühl D, Trinkner U (2006)
Pricing the last mile in the postal sector. In: Crew MA, Kleindorfer P (eds) Progress toward
liberalization of the postal and delivery sector. Springer, New York, pp 343–356
Cecchet M, Coulier J, Damgaard M, Doise N, Guillot C, Janin Lionel, Muller P, Swinand GP
(2011) Quality of service regulation and consumers needs: recent evidence from France.
Presented at the 19th Conference on Postal and Delivery Conference, Jersey, June 3, 2011
Copenhagen Economics (2010) Main developments in the postal sector (2008–2010), Final
Report, 29 Nov 2010, for DG MARKT. http://ec.europa.eu/internal_market/post/doc/studies/
2010-main-developments_en.pdf
Felisberto C, Friedli B, Finger M, Krähenbühl D, Trinkner U (2006) Pricing the last mile in the
postal sector. In: Crew MA, Kleindorfer P (eds) Progress toward liberalization of the postal and
delivery sector. Springer, New York
Jimenez L, Diakova E, Szeto C (2006) Generational analysis of mail users. In: Crew M,
Kleindorfer P (eds) Progress toward liberalization of the postal and delivery sector, vol
49, Topics in regulatory economics and policy. Springer, New York, pp 281–300
Owen BM, Willig RD (1983) Economic and postal pricing policy. In: Fleishman J (ed) The future
of the postal service. Praeger, New York
Panzar JC (1991) Is postal service a natural monopoly. In: Crew M, Kleindorfer P (eds) Compe-
tition and innovation in postal services. Lexington Books, Boston, pp 219–231
RAND Europe (2011) Technical report: study on appropriate methodologies to better measure
consumer preferences for postal services. Report for the European Union
Sawtooth Software, Technical Paper Series (2013) The CBC system for choice-based conjoint
analysis. https://sawtoothsoftware.com/download/techpap/cbctech.pdf
Schwarz-Schilling C (2001) Pricing schemes in liberalized postal markets. Presented at the IDEI
conference on Competition and Universal Service in the Postal Sector, University of Toulouse,
6–7 December
The Boston Consulting Group (2011) The postman always brings twice. BCG, Boston
TNS (2010a) Residential customer needs from a sustainable universal postal service in the UK.
TNS-BMRB Report for PostComm and Consumer Focus
TNS (2010b) Business customer needs from a sustainable universal postal service in the UK.
TNS-BMRB Report for PostComm and Consumer Focus
An Economist’s Perspective
on Terminal Dues
Henrik Ballebye Okholm, Anna M€oller Boivie,

Bruno Basalisco, and Julia Wahl
1 Introduction
The Universal Postal Union (UPU) system of terminal dues1 has at several instances
been criticized for its distortive nature. The terminal dues literature, for example
Campbell et al. (2011), Geradin (2012), Haller et al. (2012), WIK Consult (2013),
and Sorensen (2014) has recurrently underlined issues such as the distortion of
competition between delivery operators, distortions of international mail flows via
remail, financial transfers between designated operators, and distortions of compe-
tition between retailers in different countries.
These assertions become all the more relevant when considering the growth in
cross-border e-commerce. In fact, items bought online often weigh less than 2 kg
and therefore enter the letter post stream as small “packets” or bulky letters.
According to the Universal Postal Union (2015), around 80 % of the mail items
generated by e-commerce are sent in the letter post stream to which the UPU
terminal dues system applies.
To our knowledge, earlier analyses of the UPU terminal dues have only to a very
limited extent focussed on the economics of the system. The inspiration for this
chapter arose as a direct result of a study by Copenhagen Economics for the
U.S. Postal Regulatory Commission where we, based on the foundations of eco-
nomics, identified a number of market distortions created by terminal dues. Build-
ing on this study, this chapter goes one step further by investigating whether the
1
Terminal dues are the rates paid by a designated postal operator in the country of origin for the
last-mile delivery of cross-border letter post items delivered by another designated postal operator
in the destination country.
H.B. Okholm • A.M. Boivie (*) • B. Basalisco • J. Wahl
Copenhagen Economics A/S, København K, Denmark
e-mail: am@copenhageneconomics.com

284 H.B. Okholm et al.
economic distortions identified are not purely theoretical, but also are likely to
occur in practice.
The remainder of the chapter is structured as follows: After a brief introduction
to the UPU terminal dues system and the underlying economics in Sect. 2, we
identify the theoretical distortions created by the system in Sect. 3. Section 4
complements this analysis with the results of a mystery shopping experiment to
detect the actual impact that terminal dues may have on seller and buyer choices in
cross-border e-commerce. In Sect. 5, we discuss an alternative solution and the way
forward.
2 The UPU Terminal Dues System: Practice

and Underlying Economics
2.1 The UPU System for Terminal Dues
The Universal Postal Union (UPU) system of terminal dues governs payments
between designated postal operators for the transport, sorting, and delivery of
cross-border letter post items in the destination country. The scheme is made
available to the designated operators of the 192 UPU member countries. It functions
as a two-tiered system where countries are divided into target countries2 and
transition countries.3
All cross-border flows to, from, and between transition countries are subject to
fixed terminal dues rates (per item and per kilogram). Rates paid by target country
operators for last-mile delivery in other target countries are (in theory) calculated as
a share of the domestic letter post tariffs. However, in reality, over 80 % of letter
post flows by designated operators in target countries are subject to caps or floors.4
Since in most cases, the caps and floors are very close to each other, these rates are
transformed into de facto fixed rates.
This design of rates implies that terminal dues are de facto disconnected from the
cost of last-mile handling of the letter post items concerned. Furthermore, they are
not aligned with the price charged for last-mile handling of comparable domestic
letter post items in the destination country.
2
Countries that traditionally have been characterized as industrialized countries.
3
Countries that traditionally have been characterized as developing countries, but where several
countries today are highly industrialized.
4
Calculated using the UPU Tool “Terminal Dues Impact” (2012). The exact percentage of letter
post flows subject to either caps or floors depends on the subgroup of target countries, but always
exceeds 80 %.
An Economist’s Perspective on Terminal Dues 285
2.2 Economic Framework for Analysing Terminal Dues
Terminal dues are, in essence, a price paid by one designated postal operator to the
other for the last-mile handling of cross-border letter post items. Prices have an
important signalling function in the market, providing incentives to producers and
consumers to make economically efficient decisions (e.g. how many letters to send,
or from which operator to buy delivery). For competitive markets to function,
producers and consumers have to respond to the right price signals. Interventions
that change prices can thus distort price signals in the market and thereby incen-
tivize economic agents to make economically inefficient decisions.
Economists often distinguish between three types of economic inefficiencies:
allocative, productive, and dynamic inefficiency. In the context of cross-border
delivery of letter mail, distorted price signals can create three inefficiencies or
market distortions.
Firstly, distorted price signals can create an incentive for postal users to buy
more or less of a service compared to the optimal situation (allocative inefficiency).
Secondly, distorted price signals can create an incentive for postal users to buy
delivery services from a less efficient delivery operator or to mail their letters from
a less efficient location (productive inefficiency).
Thirdly, distorted price signals can prevent efficient competition in the cross-
border delivery market (dynamic inefficiency).
3 Distortions Created by the UPU System

for Terminal Dues
3.1 Methodology
To analyse the economic distortions that the current UPU terminal dues system
creates, we apply a value chain approach, analysing the cross-border journey of
three types of letter post items from sender to recipient: (1) business to consumer
(B2C) packets representing e-commerce or mail order packets, (2) bulk mail letters
(e.g. direct mail and transactional bulk mail), and (3) single piece letters
(e.g. birthday cards, invitations, and postcards).
Whereas bulk mail letters historically have represented the largest flow of cross-
border letter post items, growing e-commerce implies that a significant share of
terminal dues today apply to bulky letters or packets. This share will most likely
continue to grow.
This value chain consists of a number of decision makers whose decisions are
based on, among other things, price signals. Since terminal dues directly or indi-
rectly5 determine the payments between designated operators for last-mile handling
5
For designated postal operators who do not apply the UPU terminal dues, these rates still function
as a fall-back provision in negotiations (thereby affecting the bilaterally negotiated rates).
B2C packets
Price for cross- Price for last-

Price for cross-
border delivery mile delivery in
border delivery Terminal dues
Price decisions of e-commerce
(end-to-end)
destination
product country
First mile Last mile

E-retailer E-shopper
Decision makers delivery operator delivery operator
Fig. 1 Effects of terminal dues on prices in the value chain. Source: Copenhagen Economics
of cross-border letter post items (including packets), they may affect the decisions
made by the agents all along the value chain.
Terminal dues directly influence decisions made by the last-mile operator since
they determine the price at which he offers last-mile delivery of cross-border letter
post items originating from another designated operator. The price at which the last-
mile operator offers last-mile delivery of cross-border letter will, in turn, directly
influence the first-mile operator’s decision on which operator to use for last-mile
handling of his mail. As terminal dues paid for last-mile delivery constitute a cost to
the first-mile operator, they may also influence the price charged by the first-mile
operator for end-to-end cross-border delivery.
If terminal dues influence the price charged by the first-mile operator for end-to-
end delivery, they may also have an impact on senders’ decisions (e.g. e-retailers’
decisions regarding what products and delivery options to offer, and to what price to
offer them). By the same token, if terminal dues are reflected in the delivery prices
charged by e-retailers for items below 2 kg, they might affect buyers’
(i.e. e-shoppers’) choices when buying items online, cf. Fig. 1.
3.2 Distortions Revealed by the Value Chain Analysis
Following the value-chain approach, we identify six distortions that the current
terminal dues system induces in the value chain. In the following, we will outline
the way in which terminal dues may distort the decisions of agents in the three
phases of the value chain: the last-mile phase, the first-mile phase, and the buying
phase.
3.2.1 Distortions of Decisions in the Last-Mile Phase
The UPU terminal dues system may distort competition for the last-mile handling
of cross-border letter mail. This happens if the terminal dues charged by the
designated operator are below the long-run average incremental cost of handling
cross-border letters in the last mile. If this is the case, a non-designated operator,
which is as efficient as the designated operator (i.e. with the same costs as the
designated operator) would not be able to compete in a profitable way, since
charging a similar price for last-mile handling would lead to losses for the
non-designated operator.
The size of this distortion might be limited since it is restricted to liberalized
markets. In fact, even in most liberalized markets (e.g. in Europe), designated
operators hold market shares for last-mile delivery of letter post beyond 95 %.
In addition to the distortion of competition, if terminal dues do not cover the cost
of last-mile handling, this means that designated postal operators make losses on
inbound cross-border items. In some situations, these losses might have to be
financed by the taxpayers.6
3.2.2 Distortions of Decisions in the First-Mile Phase
The UPU terminal dues system may distort competition for the first-mile handling
of cross-border letter mail. Firstly, low UPU terminal dues constitute a cost-
advantage for designated operators vis-a-vis non-designated operators since
non-designated operators are not part of the system. This makes it more difficult
for non-designated operators to compete for the first-mile handling of cross-border
letter mail. Secondly, low terminal dues may make cross-border delivery by small
packets (i.e. bulky letters) relatively cheaper than delivery by parcel post. Hence,
parcel services (supplied by both designated and non-designated operators) may
lose competitiveness relative to the letter post services supplied by designated
operators. As UPU (2015) estimates that around 80 % of items bought online
weigh less than 2 kg and are processed in the letter-post mail stream (subject to
terminal dues), this distortion is highly relevant today and its importance will grow
with the continued growth of e-commerce.
Since there is virtually no competition on first-mile handling of single-piece
items, the main distortion for the first-mile phase most likely lies in first-mile
competition for letter mail items sent in bulk and the disadvantages of parcel
delivery compared to packet delivery.
3.2.3 Distortions of Decisions in the Buying Phase
Terminal dues may also distort buyers’ decisions. This happens if they have an
impact on the prices charged by first-mile operators for end-to-end delivery.
Buyers’ (i.e. e-retailers’ and indirectly e-shoppers’ as well as single piece and
6
This is the case, when the losses incurred by the service provider result in an increase of the net
cost of the USP. This deemed an unfair burden on the universal service provider and might
therefore have to be resolved through some form of governmental financial support.
bulk mail senders’) demand for delivery services can be distorted in three
different ways.
Firstly, the demand distortion can take the form of excessive demand for cross-
border delivery. Low terminal dues reflected in disproportionately low prices for
cross-border delivery may create incentives to retailers to offer their products on
foreign markets. Similarly, they may also push consumers to buy products cross-
border instead of domestically.
Secondly, the demand distortion might arise as inefficient remail. Low terminal
dues reflected in disproportionately low prices for cross-border delivery might
incentivize bulk mailers to inject their letter post items for the domestic market in
another country, although this is less efficient than injecting the items directly in the
domestic market. For e-commerce packets, low cross-border prices may create
incentives to e-retailers to locate their fulfilment centres or warehouses in transition
instead of target countries.
Thirdly, demand distortions might arise in the form of excessive demand for
services that the UPU framework applies to (packets versus parcels). Low terminal
dues translating into disproportionately low prices for end-to-end delivery of
packets (i.e. bulky letters) may create incentives to e-retailers to ship by packet
delivery instead of parcel delivery. Similarly, e-shoppers might get an incentive to
choose packet delivery over an express or standard parcel delivery service.
The distortion of demand for packet versus parcel delivery in the buying phase is
likely to be a growing concern in the future. We have therefore investigated the
actual incidence of these distortions in a mystery shopping experiment.
4 e-Commerce: Mystery Shopping Experiment
The Universal Postal Union (2014) estimates that global e-commerce will grow by
15–20 % per year from 2014 to 2017. While cross-border e-commerce accounts for
roughly 10–15 % of total e-commerce volumes today, this share will increase
rapidly. According to BCG (2014), by 2025, annual global cross-border e-com-
merce revenues are expected to rise between $250 billion and $350 billion (from
$80 billion today). Owing mostly to the growth of cross-border e-commerce from
China, Asia will make up around 40 % of those revenues. This will make it the
largest cross-border e-commerce region in the world.7
Since a large share of the items bought online weigh less than 2 kg, a large part of
this cross-border e-commerce could be affected by distortions created by the UPU
7
In 2013, Chinese cross-border e-commerce amounted to € 460 billion (growing by 31 % from
2012) of GMV, see for example iResearch (2015), China Cross-border E-commerce Report, http://
californiacenter.us/wp-content/uploads/2015/02/2014-China-Cross-border-E-commerce-Report-
Brief-Edition.pdf, p. 17 and BCG (2014), Cross-border E-Commerce Makes the World Flatter,
https://www.bcgperspectives.com/content/articles/transportation_travel_tourism_retail_cross_bor
der_ecommerce_makes_world_flatter/.
terminal dues system. In the last section, we define a number of ways in which
terminal dues could distort e-retailers’ and e-shoppers’ demand for delivery of
e-commerce products. The empirical significance of the distortions investigated
in our mystery shopping experiment has recently been confirmed in a Congressional
hearing in the United States on fair competition in international shipping, where
several of the testimonials witnessed about disproportionately low shipping rates
for low weight e-commerce items originating from Asia (see U.S. Congress 2015).
To see how terminal dues could have an impact on the prices paid by e-shop-
pers—and thereby influence their buying decisions—we have conducted an illus-
trative shopping experiment on some of Asia’s largest online market places. In our
experiment, we investigated the delivery options for 50 typical orders8 of different
weights when shipped from the country of origin (China/Japan) to Europe
(Denmark). We chose to shop on the four major Asian platforms for cross-border
e-commerce: Aliexpress.com, lightinthebox.com, en.jd.com, global.rakuten.com
(see McDermott 2015). Our sample includes items that are cheap in price, relatively
light, and small in size (i.e. items that typically could be sent by either letter
(packets) or parcel delivery). This allows us to achieve a marginal increase in
weight and to isolate the weight effect from the impact that other characteristics
might have on the delivery costs and the delivery services offered.
For each product, we placed two orders. For the first order, the weight was
slightly below 2 kg (“low weight order,” average of 1.9 kg). For the second order,
we added one or a few items of the same product to the shopping basket, such that
the weight reached slightly more than 2 kg (“high weight order,” average of 2.2 kg).
We then observed how the sending option offered and the delivery cost changed
between the two orders.
Our results show that the shipping prices change significantly when the order
weight exceeds 2 kg. This is especially the case where light weight orders are sent
by letter mail (by means of a “small packet” or “registered mail” service) and high
weight orders (of the same product) are sent by an “EMS” or “expedited” service. In
these cases (which account for 50 % of all orders in our experiment), the average
shipping cost per kilogram increases from 2.6 euros to 17.8 euros—i.e. an increase
by over 500 %, cf. Fig. 2.
These findings indicate that the distortions created by terminal dues are not
purely theoretical, but also highly relevant in practice. In fact, in our mystery
shopping experiment, all light weight orders at Aliexpress, the largest
e-commerce platform in Asia, were sent by letter mail.
For about 40 % of the orders placed, free shipping was offered for the order
below 2 kg. For these orders, the delivery cost increased by nearly 40 euros on
average when the weight of the order increased from 1.9 to 2.2 kg. This indicates
that e-retailers “absorb” the low packet delivery price, while passing the higher
parcel delivery price onto the e-shopper (at least partially). This further strengthens
8
For instance, stationary, light clothing, electronic devices, and toiletry products.
Orders with light weight items sent by

All orders letter mail (packets)
20 20
18 18
Shipping cost (€ per kg)

Shipping cost (€ per kg)
16 16
14 14
12 12
10 10
8 8
6 6
4 4
2 2
0 0
Low weight orders (1.9 kg) High weight orders (2.2 kg) Low weight orders (1.9 kg) High weight orders (2.3 kg)
Fig. 2 Shipping costs for “low weight” vs. “high weight” orders. Source: Copenhagen Econom-
ics, based on mystery shopping experiment
the incentive for the e-shoppers to choose a packet delivery instead of a parcel
delivery option, or to split the order into two deliveries.
In those cases where also orders below 2 kg are shipped as parcels, the price
changes are far less significant (less than 1 euro per kg price difference) and
consumers’ incentives will not be distorted by the terminal dues. This result holds
over product types and prices. For lower price products, however, this means that
the delivery price often may be at least as high as the price of the order itself. This
raises the question whether consumers would at all opt to buy these products at this
high price for parcel delivery.9
Finally, we find that terminal dues provide both e-retailers and e-shoppers with
an incentive to benefit from cheaper delivery by splitting up single (heavier) orders
into several packet shipments. While our experiment did not allow us to test this
behaviour, websites such as Chinabuye explicitly state in their shipping choices for
registered air mail by postal operators that they may ship orders over 2 kg in two
packages, (cf. Fig. 3).
The results of our mystery shopping experiment clearly indicate that the current
design of the UPU terminal dues system creates distortions in relation to global
e-commerce delivered by letter post. For the share of light weight e-commerce
items sent by parcel delivery (for instance, due to the value or size of the item),
terminal dues do not create any distortions of global e-commerce.
As illustrated in our experiment, the fact that cross-border letter mail delivery of
light weight items below 2 kg is priced disproportionately low due to the terminal
dues system creates a distortion of competition between packet and parcel delivery.
It also creates a potential distortion of competition between domestic and cross-
border e-commerce, where e-shoppers get an incentive to buy items below 2 kg
from a foreign retailer instead of from a domestic one. This aspect was highlighted
9
Rakuten is the only platform for which we could clearly identify the offer of parcel delivery
(i.e. EMS) for orders of both above and below 2 kg. The choice of parcels delivery as the only/
cheapest option might thus be specific to this platform.
Registered Airmail (HongKong Post) With Tracking Number (7-

20 working days) - Hongkong Post does not carry battery
product.Your package weight is above 2kg,we may ship it in 2
packages.
$ 7.83
Registered Airmail (China Post) - With Tracking Number (20-35

working days) Your package weight is above 2kg,we may ship
it in 2 packages.
$ 1.50
Fig. 3 Example: e-retailer splitting order due to weight. Source: Chinabuye (2015)
by several witnesses in the recent hearing on international shipping in the

U.S. Congress (2015). Further investigation into this aspect of terminal dues
would thus be useful.
5 Towards a Non-distortive UPU System
The first step in determining the features of a non-distortive terminal dues system is
to point to the three drivers for the distortions under the current system. These are
(1) discrimination (between designated and non-designated operators, and between
operators in target countries and operators in transition countries), (2) discrepancy
between terminal dues and domestic prices for last-mile handling of comparable
letter post products, and (3) terminal dues lower than the long-run average incre-
mental cost of last-mile handling of cross-border letter post items.
Table 1 provides an overview of the distortions identified, their underlying
drivers, and the solution necessary to prevent the distortion from occurring.
Hence, a non-distortionary terminal dues system would have to fulfil the fol-
lowing three conditions:
Firstly, terminal dues would have to cover the cost of last-mile handling as this
would allow as-efficient last-mile operators to compete with designated operators
without losing money.
Secondly, setting terminal dues equal to the prices for last-mile delivery of
comparable domestic delivery services will serve to mitigate three distortions: it
would ensure that postal operators do not make losses on inbound cross-border mail
(assuming that domestic tariffs are cost covering); it would eliminate postal users’
bias towards packet delivery over parcel delivery; and it would eliminate postal
users’ bias towards cross-border over domestic delivery.
Thirdly, a non-discrimination requirement is the necessary condition for apply-
ing similar rates to designated and non-designated operators, and to operators in
target and transition countries. A sufficient condition for effective
non-discrimination would however be to set terminal dues equal to the domestic
Table 1 Economics distortions, their drivers, and solutions

Distortion Drivers Solution
Last-mile TD < last-mile LRAIC TD last-mile LRAIC
competition
First-mile Discrimination (designated Non-discrimination
competition vs. non-designated operators)
Packets TD < last-mile price for TD ¼ last-mile price for comparable
vs. parcels comparable domestic services domestic services
Domestic vs. TD < last-mile price for TD ¼ last-mile price for comparable
cross-border comparable domestic services domestic services
delivery
Target Discrimination (target Non-discrimination and TD ¼ last-mile
vs. transition vs. transition countries) price for comparable domestic services
country origin
Financial TD < last-mile price for TD ¼ last-mile price for comparable
transfers comparable domestic services domestic services
Note: TD Terminal dues, LRAIC Long-run average incremental cost
Source: Copenhagen Economics
price for last-mile activities and require that this price is charged to all operators,
irrespective of their origin or status as designated or non-designated operator.
The implementation of such a terminal dues system may, however, encounter a
number of practical and political limits. From a practical standpoint, finding one or
several domestic product(s) that are comparable to the cross-border letter mix in
terms of weights and formats may be complicated. Moreover, as prices for letter
post products outside the USO scope may not be monitored by regulatory agencies,
this may create challenges with respect to ensuring compliance with the system.
Politically, the risk lies in the fact that the proposed changes may reduce the
affordability of cross-border letter mail as well as the profitability of designated
operators that may pay more for outbound cross-border mail (and potentially pass
on this higher cost to their customers in terms of higher postage rates). Although
this would be efficient from an economic point of view, it may not be socially
desirable. Finally, a change of the system would risk to reduce the attractiveness of
e-retailers in transition countries to e-shoppers in target countries, thereby reducing
their competitiveness.
However, this does not mean that the economically optimal solution should not
be implemented. Instead of trying to solve the practical and political concerns
within the terminal dues system, one could accompany the system with a number
of targeted policies outside the terminal dues system. For instance, an aid
programme or compensation scheme could ensure the continuing affordability of
cross-border letter mail in transition countries and the sustainability of provision of
universal services.10 This solution would have to be complemented by clear rules
10
At present, the Quality of Service Fund already in place for transition countries is an example of
such a compensation fund.
for the eligibility to the compensation scheme as well as trade facilitating measures
to counteract the decreased competitiveness of e-retailers in transition countries. In
this way, the non-distortive character of the terminal dues system would be ensured
while additional issues could be addressed with different policies over time.
6 Conclusion
We have identified six theoretical distortions created by the UPU system of terminal
dues and showed the practical implication of the system on e-commerce from Asia.
Our shopping experiment indicated the practical effects this system may have on
cross-border e-commerce from Asia to Europe.
In order to create a non-distortionary system, three changes are necessary.
Firstly, terminal dues would have to cover the cost of last-mile handling. Secondly,
terminal dues would have to equal the prices for last-mile delivery of comparable
domestic delivery services. Thirdly, terminal dues would have to be
non-discriminatory.
While the implementation of these changes may be less straightforward for
practical and political reasons, we propose to address these issues by means of
other targeted polices. Nevertheless, any discussion about changes to the current
terminal dues system should carefully weigh the political and practical costs of the
changes to a new system against the actual size of the distortions currently created.
Further research on the UPU terminal dues system should therefore concentrate on
qualifying and quantifying the distortions further. Several methods, from simple
comparisons of cross-border mail and trade flows over gravity models towards a
difference in differences analysis, lend themselves for such a quantification exer-
cise. Moreover, computable general equilibrium models would allow us to measure
the effects of these distortions on the wider economy, i.e. on social welfare, prices,
imports and exports, employment, and GDP. The rise of e-commerce and the
demonstrable effect on practical e-commerce of our experiment are compelling
reasons to take terminal dues research to the next level.
References
BCG (2014) Cross-border e-commerce makes the world flatter. BCG, Boston
Campbell James I, Alex Kalevi Dieke, Martin Zauner (2011) Terminal dues: winners, losers, and
the path to reform. Presentation by Alex Dieke at the 19th Conference on Postal and Delivery
Economics, CRRI, St. Helier, Jersey, 1–4 June 2011
Chinabuye (2015) http://www.chinabuye.com/checkout/cart/
Geradin Damien (2012) Legal opinion on the compatibility of the proposed target system for
terminal dues with EU law
Haller Andreas, Christian Jaag, Urs Trinkner (2012) Termination charges in the international
parcel market. Swiss Economics Working Paper
iResearch (2015) China cross-border e-commerce report, p 17

McDermott K (2015) Key-business-drivers-in-cross-border-ecommerce 2014. Payvision BV,
New York
Sorensen Janet M (2014) International terminal dues white paper, Office of Inspector General,
United States Postal Service
Universal Postal Union (2015) Activities, letter post development, about letter post development.
http://www.upu.int/en/activities/letter-post-development/about-letter-post-development.html
Universal Postal Union (2014) Fulfilling the global e-commerce promise. UPU, Berne
U.S. Congress (2015) Hearing on fair competition in international shipping on 16 June 2015.
Committee on Oversight and Government Reform, Subcommittee on Government Operations.
https://oversight.house.gov/hearing/fair-competition-in-international-shipping/
WIK Consult (2013) Main developments in the postal sector (2010–2013). Study for the European
Commission, Directorate General for Internal Market and Services
A Game-Theoretic Model of the Market
for International Mail
Edward S. Pearsall
1 Introduction
The terminal dues system, the system of rates that postal operators (POs) set for the
delivery of each other’s international mail, is the fundamental feature that must be
explained by any model of the market for international mail. In this chapter, I apply
elements of game theory to show how the terminal dues system is structured to
maximize the combined profits on international mail of the participating POs. This
leads to a system of rates that are set at marginal cost when there are no entrant/
competitors (ECs), and at or above marginal cost for those POs that face ECs.
My model extends and applies two of John Nash’s contributions to game theory.
The price paid by postal customers at the collection end of an international mail
market is determined by a Nash equilibrium for a noncooperative non-zero-sum
two-person two-stage game. The two players are the originating PO, which is
always present in the market, and a single originating EC who may or may not
enter the market. In the first stage, the PO and the EC make simultaneous decisions.
The PO commits to a price that it charges whether or not the potential entrant comes
into the market, and the EC decides whether to enter. In the second stage, if the EC
enters, it sets the price for its service given the price the PO chose in the first stage.
The Nash equilibrium takes one of three forms, depending on the underlying cost
and demand conditions: the PO sets a monopoly price with no entry by the EC, the
PO and EC form a duopoly with the EC taking the PO’s first stage price as given and
monopoly/duopoly with limit pricing by the PO. In the limit pricing case, equilib-
rium takes the form of a price for the PO that leaves the EC indifferent between
The views expressed in this chapter are those of the author. I am especially indebted to Margaret
Cigno and the editors for their help and encouragement.
E.S. Pearsall (*)
Independent Consultant, Alexandria, VA, USA
e-mail: espearsall@verizon.net

296 E.S. Pearsall
entering or not, and a mixed strategy of entry and exit for the EC with a probability
of entry consistent with the PO’s limit price.
Terminal dues are the result of a fair division of profits that can be explained by a
generalized version of the Nash bargaining model. The POs are treated as players in
a cooperative n-person game with strategies corresponding to the terminal dues
rates they charge each other. In principle, the payoffs to the individual POs are the
additional net profits they receive from participating in the system rather than
individually setting nondiscriminatory rates for delivery access. When side pay-
ments are possible, the Nash solution to the game is to set the rates to collectively
maximize the aggregate added profit that the POs derive from international mail
and then to redistribute the added profit according to pre-agreed proportions.
The model shows that the participating POs acting together have an incentive to
set terminal dues rates at levels that are at or above marginal cost. However,
terminal dues rates are generally lower than the access rates that the POs charge
ECs and domestic mailers for the delivery of comparable mail. This occurs because
delivering mail is an activity characterized by declining average costs. Virtually, all
POs set access rates well above marginal cost in order to recover some of the
non-variable cost of providing delivery service and to exploit the market power they
have over delivery in their domestic markets.
Terminal dues rates serve about the same purpose as transfer prices within a
decentralized multi-department enterprise. Transfer prices are the accounting prices
at which the departments of the enterprise transact with each other. Altogether, the
terminal dues paid by POs on the originating end of international mail markets are
exactly offset by the terminal dues received by POs on the delivery end. The role of
the terminal dues rates within the system is to create incentives for the originating
POs to set international mail rates that maximize the aggregate added profit of both
the originating and delivering POs.
In the absence of competition or if access rates do not discriminate against ECs,
the aggregate profit is maximized when the POs set terminal dues rates that equal
the marginal cost of delivery. This is the same rule that an efficient enterprise will
follow to set transfer prices. When terminal dues rates are set this way they are
nondiscriminatory.
However, international mail markets are generally open to entry by ECs at
access rates that exceed marginal cost. An important prediction of the model is
that when the POs encounter such competition they are encouraged to respond with
terminal dues rates that are also above marginal cost. The intuition is that raising the
rate increases the EC’s sales and thus the profits the terminating PO obtains from
access charges. The rates also become discriminatory with respect to the POs
themselves. Higher rates arise because setting a terminal dues rate at marginal
cost in the presence of competitors who pay higher access rates usually makes it
possible to increase the combined profit of the POs by raising the terminal dues rate.
Section 2 describes how a single liberalized international mail market can be
modeled as a noncooperative non-zero-sum two-person two-stage game between a
PO and an EC. Section 3 examines the effects of changes in the terminal dues rate
on a single international mail market under the assumption that the terminal dues
A Game-Theoretic Model of the Market for International Mail 297
rate for the originating PO and the access rate for a potential EC are predetermined.
Section 4 describes the Nash bargaining model for a multiplayer cooperative game
and applies it to the way that POs cooperate to set terminal dues rates and side
payments. In Sect. 5 I show how the terminal dues rates are related to marginal
costs. Section 6 concludes the chapter.
2 Equilibrium in a Single International Mail Market
Two markets for international mail exist for every pair of countries whose POs
exchange mail. In each country, the PO accepts mail for delivery in the other
country and sets a price for this mail. For simplicity, we treat this outgoing mail
as a single service with a single price. Similarly, we consider only a single EC that
also offers a single outgoing mail service at a single price. The service offered by
the EC may not be identical to that offered by the PO so the two goods are not
necessarily perfect substitutes. In subsequent sections, it is assumed that the two
mail services have the neoclassical demand properties of imperfect substitutes.
A single outgoing international mail market with sequential pricing may be
modeled as a noncooperative non-zero-sum two-person two-stage game. The two
players are the PO and the EC offering the substitute service. In the first stage of the
game, the PO’s pure strategies are the prices that it may set for international mail
service to the second country, denoted PI (for “Incumbent”). The PO’s price is set
before the PO knows whether the entrant has entered. The EC may be either in or
out of the market; the incumbent must always remain in. Both entry and exit by the
EC have no associated fixed costs. The EC is assumed not to know the PO’s price at
the time that it chooses to be either in or out.1 In the second stage of the game, when
both the PO and the EC are present in the market, the EC sets his price PE (for
“Entrant”), taking the PO’s price (set in the first stage) as given. The PO’s price
remains unchanged for a single play of the game.
The model is generally appropriate when a regulated enterprise is required to
remain in a liberalized market. This is the natural condition of postal markets for
international mail. Regulation makes the price of the incumbent PO sticky. Entry
and exit by an EC, although costless, is also sticky but for a different reason. Entry
and exit typically require a substantial lead time. However, the EC’s price is not
subject to regulation and can be changed much more quickly than either the POs
price or the EC’s status as in or out of the market. The two-stage game fits these
facts on the ground.
The payoffs to the two players are determined by their profit functions which are
assumed to be differentiable and concave. If the EC chooses to enter, it takes the
PO’s price, PI, as given and sets its own price, PE, to maximize its profits, fE(PI, PE).
1
This assumption differentiates this model from a conventional limit-pricing model. In a conven-
tional model, it is assumed that the EC knows the PO’s price when making this choice.
298 E.S. Pearsall
Since the EC’s profit function is concave, the EC’s profit function always reaches a
maximum at a single price. Therefore, the EC sets a price according to the reaction
function PE ðPI Þ ¼ ArgMaxPE f f E ðPI ; PE Þg; which is also differentiable.
The PO has two profit functions. The PO’s profit function, with the EC in the
market, is fI(PI, PE)in, and, with the EC out, is fI(PI)out. fI(PI)out is strictly concave
with respect to the PO’s price, PI. fI(PI, PE)in is strictly concave with respect to PI
after we insert the EC’s reaction function for PE, i.e., fI(PI, PE(Pi))in is a strictly
concave function of PI. Both functions have single maximums with respect to PI.
Therefore, if entry is forthcoming, we need not consider mixed strategies for
the PO.
However, a mixed strategy of entry and exit for the EC may arise if, at the
optimal price chosen by the PO, the EC would be indifferent to entry.2 The Nash
equilibrium in this case is an equilibrium pair consisting of a price, PI, for which the
EC is indifferent between entering or not, and a mixed strategy indexed by a
probability of entry for the EC which is profit maximizing for the PO. A mixed
strategy for the EC is represented by introducing a variable μ in the range
0 μ 1, where the EC is in with probability μ and out with probability 1 μ.
To find μ, we differentiate E½ f I ¼ μ f I ðPI ; PE Þin þ ð1 μÞ fhI ðPI Þout withi respect to
d f Iout d f Iout d f Iin
PI, set the result equal to zero, and solve for μ ¼ dPI = dPI dPI , with the
derivatives evaluated at equilibrium.
The Nash equilibrium can take one of three forms (see Pearsall 2011):

E1 (monopoly): f E ðPI ; PE Þ < 0 for PE ¼ PE ðPI Þ: PI ¼ ArgMaxPI f I ðPI Þout and
the EC stays out ðμ ¼ 0Þ.
E2 (duopoly with n sequential o pricing): f E ðPI ; PE Þ > 0 for PE ¼ PE ðPI Þ:
PI ¼ ArgMaxPI f I ðPI ; PE Þin , and the EC always enters ðμ ¼ 1Þ, setting
PE ðPI Þ ¼ ArgMaxPE f f E ðPI ; PE Þg.
E3 (mixed monopoly/duopoly with limit pricing): The PO sets PI such that f E
ðPI ; PE Þ ¼ 0 for PE ¼ PE ðPI Þ. The EC nenters with a probability of entry
o
0 < μ < 1, which satisfies PI ¼ ArgMaxPI μ f I ðPI ; PE Þin þ ð1 μÞ f I ðPI Þout .
The three forms of equilibrium conform well to what we generally observe in

international mail markets. PO monopolies are the most frequently observed form.
This is because it is almost always impossible for an EC to cover its costs and make
a profit in a small international mail market. At the other extreme, large interna-
tional mail markets usually attract one or more permanent ECs while the PO plays
the role of the price leader. Often the ECs are other POs who maintain Extraterri-
torial Offices of Exchange (ETOEs) partly for the purpose of collecting mail.
However, the number of international mail markets with permanent competitors
to the PO is relatively small and mostly limited to the large industrial countries of
2
In the standard limit-pricing model mixed strategies for the EC are made irrelevant by the
assumption that the EC knows the PO’s price when choosing to be in or out of the market.
Europe and North America. In these countries, an EC can cover its costs if it can
attract even a small proportion of the total volume of originating international mail.
In many medium-sized but developed countries ECs, in the form of ETOEs and
private enterprises, come and go but often cause scarcely a ripple in the PO’s prices.
This is exactly what we would expect from repeated plays of a game with the
equilibrium E3. The POs limit pricing in such countries tends to place the ECs on
the cusp of profitability. Entry and exit appear to be random and to have little effect
on prices. However, entry and exit without price changes is the hallmark of limit
pricing because it does not occur with other kinds of market organization. For
example, if an entering EC permanently changes a market from a PO monopoly
(E1) to a duopoly with sequential pricing (E2), we would expect to see a fall in the
PO’s price.
3 Terminal Dues and Entry
In this section, we examine how terminal dues affect an international mail market
under the assumption that both terminal dues and access rates are predetermined.
What we find, first, is that lowering terminal dues, while access rates remain fixed,
suppresses entry by potential competitors to an incumbent PO. And, second, this
will increase the incumbent’s profit on outgoing international mail.
A potential competitor is affected somewhat differently depending upon the
form of equilibrium in the market, so we proceed case-by-case after introducing
some additional notation and assumptions.
t: terminal dues rate for the PO.
a: access charge for the EC.
c: marginal cost of delivery of incoming international mail (to be used in
Sect. 5).
mI, mE: marginal cost per piece of collection, processing, and transportation of
the Incumbent PO and the Entrant EC, respectively.
FI, FE: specific fixed costs of international mail service of the PO and the EC.
QI(PI)out: demand function for the PO’s international mail with the EC out.
QI(PI, PE)in: demand function for the PO’s international mail with the EC in.
QE(PI, PE): demand function for the EC’s international mail with the EC in.
These demand functions are all assumed to be differentiable and to have partial
derivatives with signs that accord with the neoclassical properties of two substitute
goods.
For simplicity, we assume that the PO and the EC have linear cost functions. The
profit functions of Sect. 2 can now be written more specifically as:
f I ðPI Þout ¼ ðPI t mI ÞQI ðPI Þout FI the PO’s profit with the EC out.
f I ðPI ; PE Þin ¼ ðPI t mI ÞQI ðPI ; PE Þin FI the PO’s profit with the EC in.
f E ðPI ; PE Þ ¼ ðPE a mE ÞQE ðPI ; PE Þ FE the EC’s profit with the EC in.
300 E.S. Pearsall
The EC’s reaction function PE ðPI Þ ¼ ArgMaxPE f f E ðPI ; PE Þg plays a role in all
three forms of the Nash equilibrium. The function describes the solution to an
equation that is obtained by differentiating fE(PI, PE) with respect to the EC’s price
PE while the PO’s price PI remains fixed. This derivative is equal to zero at the price
PE that maximizes the EC’s profit if the EC is in.
∂fE ∂QE
¼ QE ðPI ; PE Þ þ ðPE a mE Þ ¼ 0:
∂PE ∂PE
An EC’s price will generally respond positively to changes in the incumbent’s

price, that is, the slope of the EC’s reaction function dP
dPI is nonnegative if we ignore
E
secondary effects. To show this, we take the total differential of the equation above:
2
∂QE ∂QE ∂ QE
dPI þ 2 dPE þ ðPE a mE Þ dPE ¼ 0:
∂PI ∂PE ∂P2E
From neoclassical demand theory we have ∂Q∂PE

E
0 and ∂Q
∂PI
E
0 for substitute
h i 2
∂QE ∂QE ∂ QE
goods, therefore dPI ffi ∂PI = 2 ∂PE 0 when ∂P2 ffi 0; that is, when demand is
dPE
E
roughly linear. This result demonstrates that the mail services offered by the PO and
EC are strategic complements.
The EC’s profit responds positively to changes in the incumbent’s price PI. Take
the derivative of fE(PI, PE) with respect to PI:

d f E dPE ∂QE ∂QE dPE
¼ Q ð PI ; PE Þ þ ð PE a m E Þ þ
dPI dPI E ∂PI ∂PE dPI
At the price that maximizes the EC’s profit, we have

∂QE
QE ðPI ; PE Þ ¼ ðPE a mE Þ ∂PE . Substituting for QE(PI, PE), we have:

d f E dPE ∂QE ∂QE ∂QE dPE
¼ ðPE a mE Þ þ ðPE a mE Þ þ :
dPI dPI ∂PE ∂PI ∂PE dPI
The right-hand side simplifies to leave d fE

dPI ¼ ðPE a mE Þ ∂Q
∂PI
E
with the EC in.
∂QE
∂PI
0 for substitute goods. PE a mE 0 on any occasion that finds the entrant
in because the EC will never set a price that does not cover his variable cost.
Therefore, ddPf EI 0:
E1 (monopoly): Terminal dues affect the market by changing PI. This, in turn,
affects the EC’s non-entry condition f E ðPI ; PE Þ < 0 for PE ¼ PE ðPI Þ. This chain of
events is described by the derivative ddtf E ¼ ddPf EI ∂P
∂t
I
. From above ddPf EI 0, so a
decrease in terminal dues decreases the EC’s profit on entering if ∂P
∂t
I
> 0.

From E1 we have PI ¼ ArgMaxPI f I ðPI Þout . The price PI that maximizes the
incumbent’s profit solves:
∂ f Iout ∂QIout
¼ QI ðPI Þout þ ðPI t mI Þ ¼ 0:
∂PI ∂PI
∂QIout
∂PI
0 at a maximum because QI ðPI Þout 0 and because the PO will always set a
price to cover his variable costs, so PI t mI 0. The total differential of the
maximum profit condition is:
2
∂QIout ∂QIout ∂ QIout
2 dPI dt þ ðPI t mI Þ ¼ 0:
∂PI ∂PI ∂P2I
With linear demand, half of a change in the terminal dues rate will be passed on to
2
∂ Q out
the PO’s customers as a decrease in price, dP dt ffi 1=2 when ffi 0. Also,
I I
∂P2I
h i
∂QE ∂QE
dt ffi ∂PI = 4 ∂PE 0, the EC’s price after entry falls when terminal dues are
dPE
lowered.
We can see now that lower terminal dues deter entry by suppressing the EC’s
profit on entry. Recall that PE a mE 0 and ∂Q
∂PI
E
0 for substitute goods.
d f E d f E ∂PI ∂QE
¼ ffi ðPE a mE Þ =2 0:
dt dPI ∂t ∂PI
Lower terminal dues will also increase the PO’s profit. Differentiate the PO’s profit
function with respect to t:

∂ f Iout dPI dQ out
¼ 1 QI ðPI Þout þ ðPI t mI Þ I :
∂t dt dt
dQIout ∂QIout dPI ∂QIout

dt ¼ ∂PI dt ffi ∂PI =2 ffi 1=2
dPI
Substitute and dt to obtain
∂ out
I Þ ∂QI
QI ðP2I Þ þ ðPI tm
f Iout out
∂t
ffi 2 ∂PI
. Substituting from the differential equation that
∂QIout
maximizes the incumbent’s profit ð PI t m I Þ ∂PI
¼ QI ðPI Þout leaves
∂ f Iout out
∂t
ffi QI ðPI Þ . This shows that the PO captures approximately the entire cost
saving from a decrease in terminal dues when the international mail market is a PO
monopoly.
E2 (duopoly with sequential pricing): The analysis of E2 parallels that of E1. To
see why, we insert the EC’s reaction function PE(PI) in the PO’s demand function
QI(PI, PE)in with the EC in. With this insertion, the demand function accounts for
the presence of the EC implicitly and becomes a function of only PI. Likewise, PE
no longer appears as a variable in the PO’s profit function
302 E.S. Pearsall
f I ðPI , PE ðPI ÞÞin ¼ ðPI t mI ÞQI ðPI , PE ðPI ÞÞin FI . Now, when we go through
the same steps as in the monopoly case we get virtually the same results.
Again, terminal dues affect the market by altering the EC’s entry condition
f E ðPI ; PE Þ > 0 for PE ¼ PE ðPI Þ. A decrease in terminal dues will depress the EC
profit if ∂P I
∂t n
> 0. With linear demand, ∂P I
ffi 1=2; as with E1. We have
o ∂t
in
PI ¼ ArgMaxPI f I ðPI ; PE Þ . The price PI that maximizes the PO’s profit solves
the differential equation:
in
∂ f Iin in ∂QI ∂QIin dPE
¼ QI ðPI ; PE Þ þ ðPI t mI Þ þ ¼ 0:
∂PI ∂PI ∂PE dPI
∂QIin ∂Q in in
∂PI
þ ∂PIE dP
dPI 0 at a maximum because QI ðPI Þ 0 and because the PO will
E
always cover his variable costs, so PI t mI 0. Ignoring second-order effects,

the total differential of the maximum profit condition is:
in
∂QIin ∂QIin dPE ∂QI ∂QIin dPE
2 þ dPI þ dt ffi 0 :
∂PI ∂PE dPI ∂PI ∂PE dPI
From which we again obtain the simple result, dPI

dt ffi 1=2, and
dPE
dt ffi ∂Q
∂PI
E
=4 ∂Q
0.E
∂PE
All of the results for E1 also hold for E2. As before,
d fE d f E ∂PI ∂QE
dt ¼ dPI ∂t ¼ ðPE a mE Þ ∂PI =2 0. Lower terminal dues suppress the
h i
EC’s profit. And dPdtE ffi ∂Q
∂PI
E
= 4 ∂QE
∂PE
0 ; the EC’s price falls when terminal
dues are lowered. Lower terminal dues will also increase the PO’s profit from
international mail. Differentiate the PO’s profit function with respect to terminal
∂ f Iin dQ in
dues, ¼ dP
dt
I
1 QI ðPI ; PE Þin þ ðPI t mI Þ dtI . Substitute
h ∂t i h i
dQIin ∂QIin ∂QIin dPE dPI ∂QIin ∂QIin dPE
dt ¼ ∂PI þ ∂PE dPI dt ∂PI þ ∂PE dPI =2 dt ffi 1=2
dPI
and to obtain
h in i
∂ fI in
I Þ ∂QI ∂Q
¼ QI ðPI2;PE Þ þ ðPI tm
in in
∂t 2 ∂PI
þ ∂PIE dPdPI . Substituting from the differential
E
hequation that
i maximizes the PO’s profit, ðPI t mI Þ
∂QIin ∂QIin dPE in ∂ f Iin in
∂PI
þ ∂PE dPI ¼ QI ðPI ; PE Þ leaves ∂t ffi QI ðPI ; PE Þ . Again, the PO cap-
tures approximately the entire cost saving from a decrease in terminal dues when
the international mail market is a duopoly with sequential pricing.
E3 (mixed monopoly/duopoly with limit pricing): A small change in the termi-
nal dues rate has no effect on the prices of either the PO or the EC. Both PI and PE
are determined by the entry condition f E ðPI ; PE Þ ¼ 0 and the reaction function
PE ¼ PE ðPI Þ. PI is found by substituting the reaction function into the zero-profit
condition
f E ðPI , PE ðPI ÞÞ ¼ ðPE a mE ÞQE ðPI , PE ðPI ÞÞ FE ¼ 0
and solving for PI. The equilibrium prices are not functions of terminal dues
because t does not appear anywhere in the EC’s profit function or in the reaction
function.
Instead, small changes in the terminal duesh rate affectithe market by altering the
d f Iout d f Iout d f Iin
state probability of entry μ. At E3, μ ¼ dPI = dPI dPI . In order to have μ in the
interval 0 < μ < 1; the two derivatives in this formula must be of opposite sign.
d f Iout
Since dPI > 0 is positive for any price that is lower than PI for the equilibrium E1,
d f Iin
we must have dPI < 0.
4 The Bargaining Problem
Terminal dues and side payments are part of the solution to a cooperative game in
which the payoff to each PO is the added profit and side payments it derives from
outgoing and incoming international mail, including mail that originates with ECs.
The game’s strategies are parameterized by the bargained terminal dues rates.
The profit any single PO derives from its outgoing mail depends on the terminal
dues imposed by the countries to which the mail is sent. A net surplus (or deficit)
arises, first, because each PO retains a profit (or takes a loss) after paying terminal
dues on its outgoing mail, second, because the POs each make profits and losses on
the difference between their own terminal dues receipts and the cost of delivering
incoming international mail from other POs, and third, because some POs also gain
a profit from delivering mail that has originated with ECs. Terminal dues rates
affect these mail flows by affecting the equilibria in the outgoing international mail
markets.
We use a generalized version of the Nash bargaining model to characterize how
terminal dues rates and side payments would be structured.3 The solution to the
bargainingQ problem for profit-maximizing firms is to maximize
W ðπ Þ ¼ i ðπ i π i Þwi , where π i is a von Neumann-Morgenstern index of the i-th
player’s profit, and π i is his profit in the event there is no P
bargain. The parameters wi
are agreed proportions for a fair division; wi 0 8i and i wi ¼ 1. The uniqueness
of W(π) follows directly from two theorems for which the proofs may be found in
Pearsall (1965).
3
The generalizations extend the bargaining model to n-person cooperative games and permit the
players to choose asymmetric fair divisions. The model applies to profit maximizers because an
enterprise’s profit has the properties of a von Neumann-Morgenstern cardinal utility index. The
generalized Nash bargaining model is described at length in Pearsall (1965).
304 E.S. Pearsall
We can apply this framework to the system that has evolved for setting terminal
dues and side payments. There are three institutional systems that POs currently
follow to cooperatively set terminal dues rates and side payments.4
The Universal Postal Union (UPU)—The UPU administers a fallback system of
terminal dues and side payments for its 192 member countries. Members are
assigned to one of six categories graduated by size and development level. Terminal
dues rates for incoming mail differ by category and are also subject to moving caps
and floors that are mostly effective. Side payments that range from 2 to 20 % of the
terminal dues receipts from the larger developed countries are part of the system.
These take the form of payments to “transitional” countries and receipts from
“target” countries that pass through a UPU-administered “Quality of Service”
fund. The fund’s transactions are publicly reported. Side payments received by
small and under-developed countries from the fund form a significant part of the net
profit that the POs of these countries derive from international mail.
REIMS V—This is an agreement (the fifth) among 26 mostly European POs. The
system is administered by the Interpostal Corporation, and supercedes the UPU
rates and side payments among its participants. The precise terms are not public;
however, it is known that REIMS V sets terminal dues rates that are typically higher
than those of the UPU. REIMS V also includes nonpublic “quality of service
incentives with individual penalties and bonuses.”5 This is about the same euphe-
mism for side payments employed by the UPU.
Bilateral Agreements—Almost nothing specific is known about bilateral agree-
ments except that several large countries exchange international mail and pay
terminal dues at rates covered by such agreements, rather than at the rates stipulated
by the UPU or REIMS V. It is not known whether any of the currently existing
bilateral agreements specifically provide for side payments. However, bilateral
agreements often also arrange for transportation and processing services to be
performed by one PO for another. Such services, when performed below or above
cost, effectively become side payments in kind.
Consider now a cooperative game using the players’ monetary profits as surro-
gates for the von Neumann-Morgenstern utility indices. We parameterize the
players’ strategies by basing the players’ profits on a vector of terminal dues
rates. The method of solution dictated by the generalized Nash bargaining model
is to choose terminal dues and side payments to solve the constrained maximization
problem:
4
It is not necessary for the agreed rates and side payments to be the result of a single global
arrangement. The same result can be achieved constructively through a hierarchy of
interconnected multilateral and bilateral agreements.
5
This is one of four “principles” of REIMS cited by Copenhagen Economics (2014).
( )
X X
Maxt, s ln W ¼ wi lnðπ i π i si Þ π ¼ π ðtÞ, gðtÞ ¼ 0, si ¼ 0 :
i i
The players in the game are the POs of countries indexed i. They solve the
cooperative game by simultaneously agreeing on the selection of a vector of
terminal dues rates t and a vector of side payments s
The objective function of the maximization is the log transform of W ðÞ. The
POs’ profits are the elements of the vector π ¼ fπ i g: Profits are functions only of the
terminal dues rates, π ¼ π ðtÞ. π ¼ fπ i g denotes the POs’ profits if they fail to reach
an agreement.
The vector s ¼ fsi g is the net side payments for each PO. si is the sum of PO i’s
payments to all of the other POs minus its receipts from them. There are no side
payments if the POs fail to reach an agreement. Altogether, the net sidePpayments
among the POs must sum to zero but are not otherwise constrained, so i si ¼ 0.
The vector t ¼ ftk g is the collection of terminal dues rates that apply to mail from
an originating PO that is delivered by the PO of a second country. Any limitations
on the choice of the vector t are represented generally by the vector equation
gðtÞ ¼ 0.
Allowing side payments in a cooperative game is desirable from the point of
view of the players. When side payments are prohibited, this prohibition is equiv-
alent to adding the vector equation s ¼ 0 to the constraint set. When ln W is
maximized subject to the augmented set, its solution value decreases if any of the
added constraints are effective. A similar argument explains why the terminal dues
system permits rates that discriminate between the POs. Constraining the rates by
prohibiting discrimination among the POs is equivalent to adding many more
constraints of the form t j ¼ tk to the vector equation gðtÞ ¼ 0. Again, when any
of these added constraints are effective, they decrease the solution value of ln W.
The Nash solution to this cooperative game with side payments has two impor-
tant simplifying properties. First, the ratios of each PO’s gain from profits and side
payments to the total combined gain in profits for all POs are given by the elements
of the parameter vector w ¼ fwi g. It is this property of the solution that enables the
UPU and REIMS V to pool receipts and make side payments from pooled funds. It
also explains why the various agreements establishing rates and side payment rules
can be both simple and durable. Second, the solution vector of terminal dues, t,
maximizes the combined added profit of the players regardless of how the profit is
distributed. The vector t is determined independently of the vector w. This is the
Coase theorem on steroids. P
Both of these results are proven
P by differentiating the Lagrangian L ¼ i wi ln
ðπ i ðtÞ π i si Þ þ θgðtÞ þ ϑ i si with respect to the side payments, si, and the
elements, tk, of the vector of terminal dues rates. θ is a vector of Lagrangian
multipliersP for the constraints gðtÞ ¼ 0, and ϑ is a scalar multiplier for the single
constraint i si ¼ 0. Setting the derivatives to zero gives the first-order conditions
for a maximum of ln W subject to the constraints.
306 E.S. Pearsall
∂L wi
¼ þ ϑ ¼ 0 8i and
∂si ðπ i π i si Þ

∂L X wi ∂ui ∂g
¼ þθ ¼ 0 8k:
∂tk i
ð π i π i s i Þ ∂tk ∂tk
From set of conditions we have wP

Pthe first P i ¼ ϑðπ i π i Psi Þ 8i. Sum on i to
obtain i wi ¼ ϑ i ðπ i π i si Þ. Recall that i wi ¼ 1 and i si ¼ 0. Substitute
and solve for the multiplier ϑ ¼ P ðπ1 π Þ. Substitute back into the first set of
i i i
ðπ i π i si Þ
conditions to get P ðπ π Þ
¼ wi 8i. This proves the first result. The side payments
i i i
redistributed the total gain in profit in proportions according to the elements of w.

To show that the POs will maximize their combined profit, we substitute
h i from
P ∂π i ∂g
the first set of conditions into the second set to obtain ϑ i ∂tk þ θ ∂t ¼ 0 8k.
P ∂πi θ ∂g h i k
Divide through by ϑ to obtain i ∂tk þ ϑ ∂tk ¼ 0 8k. These turn out to be the first-
order conditions for the constrained maximization problem
P
Maxt i π i π ¼ π ð t Þ, g ð tÞ ¼ 0 . θ/ϑ is the vector of Lagrange multipliers for
gðtÞ ¼ 0. The Nash solution P for the vector of decision variables t is just to maximize
the total combined profit i π i subject to whatever additional limitations are
imposed by the constraint set gðtÞ ¼ 0. The distribution parameters w ¼ fwi g
play no part in this maximization.
5 International Mail Markets
The solution to the generalized Nash bargaining model has important implications
for the prices that bring equilibrium in the individual markets for international mail.
These markets were analyzed in Sect. 4 under the assumption that the terminal dues
rates are predetermined. We now know that when these rates are set cooperatively
the result maximizes the combined profit that the POs derive from all international
mail markets.
Excepting side payments, the total profit derived from a single international mail
market has three parts. These are the originating PO’s profit on the mail as it is
outgoing, the profit or loss of the PO delivering the mail as it is incoming, and,
sometimes, the delivering PO’s profit or loss on mail that is incoming from an EC.
Using the notation of Sect. 3, the originating
h PO’s expected profit oni outgoing
mail to the second country is μ ðPI t mI ÞQI ðPI ; PE Þin FI þ ð1 μÞ
ðPI t mI ÞQI ðPI Þout FI . The delivering PO’s expected profit or loss from
h i
delivering incoming mail from the originating PO is μ ðt cÞQI ðPI ; PE Þin þ
ð1 μÞ ðt cÞQI ðPI Þout . The delivering PO’s expected profit or loss from deliv-
ering incoming mail from the EC is: μ½ða cÞQE ðPI ; PE Þ.
When we sum these parts, the terminal dues paid by the originating PO to the
delivering
h PO disappear. The i resulting formula for the combined profit is:
μ ðPI c mI ÞQI ðPI ; PE Þin þ ð1 μÞ ðPI c mI ÞQI ðPI Þout FI þ μ½ða cÞ
QE ðPI ; PE Þ. With the appropriate probability of entry, this formula applies to all
three forms of the Nash equilibrium.
The effect of the terminal dues rate, t, on the combined profit of the originating
and delivering POs turns out to be entirely indirect. The rate t is just an instrument
that controls the two prices, PI and PE, and the probability of entry, μ, as we have
described in Sect. 3. Altogether, the total profit from international mail is the sum of
the profits for every origin and destination ði, j : j 6¼ iÞ pair of countries. Omitting
subscripts except for the marginal cost of delivery cj and the access charge rate aj
we have:
X X X h h i
πi ¼ μ PI c j mI QI ðPI ; PE Þin
j6¼i
i i

þ ð1 μÞ PI c j mI QI ðPI Þout

FI þ μ a j c j QE ðPI ; PE Þ i, j
Total profit is a separable function because each terminal dues rate applies to only
one combination of outgoing (i) and incoming ( j) countries. That is, each origin
and destination pair of countries has its own terminal dues rate tij which is a separate
element of the vector of control variables t.
Aside from the constraints gðtÞ ¼ 0, there is nothing about the cooperative game
to prevent the POs from setting terminal dues market-by-market to maximize the
profit derived from each individual market. Let us assume that there are no
restrictions gðtÞ ¼ 0 on the POs’ freedom to select terminal dues to maximize
their combined profit and proceed case by case through the three forms of equilib-
rium to see how the price of international mail will be set.
E1: The EC is always out of the market (μ ¼ 0). The combinedprofit derived
from the international mail from country i to country j is PI cj mI
QI ðPI Þout FI . The first-order condition with respect to the terminal dues rate is
found by differentiating with respect to tij and setting the result equal to zero:

dð Þ ∂QIout dPI
¼ QI ðPI Þout þ PI cj mi ¼ 0:
dti j ∂PI dti j
From Sect. 3:
∂ f Iout ∂QIout ∂QIout dPI

¼ QI ðPI Þout þ PI ti j mI ¼ 0, 0 and ffi 1=2.
∂PI ∂PI ∂PI dti j
308 E.S. Pearsall
Substituting in the first-order condition, we get:
dð Þ ∂QIout dPI
¼ ti j cj ¼ 0:
dti j ∂PI dti j
The first-order condition for a maximum of the combined profit is met when ti j ¼ cj .
Therefore, when the market is a monopoly for the originating PO, the solution to the
cooperative game is to set the terminal dues rate equal to marginal cost.
The intuition for this result is that the PO can increase the combined profit by
raising PI when ti j < cj and by lowering PI when ti j > cj . In order to encourage the
PO to make changes in PI to maximize the combined profit on international mail,
we make changes in the terminal dues rate tij knowing that about half the change
will be reflected in PI. Changing tij to increase the combined profit only becomes
impossible when ti j ¼ cj .
E2: The EC is always in the market (μ ¼ 1). The combined profit derived from
the international mail going from country i to country j becomes

PI cj mI QI ðPI ; PE Þin FI þ a j c j QE ðPI ; PE Þ:
Recall that the EC’s price is determined by the reaction function PE(PI). The first-
order condition for a maximum with respect to the terminal dues rate is:
in
dðÞ ∂QI ∂QIin dPE
¼ QI ðPI ; PE Þin þ PI c j mi þ
dti j ∂PI ∂PE dPI

∂QE ∂QE dPE dPI
þ a j cj þ ¼ 0:
∂PI ∂PE dPI dti j
From Sect. 3,
in
∂ f Iin ∂QI ∂QIin dPE
¼ QI ðPI ; PE Þin þ PI ti j mI þ
∂PI ∂PI ∂PE dPI
∂QIin ∂QIin dPE dPI
¼ 0, þ 0 and ffi 1=2:
∂PI ∂PE dPI dti j
Substituting in the first-order condition, we get:

in
dð Þ ∂QI ∂QIin dPE ∂QE ∂QE dPE dPI
¼ ti j cj þ þ aj cj þ ¼ 0:
dti j ∂PI ∂PE dPI ∂PI ∂PE dPI dti j
We consider first the case when a j ¼ ti j , i.e., price discrimination against the EC
is prevented by equating the delivery access rate to the terminal dues rate. The first-
order condition simplifies to become:
in
dðÞ ∂QI ∂QIin dPE ∂QE ∂QE dPE dPI
¼ ti j cj þ þ þ ¼ 0:
dti j ∂PI ∂PE dPI ∂PI ∂PE dPI dti j
The solution to the cooperative game is again to set terminal dues equal to marginal
cost, ti j ¼ cj . The intuition runs parallel to E1.
More commonly, however, aj cj > 0 as the delivering PO seeks to recover
some of its non-variable costs from the EC. From Sect. 3 we obtain the
approximation

dPE ∂QE ∂QE
ffi = 2 :
dPI ∂PI ∂PE
∂QE ∂QE
dPI , we find that ∂PI þ ∂PE
Substituting for dP E dPE
dPI ffi ∂QE
∂PI
=2. Insert this in the first-order
condition:
in
dð Þ ∂QI ∂QIin dPE ∂QE dPI
ffi ti j cj þ þ aj cj =2 ¼ 0:
dti j ∂PI ∂PE dPI ∂PI dti j
∂Q in ∂Q in ∂QE
We have ∂PII þ ∂PIE dP dPI 0, aj cj > 0 and ∂PI 0 for substitute goods.
E
Therefore, ti j cj 0 and the terminal dues rate tij is set at or above the marginal
cost of delivery cj. When the market is a duopoly with sequential pricing, the
terminal dues rate may rise above the marginal cost of delivery but cannot fall

below it. This result is due entirely to the presence of the term aj cj ∂Q E
∂PI
=2 in the
∂QE ∂QIin
∂QIin dPE
first-order condition. When aj cj ∂PI =2 > 0, we must have ∂PI þ ∂PE dPI < 0
and ti j cj > 0 to satisfy the first-order condition.
The intuition for this result is an extension of the intuition for E1. We can change

PI indirectly by changing tij because dt dPI
ij
ffi 1=2. However, if aj cj ∂Q E
∂PI
=2 > 0; the
dividing line between increasing and decreasing PI to improve the combined profit
moves up. If we evaluate the derivative ddtðiÞj at ti j ¼ cj , we get:

dð Þ ∂QE dPI
¼ aj cj =2 > 0:
dti j ∂PI dti j
When the terminal dues rate is set equal to marginal cost, it remains possible to
increase the combined profit by raising tij.
E3: The originating PO’s price leaves the EC with a zero profit if it is either in or out
of the market. With E3 we have 0 < μ < 1. The combined h profit expected from the
international mail from country i to country j is E½ ¼ μ PI cj mI QI ðPI ; PE Þin þ

aj cj QE ðPI ; PE Þ þ ð1 μÞ PI cj mI QI ðPI Þout FI . Equilibrium requires
that the terminal dues rate be set to maximize expected profit against the EC’s optimal
310 E.S. Pearsall
choice of a probability of entry μ. The math for E3 parallels that for E2. dðdtEi½jÞ is derived
with μ held fixed to obtain the first-order condition for a maximum of the combined
expected profit. Substitutions are made from Sect. 3 and we eventually obtain:
in
dðE½Þ ∂Qi ∂Qiin dPE ∂Qiout
¼ ti j c j μ þ þ ð1 μÞ
dti j ∂PI ∂PE dPI ∂PI

∂QE dPI
þ μ aj cj =2 ¼ 0:
∂PI dti j
The E3 solution is a weighted average of the E1 and E2 solutions with ð1 μÞ

and μ as the respective weights. When a j ¼ ti j, the terminal dues rate is set equal to
marginal cost, but when a j > c j we obtain ti j c j 0. And again, tij is set above cj

if a j cj ∂Q E
∂PI
=2 > 0.
The E3 solution differs from the E2 solution only with respect to the strength of
the effect that takes the terminal dues rate above marginal cost. This effect depends
upon the probability of entry. As μ approaches zero, the effect diminishes and tij
approaches cj.
The model predicts a general pattern for terminal dues rates that parallels what
we actually observe. Campbell et al. (2012), Campbell (2014) and Copenhagen
Economics (2014) have documented the widespread discrimination found in cur-
rent terminal dues rates. The pattern heavily favors incoming mail from other POs.
The terminal dues rates for this mail are universally lower than the access rates that
the POs charge domestic mailers and/or potential ECs for comparable delivery
services.
When we analyze the solution to the Nash bargaining model in the absence of
competition, we find that terminal dues rates gravitate towards the marginal cost of
delivery, cj, rather than towards the access rate, aj. This explains why terminal dues
are so much lower than the commercial access rates set by most delivering POs. The
access rates ordinarily include a substantial component to offset delivery costs that
are not variable at the margin. For example, this happens when access rates are set
efficiently to maximize welfare subject to a zero-profit constraint. The resulting
access rate is a Boiteaux-Ramsey price that will always exceed the marginal cost of
delivery. Marginal cost is lower than average cost because mail delivery is an
activity with an average cost function that decreases over the relevant range of
postal volumes. It is not terribly imprecise to say that, where there is no competition
for international mail, terminal dues rates are based on marginal cost rather than
average cost.
Rates set at marginal cost are nondiscriminatory among the POs themselves. The
E1 case is most applicable to countries that are small and/or underdeveloped. In
practice, the UPU simply lumps these countries together for the purpose of
assigning terminal dues rates. The UPU rates among these countries are not
particularly discriminatory.
However, this picture changes when potential competitors are present in an

international mail market. When entry always happens (E2) or happens just occa-
sionally (E3) the model predicts that the terminal dues system may discriminate
among the POs themselves. Here, the general pattern is based upon the threat of
entry. When an international mail market is large enough to be attractive to an EC,
the solution predicts that the terminal dues rate will rise above marginal cost
depending upon the demand conditions confronted by the EC. For both E2 and
E3, the relationship is similar. For substitute services, terminal dues are greater-
than-or-equal to marginal cost when access rates exceed marginal cost. The strength
of this effect is directly linked to the threat of entry as measured by the probability
of entry, μ. When μ is close to zero, the solution is close to that of a PO monopoly
(E1); when μ is close to one the solution differs little from a duopoly with sequential
pricing (E2); and when μ is between zero and one the mixed strategy with limit
pricing (E3) splits the difference.
Generally, current terminal dues rates also vary this way. The default rates set by
the UPU progress through six levels of increasing size and development. The lower
levels include far more POs than the higher levels, so the UPU rates mostly
discriminate very broadly between countries that are large and developed and
those that are either not large or not developed. However, the larger and more
developed the market for international mail, the more likely it is that the UPU
default rate will be superseded by a rate from REIMS V or a bilateral agreement.
These rates are generally higher than the corresponding UPU rates. In any case, the
rates discriminate much more between the largest and most developed POs than the
others.
The model lets us discover the outcome of eliminating price discrimination in
the provision of delivery services to ECs. Such price discrimination violates
competition law for the European Union’s countries and may be illegal elsewhere
as well. When an access rate for an EC is set equal to the corresponding terminal
dues rate for an originating PO, both the terminal dues rate and the access rate
gravitate together towards the marginal cost of delivery.
6 Conclusion
The model presented in this chapter unites two well-known contributions to game
theory attributable to John Nash. The model explains reasonably well the workings
of the current terminal dues system in-so-far as these workings are publicly known.
The model predicts that international mail markets will fall into one of three
categories of market organization: monopoly, duopoly with sequential pricing
and mixed monopoly/duopoly with limit pricing. The categories are distinguished
by differing probabilities of entry by an EC. Reducing terminal dues rates discour-
ages the entry of ECs and increases the profits of an originating PO. The POs have a
collective incentive to set terminal dues rates cooperatively. When the rates are set
cooperatively, they will be accompanied by side payments and may be
312 E.S. Pearsall
discriminatory. The total added profits derived by the POs from setting rates
cooperatively will be fairly divided according to pre-agreed proportions. The
terminal dues system will maximize the aggregate added profit of all of the POs
taken together. The resultant terminal dues rates will be set at marginal cost in the
absence of competition or if there is no price discrimination directed against the
ECs. And, the rates will be set at or above marginal cost in markets where actual or
potential competitors are present and confront access rates that exceed marginal
costs.
References
Campbell JI Jr, Dieke AK, Zauner M (2012) UPU terminal dues: winners and losers’. In: Crew
MA, Kleindorfer PR (eds) Multi-modal competition and the future of mail. Edward Elgar,
Cheltenham, pp 278–289
Campbell JI Jr (2014) Estimating the effects of UPU terminal dues, 2014–2017. Paper presented at
the 8th Bi-annual Postal Economics Conference, Toulouse, France, 3–4 Apr 2014
Copenhagen Economics (2014) The economics of terminal dues. Final Report, U.S. Postal Reg-
ulatory Commission, 30 Sept 2014
Pearsall ES (1965) Social welfare functions—an axiomatic approach. Zeitschrift fur
Nationalokonomie XXV/3–4: 278–286
Pearsall ES (2011) On equilibrium in a liberalized market. Paper presented at the CRRI advanced
workshop on regulation and competition, 30th annual eastern conference, Skytop, PA, USA,
18–20 May 2011
Quantifying the Distortive Effects of UPU
Terminal Dues
James I. Campbell, Jr.
1 Introduction
The terminal dues system of the Universal Postal Union (UPU) is a signal failure in
modern international trade policy. Over the last three decades, rising competition
and regulatory reform have rendered domestic postal services more cost based and
efficient. Not so internationally. More than a quarter century after the above
observation by the U.S. Department of Justice, the UPU’s terminal dues system
continues to produce large distortions in the exchange of international postal
delivery services, distortions which also skew the larger market for international
delivery services. The root of the problem is that postal operators (POs)1 do not
charge each other domestic postage for delivery of inbound international docu-
ments and small packages. They charge “terminal dues” fixed by the UPU to serve
political and anticompetitive ends.
As the U.S. Department of Justice, the European Commission and other
observers have recognized, economic efficiency and competitive fairness dictate
Independent lawyer and consultant whose clients in matters relating to UPU issues have included
the major private express operators and the European Commission; views expressed in this chapter
are the personal views of the author.The current terminal dues structure produces distortions in the
economic structure of the international mail system. Since terminal dues do not accurately reflect
costs, the current system causes a subsidy to flow from some parties to others. . . and generally
impairs the efficient operation of the international mail system.—U.S. Department of Justice (1988)
1
In 2008, UPU terminology substituted the term “designated operator” for “postal administration”
due to widespread corporatization and privatization of government POs. In almost all countries,
however, the only “designated operator” remains the postal administration or its corporatized
successor. For simplicity, this chapter uses “postal operator” (PO) instead of “designated
operator.”
J.I. Campbell, Jr. (*)
Independent Lawyer and Consultant, Potomac, MD, USA
e-mail: jcampbell@jcampbell.com

314 J.I. Campbell, Jr.
that the terminal dues charged for delivery of a shipment of inbound international
mail should be equivalent to the domestic postage rates that the destination PO
would charge for delivery of a similar shipment of domestic mail (U.S. Department
of Justice 1988, pp. 25–26; European Commission 1992, pp. 220–21; Copenhagen
Economics 2014, pp. 13–14). What is the “equivalent domestic postage”? Terminal
dues compensate only for local delivery of inbound international mail since inter-
national mail is collected by the origin PO and long distance transportation in the
destination country, if any, is compensated by other UPU fees.2 On the other hand,
inbound international mail is typically more differentiated and less well addressed
that domestic bulk mail and may require additional handling for customs or other
import formalities. In principle, therefore, the domestic postage that is equivalent to
terminal dues is likely between retail domestic postage for “priority” or “first class”
mail3 and domestic postage for bulk mail. The UPU uses 70 % of priority domestic
postage as the appropriate standard for calculating the domestic postage equivalent
to terminal dues, and this chapter generally adopts this standard.4
In general, however, the UPU sets terminal dues substantially below equivalent
domestic postage. Terminal dues thus favor international mailers over domestic
mailers. Competition between POs and private carriers is restrained since terminal
dues are available only to POs. Competition between POs is also limited because POs
charge different terminal dues rates to different POs and discriminate against mail that
is not posted by a mailer physically located in the national territory of the originating
PO. As the effects of terminal dues vary by country, large economic transfers occur
between POs, primarily to the benefit of POs from a handful of large industrialized
countries and Asian POs heavily engaged in international e-commerce.
Despite manifest flaws in the UPU terminal dues system, there have been few
attempts to quantify the magnitude of resulting distortions. This chapter describes a
methodology for developing a “terminal dues model” that estimates the economic
distortions implied by the current UPU agreement on terminal dues, in effect from
2014 through 2017. This remainder of the chapter is divided into four sections.
Section 2 describes the UPU and its terminal dues system. The third section pro-
poses an approach for developing a terminal dues model derived from UPU studies
and statistics. The Sect. 4 summarizes illustrative findings from the terminal dues
model. The final section summarizes conclusions.
Two caveats are in order at the outset. First, the chapter estimates only distortions
implied by the 2012 UPU terminal dues system. These estimates do not take account
of alternative bilateral arrangements agreed by some POs. Such side agreements
2
2012 UPU Convention, Art. 33 (transit charges) and Art. 34 (air conveyance dues).
3
The UPU Convention refers to “charges corresponding to priority items in the domestic service”
as the proper basis for terminal dues. 2012 UPU Convention, Art. 30(1). in this paper, “priority
postage rates” are the same as postage for “first class mail” (in the US and UK) and for “the fastest
standard category” of mail (in the EU Postal Directive).
4
2012 UPU Convention, Art. 30(5). In 2013, a limited UPU survey indicated that the cost of
delivering inbound international letter mail is between 60 and 120 % of the equivalent cost for
domestic mail (UPU 2013b).
Quantifying the Distortive Effects of UPU Terminal Dues 315
cannot justify a distortive UPU system in the first place and, in any case, are
constrained by the default rates established by the UPU. Second, all estimates are
highly approximate. As explained below, the more specific the mail flows, the more
they should be interpreted as illustrative rather than actual quantities.
2 Universal Postal Union and the Terminal Dues System
The Universal Postal Union (UPU) is intergovernmental organization with

192 member countries. National delegations meet in a general Congress every
4 years to amend or reenact the agreements governing international postal services.
The most recent Congress was held in Doha in 2012; it adopted the current terminal
dues provisions. The next Congress convenes in Istanbul in September 2016.
Terminal dues apply to delivery of “letter post” items.5 Despite its name, the
“letter post” is not limited to what are normally called “letters.” The letter post
conveys all types of documents and packages up to 2 kg, which the UPU classifies
according to three “shapes”: ordinary small letters, large envelopes (or “flats”), and
small packets, which are packages weighing up to 2 kg. In this chapter, these shapes
will be referred to simply as letters, flats, and small packets.
It should be emphasized that small packets—packages up to 2 kg—are espe-
cially important for the future of the international post and, indeed, of international
commerce. Small packets are the primary postal conduit for e-commerce. With
paper mail declining rapidly (more rapidly than in domestic services), e-commerce
is widely seen as the best hope for sustaining international postal services. Growth
in e-commerce traffic from Asia has been especially significant. In the last few
years, many industrialized countries have seen increases of 100 % or more each
year in e-commerce items received from POs in China, Hong Kong, and Singapore,
partly because these POs consolidate shipments from other Asian countries. Com-
petition in the e-commerce market is also heating up. Commercially minded POs
are extending operations outside their national territories, and global express
companies are actively developing services for e-commerce.
The current terminal dues system establishes three schedules of terminal dues.
These schedules apply to six UPU-defined “groups” of countries distinguished by
their level of economic development. One schedule applies to Group 1.1, the
24 major (and 4 minor) industrialized countries. The second schedule applies to
Groups 1.2 and 2, collectively 24 advanced developing countries. The third schedule
applies to Groups 3, 4, and 5, the less developed countries. For simplicity, this chapter
refers to these collections of countries as Groups 110, 122, and 345, respectively.
5
The UPU also fixes delivery rates, called “inward land rates,” for “parcel post” items, i.e.,
packages weighing up to 20 kg. Most postal packages weighing less than 2 kg are conveyed by
letter post rather than parcel post since terminal dues are generally less than inward land rates. For
a discussion of inward land rates, see the chapter by Ralf Wojtek.
The UPU also classifies terminal dues groups according to two larger categories,
the “target system” and “transitional system.” These terms refer to the UPU’s
commitment, adopted in the 1999 Congress and reaffirmed at each subsequent
Congress, to implement a worldwide “country-specific payment system,” that is,
a system in which terminal dues are related to domestic postage rates in each
country. Since 1999, however, the UPU has made no progress towards this objec-
tive. In 2014, the target system included two terminal dues schedules, for Groups
110 and 122, while the transitional system referred to the schedule for Group 345.
In 2016, UPU Group 3 (“Group 300”) will move from the transitional system to the
target system but keep the same terminal dues schedule, so after 2016 the “target
system” will include all three terminal dues schedules, none of which is “country
specific.”
Rates for the three terminal dues schedules are shown in Table 1.6 In the target
system, a rate formula purports to align terminal dues to 70 % of retail dominant
postage rates. In almost all target system countries, however, the target rate formula
is limited by a cap or floor, so that the formula is not actually used.
3 Developing a Global Terminal Dues Model
Terminal dues would cause no distortions between domestic and international mail
from POs if, in each destination country, terminal dues equaled equivalent domestic
postage. The financial effects of the distortions implied by the UPU terminal dues
system may be estimated with a spreadsheet, a “terminal dues model,” that calcu-
lates the postal volumes, weights, terminal dues charges, and equivalent domestic
postage charges for the flow of letter post items, in each direction, between all UPU
countries and/or groups countries. The terminal dues model proposed in this chapter
is based on five preparatory spreadsheets or modules relating to volumes, flows,
mail structure, terminal dues, and domestic postage.
The volumes module develops outbound and inbound letter post volumes for
2011 for all UPU member countries for which estimates can be obtained. The
starting point is the UPU database of postal statistics, even though incomplete and
often unreliable. Statistics for 2011 include outbound volumes for about 139 of
220 countries and territories. These 139 countries accounted for 1.57 billion letter
post items, less than half of the UPU’s estimated global total of four billion items
(UPU 2013a, p. 5). By relying on data for the year before, the year after, or, when
necessary, 2 years before, estimates for another 31 countries can be included,
adding 0.41 billion. What is missing are volumes from a small number of large,
highly commercialized countries. By adding data from national regulators (five
countries, 1.11 billion items) and estimating international volumes based on a
6
Terminal dues are expressed in Special Drawing Rights. On May 14, 2015, 1 SDR was equal to
USD 1.4199 and EUR 1.1425.
Table 1 Terminal dues schedules, 2014–2017

% volume Annual
Schedule 2014 (%) Formulae (SDR) increase (%)
Intra-110 67 Target rate formula constrained by: 1.9
– Cap: 2.294/kg + 0.294/item; maximum 3.0
annual increase 13 %
– Floor: 1.591/kg + 0.203/item 2.8
Intra-122 and 11 Target rate formula constrained by: 1.9
to/from 110 – Cap: 1.641/kg + 0.209/item 6.0
– Floor: 1.591/kg + 0.203/item 2.8
To/from/ 22 4.192/kg (<75 tonnes) 2.8
between 345 Revision mechanism (optional, 2.8
> 75 tonnes): 0.591/kg + 0.203/item
Source: UPU, 2012 Convention, Arts. 29–31; 2013 Letter Post Regulations, Art. RL
220 (target rate formula)
plausible ratios to domestic mail volumes (seven countries, 0.79 billion), the result
is a global estimate of 3.85 billion outbound letter post items in 2011 (182 coun-
tries). This agrees reasonably well with the UPU estimate. A similar process was
used to develop estimates of inbound volumes by country.
The mail structure module summarizes information about the structure of the
international letter post in 2014. This information is used in the terminal dues model
to calculate terminal dues and equivalent domestic postage charges. The source is
the UPU’s 2014 IPK Study (UPU 2014b) (“IPK” refers to “items per kilogram”).
This study estimates the average weight and distribution of letter post mail by shape
for exchanges between Groups 110, 122, 300, and 450. The study reports, for
example, that for mail sent from Group 110 to Group 300 in 2014, the average
weights of letters, flats, and small packets, were 14.79, 154.32, 308.64 g, respec-
tively, and these shapes accounted for 59.3 %, 11.9 %, and 28.8 % of the volume of
letter post, respectively. The study is based on inbound letter post reports from
49 post offices. The 2014 IPK Study also provides a detailed profile of international
letter post by shape and weight step, but it is based on reports from only 15 posts.
The flows module is the heart of the terminal dues model. It develops estimates of
bilateral letter post flows between UPU countries and groups of countries in 2014.
Bilateral flows are not publicly available. Recent studies by the UPU, however,
estimate flows from or between geographic regions. These studies can be pieced
together with plausible assumptions to allow calculation of approximate bilateral
flows that appear adequate for the aims of a terminal dues model.
A key UPU study estimates the relative weights of letter post flows between
world regions in 2011 (Anson and Helble 2014). The “2011 Regional Weight
Study” encompasses six world regions: Asia; Eastern Europe (and Central Asia);
Middle East (and North Africa); North America; South America (and Central
America and Caribbean); Africa (sub-Saharan); and Western Europe. The study
reports, for example, that in 2011, if the weight of mail exported North Africa to
Western Europe was 1, the weight of mail from Western Europe to North America
was 55. These relative weights can converted into relative volumes using a 2010
IPK study (UPU 2011, Annex 4).
In the flows module, interregional flows are divided into bilateral country flows
using an assumption of “proportional regional participation.” For example, sup-
pose that, based on the 2011 Regional Weight Study, Asia sent to Western Europe
12.24 % of global letter post and that, based on the volumes module, Australia
accounts for 10.4 % of all mail exported from Asia and France accounts for 18.4 %
of all letter post imported by Western Europe. Then, the flow from Australia to
France may be estimated as 0.234 % of global mail. More precisely, the flow from
Country i in Region A to Country j in Region B, Fij, can be derived from the
interregional flow as follows:
Oi Ij
Fij ¼ FAB
OA I B
where FAB is the flow from Region A to Region B; Oi and OA are, respectively, the
total outbound volumes from Country i and Region A; and Ij and IB are, respec-
tively, the total inbound volumes to Country j and Region B. If Country i and
Country j are in the same Region, A, and FAA is the flow between countries in
Region A, the equation is:
Oi Ij
Fij ¼ FAA
OA I A ð1 ðI i =I A ÞÞ
The result is a complete set of relative bilateral international mail flows that is
internally consistent and consistent with available UPU volume data.7
The assumption of proportional regional participation is obviously a rough and
ready approximation. It relies only on total volumes of outbound and inbound mail
sent and received by Countries i and j. In reality, the volume of mail between two
countries is also affected by such factors as distance, historical relations, language,
and currency.8 However, the assumption of proportional regional participation is
substantially constrained by the UPU’s regional flow studies. In the flows module,
the estimated volume of mail that Country i sends to Country j is based first on the
overall interregional volume and only secondarily on the assumption of the pro-
portional regional participation. This constraint limits potential errors implicit in
the assumption of proportional regional participation.
7
The assumption of proportional regional participation is also used to subdivide regional flows
from Asia in the 2011 Regional Weight Study into flows associated with Groups 110, 122, and
345 before using the 2010 IPK study to convert weights into volumes. In the North American
region, mail flows between the United States and Canada have been adjusted to reflect 2011 data
from the U.S. Postal Regulatory Commission.
8
Anson and Helble 2013. Unfortunately, the authors’ “gravity model” depends on at least one
variable that requires access to nonpublic UPU data.
In October 2014, the UPU published an annual survey of the previous year’s
global and regional mail volumes called “Development of Postal Services” (DPS).
It estimates the volume of international letter post in 2013 was 3.5 billion items,
5.6 % less than in 2012 (UPU 2014a, p. 4). The 2013 DPS survey also provides a
breakdown for outbound letter post by geographic region.
In the flows module, the total volume of mail for 2014 is assumed to be 3.5
billion items, the same as 2013. The outbound volumes per geographic region are
taken from 2013 DPS. The allocation of outbound regional volumes among the
other regions is based on the 2011 Regional Weight Study. Division of bilateral
regional flows into bilateral country flows (using the principle of regional propor-
tionality) relies on estimates from the volumes module.
Bilateral country flows are then collected into “regional/TD Groups” defined by
geographic region and terminal dues group. Groups 122 and 300 have been
subdivided to allow separate estimates for POs heavily engaged in e-commerce
(China, Hong Kong, and Singapore). Not every geographic region includes all
terminal dues groups. In total, there are 19 regional/TD Groups, three of which
are composed solely of industrialized countries in regions Asia, North America, and
Western Europe.
The terminal dues module is slightly revised version of the UPU’s 2012 “termi-
nal dues tool” (UPU 2012a). The “tool” is a UPU spreadsheet that allowed POs to
forecast terminal dues they could expect to pay and be paid in 2014–2017.
The domestic postage module synthesizes available information on domestic
postage rates from several sources and estimates the 2014 priority domestic postage
rates by shape and weight step. For industrialized countries and other European
countries, the primary source is survey of 2013 priority domestic rates by WIK. For
other countries, the original source is a UPU database of 2008 priority domestic
rates all countries, by shape and weight step. Rates for countries in Group 122 have
been updated to 2014 using selected domestic postage rates reported annually by
the UPU. Rates for countries in Group 345 have been updated to 2014 by assuming
rate changes in line with inflation.
Output tables from these modules are brought together in the terminal dues
model. The model estimates bilateral volumes, weights, terminal dues charges, and
equivalent domestic postage charges in 2014 for 41 origins and destinations:
25 industrialized countries9 and 16 regional/TD Groups of non-industrialized
countries. Separate estimates are calculated for letters, flats, and small packets.
Estimates for the years 2015, 2016, and 2017 reflect user inputs defining anticipated
growth rates, shifts in average weight per item, and changes in domestic postage
rates.
9
Minor industrialized countries and territories are consolidated into a composite country, “Europe
Minor.”
4 Implications of the Terminal Dues Model
Implications of the terminal dues model can be illustrated by adopting a “base

scenario” or baseline set of assumptions about development of the letter post after
2014. For purposes of this chapter, the base scenario assumes continuation of
current trends in the years 2015–2017, resolving doubts in favor of the status quo.
Volumes of letters and flats sent between industrialized country POs (ICs) are
expected to decline annually by 5 % and 3 %, respectively, while small packets
are expected to grow at 15 % per year. Volumes of flats and small packets sent by
the major e-commerce POs (ECs)—China, Hong Kong, and Singapore—to the ICs
are expected grow by 20 % and 50 % per year, respectively. Similar but more muted
trends are expected for shipments from ICs and ECs to ECs and developing country
POs (DCs).
Under the base scenario, the composition and global distribution of letter post
will change significantly from 2014 to 2017. Letters will decline from 59 % of the
letter post to 49 %, while small packets increase from 25 to 37 %. The volume of
letter post will increase modestly from 3.5 billion to 3.8 billion items. The main
engine for this growth is the increasing e-commerce traffic from the Asia, which
will increase its share of global postal letter post from 9 to 17 %.
4.1 Terminal Dues vs. Equivalent Domestic Postage
The primary cause of economic and competitive distortions arising from the current
terminal dues system is the failure of the ICs to charge terminal dues equal to
equivalent domestic postage. In 2014, about 87 % of all international mail was
destined for the ICs: 68 % from other ICs, 6.6 % from ECs, and 12.6 % from DCs.
Virtually, all of this mail is subject to the target system rate cap or the per kilogram
rate available to DCs.
Table 2 compares the terminal dues charges and equivalent domestic postage
rates for delivery of a kilogram of typical letter post items received by ICs in 2014.
For example, under the UPU terminal dues system, the Austrian PO charges SDR
5.42 for a typical kilogram of letter post mail sent by an IC, whereas the equivalent
domestic postage (70 % of normal priority rates) would imply a charge of SDR
9.03. The “terminal dues discount” is 40 %. Overall, ICs give each other an average
discount of about 46 % from equivalent domestic postage. The discount for small
packets is about 56 %. ICs give larger discounts for delivery of letter post items
received from ECs and DCs, in both cases averaging about 75 %. These discrep-
ancies between terminal dues and equivalent domestic postage are not rough
approximations. They are derived from the 2012 UPU Convention and public
tariffs. Nor will these discrepancies decline significantly over the course of the
current terminal dues system.
Table 2 Volumes and terminal dues discounts, 2014

TD per TD TD TD
LP vol LP kg for EDP per discount discount discount
out LP vol ICs LP kg for for ICs for ECs for DCs
Destination (mil) in (mil) (SDR) ICs (SDR) (%) (%) (%)
Austria 113.4 158.1 5.42 9.03 40 83 82
Australia 56.6 156.7 5.20 7.31 29 79 77
Belgium 70.1 82.9 5.42 10.51 48 84 83
Canada 85.4 248.1 4.67 14.87 69 92 90
Switzerland 107.1 177.8 5.42 8.97 40 74 73
Germany 361.1 280.2 5.41 7.73 30 76 75
Denmark 44.0 50.7 5.42 14.04 61 82 81
Spain 83.1 61.6 5.45 5.93 8 59 57
Finland 15.5 27.6 5.42 15.09 64 89 89
France 190.4 357.3 5.41 9.24 41 73 72
Great Britain 444.1 183.2 5.42 9.78 45 83 82
Greece 36.3 18.3 5.42 7.94 32 70 69
Ireland 32.5 68.8 5.42 14.04 61 89 88
Israel 12.6 18.6 5.43 6.73 19 60 59
Iceland 1.4 1.8 5.42 6.63 18 61 59
Italy 70.2 128.8 5.42 13.13 59 83 82
Japan 36.5 171.4 5.08 7.70 34 69 67
Luxembourg 30.2 18.3 5.42 6.94 22 74 73
Netherlands 137.9 104.2 5.42 8.91 39 72 71
Norway 18.1 103.2 5.42 20.53 74 92 92
New Zealand 16.1 34.7 5.15 9.11 43 77 75
Portugal 30.7 24.0 5.42 6.62 18 62 61
Sweden 41.8 57.0 5.42 11.65 53 79 78
United States 505.1 457.2 5.49 7.05 22 79 76
Europe 31.3 42.8 5.43 10.96 50 85 84
Minor
Totals/ 2571.4 3033.4 5.36 9.98 40 77 75
averages
Key: LP letter post, TD terminal dues, EDP equivalent domestic postage (70 % of priority
postage), SDR special drawing rights, ICs industrialized countries, ECs e-commerce countries,
DCs developing countries
In light of such large rate differences, it is obvious that for any set of mail flows,
the total terminal dues charges levied by ICs will be significantly less than equiv-
alent domestic postage. How much will depend on volumes and structures of
specific flows? The cumulative economic distortions produced by discrepancies
between terminal dues and equivalent domestic postage could be analyzed quanti-
tatively using hypothetical flow data. The terminal dues model, however, aims to
produce a more realistic picture of the distortions resulting by the UPU terminal
dues system by estimating bilateral mail flows from UPU data. While the estimates
are necessarily imperfect and incomplete, the model provides a reasonably clear
picture of the effects of differences between terminal dues and equivalent domestic
postage on the global postal market.
4.2 First: Total Undercharges
Whether using hypothetical or estimated flow data, analysis of cumulative effects

suggests that distortions created by the UPU terminal dues system can be summed
up by two measures: (1) the total amount of undercharges implied by the terminal
dues system and (2) the net transfer of economic value between POs. The total
amount of undercharges provides a rough indication of the total potential compet-
itive distortion implied by the system of terminal dues.10 The net transfer of
economic value between POs measures the extent to which some POs actually
benefit at the expense of others. Total undercharges equal the sum of net transfers
and revenues from other sources, primarily higher rates on other postal items. In
other words, when a PO undercharges for delivery of inbound international mail, it
must make up the shortfall by a net advantage in the exchange of letter post with
other POs and/or by charging higher rates to other mailers (or by getting money
from other sources).
Undercharges implied by the UPU terminal dues system may be calculated by
multiplying the total weight of each bilateral flow by, first, the terminal dues charge
for a typical kilogram of letter post for that flow and, second, by the equivalent
domestic postage that would be charged for the same kilogram. The undercharge is
the latter minus the former.
An undercharge for delivery of a specific international postal item does not
necessarily result in a distortion in the price which the origin PO charges the mailer
for that item. If the price which the origin PO charges the mailer does not fully
reflect the artificially low “cost” represented by terminal dues, then the distortion
will be less than the full amount of the undercharge. The total undercharge
measures how much an origin PO could undercharge outbound customers for
international postal services and, similarly, the extent to which domestic mailers
could be disadvantaged compared to foreign mailers.11
Undercharging facilitated by the UPU terminal dues system is potentially anti-
competitive. In almost all industrialized countries, domestic rates are required to be
cost based (with some exceptions), i.e., rates must cover marginal costs plus an
appropriate share of common costs. If the origin PO sets its prices for international
mailers based on the terminal dues charges of the destination PO, the prices of the
10
The amount of total undercharges does not take into account factors such as own and cross-price
elasticities which would be needed to estimate overall effects on the market.
11
If undercharges are not passed on in lower rates for mailers, the distortive effects of
undercharging will appear elsewhere. The origin PO will book an extraordinary profit on the
outboard item that will, ceteris paribus, artificially lower rates for other mailers.
origin PO will not fully reflect what the destination country considers to be the
appropriate level of costs and what the domestic mailer must pay for similar
domestic services.12 Under such circumstances, competitors of POs would be
disadvantaged in the end-to-end market and in the upstream and downstream
portions of that market. Similarly, such prices would favor foreign mailers to the
disadvantage of domestic mailers in the destination country. If foreign mailers and
domestic mailers are competitors—e.g., competing e-commerce merchants—the
anticompetitive effects at the buyers’ level are apparent (Copenhagen Economics
2014, Chap. 2). While the distortive effects of undercharging are potential, it is
clear that the UPU’s cap on IC terminal dues is intended “to moderate the impact of
TD on international tariffs” (UPU 2012b, para. 37).13
The magnitude of undercharging relative to equivalent domestic postage by ICs
permitted by the UPU terminal dues in 2014 is summarized in Table 3. In 2014, the
total market for delivery of international letter post items was about SDR 3.0
billion, measured by the sum of equivalent domestic postage charges. ICs
undercharged for delivery services by about SDR 1.53 billion, roughly 53 % of
the value of the market. Most of the “terminal dues discount,” SDR 937 million,
went to other ICs, and most of this amount was due to discounts for delivery of
small packets (SDR 502 million). ICs undercharged for delivery of letter post items
received from ECs by about SDR 206 million and from DCs by about SDR
384 million, effectively discounts of 72 % for each.
Under the base scenario, the extent of undercharging will not improve over the
4-year course of the terminal dues agreement. Total undercharging will grow to 2.2
billion in 2017, 56 % of the market. For the 4-year period, total undercharges
amount to SDR 7.35 billion, two-thirds of which is due to undercharging for small
packets.
4.3 Second: Net Transfers Between POs
Although undercharging creates a loss on inbound mail relative to equivalent

charges for domestic mail, it is a benefit for outbound mail. The net effect on a
PO is the combination of undercharges (and occasional overpayments) in outbound
12
Terminal dues result in what amounts to false or at least inaccurate accounting because terminal
dues are essentially barter transactions at prices that fail to reflect fair market value. A discussion
of cost accounting and measures of cost is beyond the scope of this chapter. It should be noted,
however, that whether prices charged between undertakings cover “attributable costs” (or some
other measure of direct costs) is only one factor in determining whether a price agreement among
competitors is unlawfully discriminatory or anticompetitive.
13
See also USPS (2012): “Increasing terminal dues rates, especially significant increases resulting
from elimination of the cap or directly tying the rates to domestic mail rates, would result in a
considerable increase in the cost of delivery of letter post mail abroad. To offset this increase, the
Postal Service would be forced to cover its costs with rate increases to the mailers.”
Table 3 Undercharges by industrialized countries, letter post 2014

TD for EDP EDP EDP Under-
LP for LP Under- for LP Under- for LP charge
from from charge from charge from for
ICs ICs for ICs ECs for ECs DCs DCs Total
(mil (mil (mil (mil (mil (mil (mil under-
Destination SDR) SDR) SDR) SDR) SDR) SDR) SDR) charge
Austria 64.7 102.9 38.2 7.5 5.6 29.2 21.9 65.7
Australia 42.6 60.7 18.1 39.2 27.5 28.4 19.5 65.0
Belgium 34.0 63.0 29.0 4.1 3.2 16.1 12.3 44.5
Canada 94.9 264.8 169.9 54.7 48.1 65.0 57.0 275.0
Switzerland 72.8 115.0 42.3 5.3 3.3 21.2 13.0 58.5
Germany 111.8 152.3 40.6 10.7 7.0 41.8 27.3 74.8
Denmark 20.9 51.6 30.8 2.1 1.6 8.5 6.2 38.5
Spain 19.2 26.4 7.2 1.2 0.5 4.6 1.8 9.5
Finland 11.4 30.3 18.9 1.9 1.6 7.6 6.4 26.9
France 145.0 236.0 91.0 10.9 6.6 43.0 25.8 123.4
Great Britain 72.6 125.2 52.6 9.9 7.4 38.8 29.1 89.1
Greece 7.5 10.6 3.0 0.5 0.3 1.9 1.0 4.3
Ireland 28.3 70.1 41.8 4.7 4.0 18.5 15.4 61.2
Israel 5.8 9.1 3.3 0.4 0.1 1.4 0.6 4.0
Iceland 0.7 0.9 0.1 0.0 0.0 0.1 0.1 0.2
Italy 52.9 122.4 69.5 6.0 4.5 23.6 17.7 91.8
Japan 52.2 70.8 18.6 28.0 15.5 21.1 11.6 45.7
Luxembourg 7.5 9.2 1.7 0.5 0.3 2.1 1.3 3.3
Netherlands 42.6 66.9 24.3 3.0 1.8 11.8 6.9 33.0
Norway 42.6 154.0 111.4 10.0 8.8 39.1 34.5 154.7
New Zealand 10.8 17.3 6.5 7.5 5.1 5.6 3.7 15.3
Portugal 9.9 11.5 1.6 0.5 0.2 1.9 0.9 2.7
Sweden 23.5 48.2 24.7 2.1 1.4 8.2 5.6 31.7
United States 117.4 192.2 74.8 71.1 49.4 84.8 58.0 182.3
Europe 16.5 34.0 17.5 2.1 1.7 8.3 6.5 25.7
Minor
Totals 1108.0 2045.4 937.4 284.0 205.5 532.6 384.1 1527.0
Key: LP letter post, TD terminal dues, EDP equivalent domestic postage (70 % of priority
postage), SDR special drawing rights, ICs industrialized countries, ECs e-commerce countries,
DCs developing countries
and inbound mail flows. Net transfers reflect actual economic transfers, and they
affect different POs very differently.
The overall effect of the UPU terminal dues system may appear benign to the
casual observer. DCs pay lower terminal dues than ICs. Do terminal dues not foster
an international subsidy of poor countries by rich countries that is similar to a
desirable cross-subsidy between urban and rural areas in domestic postal services?
Basically no. While the UPU terminal dues system does benefit POs in the neediest
developing countries by reducing the cost of postal communications with industri-

alized countries, this positive effect is a minor by-product of much larger effects.
Table 4 shows the estimated net gains or losses of individual ICs in 2014.
The major transfer of economic value implied by the UPU terminal dues system
flows is between the ICs for the simple reason most international mail is sent
between ICs. ICs are trading discounts for delivery of inbound mail at more or
less fixed rates (fixed by the target system “cap”). This scheme benefits net
exporting countries and hurts net importing countries, since exporters are getting
more discounted services. It also benefits POs with low unit costs and hurts those
with high unit costs, since the low-cost POs are trading relatively cheap services for
relatively more expensive services. In this manner, the terminal dues system creates
“winners” and “losers” among ICs. The main winners are low-cost exporters, and
the main losers are high-cost importers. While individual country data must be
regarded as highly approximate in the terminal dues model, it appears that the main
winners in this intra-IC exchange are the United States, Germany, the United
Kingdom, the Netherlands, and Spain. The main losers are Canada, the Nordic
countries, Switzerland, Italy, France, Japan, and Ireland. In 2014, the implied intra-
IC subsidy was about SDR 418 million.14
As a group, ICs also subsidized ECs and DCs. The total net transfer from ICs to
the three ECs—China, Hong Kong, and Singapore—was about SDR 209 million.
The net transfer to the 138 DCs was about SDR 386 million. About 90 % of this net
transfer was derived from IC undercharges for delivery of inbound items received
from ECs and DCs. A small amount is due to IC overpayments for delivery of letter
post items sent from ICs to DCs. Among DCs the major beneficiaries are 20 rela-
tively large countries that account for 70 % of the outbound letter post.15 About
80 % of the IC net transfer to ECs and DCs is due to the exchange of small packets.
The burden of the IC transfers to ECs and DCs falls unevenly on the intra-IC
winners and losers. The IC losers tend to be high-cost countries, so the economic
cost of delivering letter post mail from ECs and DCs at low, fixed rates is greater for
IC losers than IC winners. Overall, the net economic transfer from IC losers to ECs
and DCs in 2014 was about SDR 372 million, while the net transfer from IC
winners was about SDR 231 million.
The different treatment of IC winners and losers is still more pronounced if one
considers the relative sizes of the POs since the winners tend to be the largest ICs.
One way to take size into account is to divide the net gain or loss by the volume of
outbound letter post items. Outbound letter post is an appropriate divisor since
outbound mailers constitute the only available “tax base” if a PO operates its
14
In the model, results for the United States and Canada are substantially affected by the fact that
the US ships a large volume of small packets to Canada, and the equivalent domestic postage rate
in Canada is established by reference to relatively high parcel post rates since Canada Post offers
no letter post services for small packets. Omitting the US–Canada exchange, the USPO received a
net transfer from other ICs of about SDR 43 million in 2014.
15
Five countries account for one-third of the outbound letter post generated by DCs: India,
Thailand, Czech Republic, South Africa, and Poland.
Table 4 Net transfers by industrialized countries, letter post 2014
326
Gain on Loss on inbound Net gain (loss) Net gain (loss) Net gain (loss) Total net Total net gain
outbound LP to LP from ICs (mil to/from ICs Winner/ to/from ECs to/from DCs gain (loss) (loss) per outbd
Destination ICs (mil SDR) SDR) (mil SDR) loser (mil SDR) (mil SDR) (mil SDR) LP (SDR)
Austria 40.8 38.2 2.6 W 5.7 22.2 25.4 0.22
Australia 15.8 18.1 2.2 L 27.9 19.8 50.0 0.88
Belgium 24.7 29.0 4.3 L 3.2 12.5 20.0 0.29
Canada 21.4 169.9 148.5 L 48.3 57.1 253.9 2.97
Switzerland 38.6 42.3 3.6 L 3.4 13.3 20.3 0.19
Germany 137.9 40.6 97.3 W 7.3 28.3 61.6 0.17
Denmark 15.3 30.8 15.5 L 1.6 6.3 23.4 0.53
Spain 29.8 7.2 22.6 W 0.5 2.1 20.0 0.24
Finland 5.4 18.9 13.5 L 1.6 6.4 21.5 1.39
France 69.8 91.0 21.2 L 6.8 26.4 54.3 0.29
Great Britain 158.2 52.6 105.7 W 7.9 30.3 67.5 0.15
Greece 12.9 3.0 9.9 W 0.3 1.1 8.4 0.23
Ireland 11.2 41.8 30.6 L 4.0 15.5 50.1 1.54
Israel 4.5 3.3 1.2 W 0.2 0.6 0.4 0.03
Iceland 0.5 0.1 0.4 W 0.0 0.1 0.3 0.20
Italy 23.9 69.5 45.6 L 4.6 17.9 68.2 0.97
Japan 10.3 18.6 8.3 L 15.8 11.8 35.9 0.98
Luxembourg 10.7 1.7 9.0 W 0.4 1.4 7.3 0.24
Netherlands 49.3 24.3 25.0 W 1.9 7.3 15.8 0.11
Norway 5.8 111.4 105.7 L 8.9 34.5 149.0 8.24
New Zealand 4.5 6.5 2.1 L 5.2 3.8 11.0 0.68
Portugal 10.9 1.6 9.3 W 0.3 1.0 8.1 0.26
Sweden 14.7 24.7 10.0 L 1.5 5.7 17.2 0.41
United States 209.5 74.8 134.7 W 51.5 60.4 22.7 0.04
J.I. Campbell, Jr.
Europe 11.0 17.5 6.5 L 1.7 6.6 14.8 0.47
Minor
Totals 937.4 937.4 0.0 208.7 386.0 588.2 0.23
– IC winners 665.0 247.5 417.5 W 76.0 154.8 186.7 0.11
– IC losers 272.4 689.9 417.5 L 134.4 237.7 789.7 0.97
Key: same as previous table
Quantifying the Distortive Effects of UPU Terminal Dues
327
international service on a break even basis, i.e., without surcharges on domestic

mailers or payments from government. The last column in Table 4 shows the net
gain or loss from net transfers per outbound letter post item for each IC. Per
outbound letter post item, IC losers lost SDR 0.97 per item, while the IC winners
gained SDR 0.11 per item.
The bottom line in all this is remarkable. The entire global net subsidy fostered
by the UPU terminal dues system in 2014 was paid by 14 or so IC losers and their
mailers. IC losers bear a disproportionate share of the net transfer from ICs to DCs
and ECs, and they transfer to IC winners a sufficient amount to offset all of their net
transfer to DCs and ECs (SDR 231 million) while leaving the IC winners with a
significant net gain (SDR 187 million). All together, in 2014, the UPU terminal
dues system implied a net loss for IC losers of SDR 790 million. About two-thirds
of this net loss was due to the exchange of small packets.
Under the base scenario, the position of the IC losers will not improve over the
next 3 years. In 2017, IC losers will suffer a net negative transfer of SDR 1.3 billion;
almost 80 % will be due to small packets. Over the 4-year course of the agreement,
IC losers will transfer about SDR 4.0 billion to other groups. IC winners will have a
net gain, SDR 779 million because all their transfers to ECs and DCs are more than
covered by a net transfer from IC losers of SDR 1.9 billion. The burden on the IC
losers is about SDR 1.68 per outbound letter post item, while the IC winners gain
about SDR 0.12 per outbound item.
The portrait of economic distortions that emerges from the terminal dues model
and base scenario depends to a significant degree on approximations and assump-
tions. To test the overall correctness of this portrait, it is necessary to consider to
what extent different assumptions result in materially different patterns of
distortion.
While a complete sensitivity analysis is beyond the scope of this chapter, a few
examples may be informative. If the annual growth in small packets dispatched by
ICs and ECs is 20 % greater than that assumed in the base scenario, the 4-year total
of undercharges will increase by about SDR 322 million or 4.4 %. The net transfer
from IC losers will increase by SDR 192 million or 4.7 %. If the percentage of
priority domestic postage rates equivalent to terminal dues is assumed to be 60 %
(instead of 70 %), total undercharges will decrease by SDR 1.9 billion (26 %)—
47 % of the estimated market value—and the net transfer from 12 IC losers (two IC
losers are changed to winners) will be reduced to SDR 2.9 billion (28 %).
Conversely, if equivalent domestic postage is taken to be 80 % of priority domestic
postage, total undercharges rise to SDR 9.3 billion (+26 % increase)—60 % of the
market—and net transfers by 15 IC losers increase by SDR 882 million (+22 %). If
domestic postage rates in the ICs increase by 4 % per year (instead of 2.3 % as
assumed in the base scenario), total undercharges will increase by SDR 372 million
(+5.1 %) and the net transfer by the IC losers will increase by SDR 171 million
(+4.2 %). Such considerations suggest the overall pattern of distortions revealed by
the base scenario is substantially correct even though quantities will vary depending
on assumptions.
5 Conclusions
The proposition that the UPU terminal dues create significant economic distortions
and anticompetitive effects cannot be seriously contested. This chapter considers
the further problem of estimating the quantity of economic distortions implied by
the current UPU terminal dues system, in effect from 2014 through the end of 2017.
The chapter describes a plausible, if approximate, model for evaluating these
distortions and suggests that the pattern of economic distortions that is revealed is
not particularly sensitive the assumptions adopted. This analysis suggests the
following conclusions, simplified by use of round numbers:
First, the economic distortions implied by the UPU terminal dues may be
summarized using two measures. Total undercharges measures the potential for
economic distortions and anticompetitive effects resulting from the pricing behav-
ior by POs. Net transfers between PO measures the portion of total undercharges
financed by transfers of economic value between POs, i.e., the extent to which some
POs are benefitting at the expense of others.
Second, total undercharges fostered by the UPU terminal dues system are
substantial compared to the real value of the market, i.e., the domestic postage
that would be charged for delivery services provided. When terminal dues are
compared to equivalent domestic postage rates, it appears that POs are
undercharging each other for delivery of inbound international letter post by
40–60 %. Undercharges for letter post items from EC and DCs approach 70 % or
more. Total undercharges will amount to approximately SDR 7.5 billion over
4 years, plus or minus perhaps two billion. About two-thirds of this is due to
undercharging for delivery of small packets.
Third, the net effects of undercharging vary very substantially among post
offices. POs (and their mailers) are, in effect, subsidizing each other. ICs are
subsidizing ECs and DCs. Some ICs are subsidizing other ICs. The total amount
of implied net transfers is on the order of SDR 4 billion over 4 years. Approxi-
mately, SDR 1.5 billion in net transfer benefits three POs specializing in
e-commerce: China, Hong Kong, and Singapore. Another SDR 1.5 billion is
divided among 138 DCs, most of it going to the largest and least needy POs. The
remaining net transfer, approaching SDR 1 billion, benefits a dozen or so, generally
very large, ICs. Under the UPU terminal dues system, the entire cost of this global
network of postal subsidies is paid by 12–15 ICs and their mailers, including the
POs of Canada, the Nordic countries, Switzerland, Italy, France, Japan, and Ireland.
References
Anson J, Helble M (2013) A gravity model of international postal exchanges. In: Crew M,
Kleindorfer P (eds) Reforming the postal sector in the face of electronic competition. Edward
Elgar, Cheltenham
Anson J, Helble M (2014) Global postal connectedness. In: Development strategies for the postal
sector: an economic perspective. UPU, Bern
Copenhagen Economics (2014) The economics of terminal dues. Report for the U.S. Postal
Regulatory Commission
European Commission (1992) Green paper on the development of the single market for postal
services. COM(91) 476
Universal Postal Union (2011) POC C 1 TDG 2011.1 Doc 4a. In: Terminal dues model—results of
the studies on costs, flows, tariffs and items per kilogramme (committee document). UPU, Bern
UPU (2012a) TDG terminal dues proposed options—impact assessment tool, Version 3 (Excel
spreadsheet). UPU, Bern
UPU (2012b) Doha congress, Doc 20b. UPU terminal dues for the period 2014–2017 (Congress
document). UPU, Bern
UPU (2013a) Development of postal services in 2012 (Powerpoint presentation). UPU, Bern
UPU (2013b) POC C3 LPRG 2013.2 Doc 3. Results of the cost study (committee document). UPU,
Bern
UPU (2014a) Development of postal services in 2013 (Powerpoint presentation). UPU, Bern
UPU (2014b) POC C3 LPRG 2014.2 Doc 4a. Results of the items per kilogramme (IPK) study
(committee document). UPU, Bern
United States, Department of Justice (1988) Evaluating a proposed agreement on terminal dues
(unpublished study)
United States Postal Service (2012) Comments of United States postal service (27 Aug 2012)
Postal Regulatory Commission, Docket PI2012-1
UPU Compensation Rates for Packages
Under EU Competition Law: Are There
Lessons to Be Learned from Other
International Fee Arrangements?
Ralf Wojtek
1 Introduction
The EU Green Paper on Parcels (2012) and the follow-up “Roadmap” (2013a) raise
concerns about the effectiveness of the international flow of parcels throughout the
European Union. According to the Green Paper, the availability of effective cross-
border parcel services is key to the success of e-commerce in Europe. The Roadmap
laid out a series of steps that the Commission believed postal operators and other
parties should take to improve cross-border parcel services. Although the Roadmap
left the initiative to industry, the Commission warned that it “will monitor devel-
opments during an 18 months period” and will then “consider appropriate correc-
tive or additional actions to remedy the market failures” (Roadmap, p. 7). As part of
its review process, the Commission in May 2015 opened a public “Consultation on
cross-border parcel delivery.”1
The Roadmap includes a call for national postal operators to improve interop-
erability. In August 2014, WIK-Consult completed a study for the Commission on
“Design and development of initiatives to support the growth of e-commerce via
better functioning parcel delivery systems in Europe.” As WIK-Consult (2014)
points out, “While national posts and private carriers need to improve interopera-
bility of cross-border parcel operations, they must do so in a manner that is
consistent with the principles and requirements of EU competition rules”
(p. 186). The study concludes, “In sum, there is an apparent need for greater
cooperation among parcel service providers in the cross-border market and, at the
This presentation reflects the personal views of the author.

1
Available at: http://ec.europa.eu/growth/tools-databases/newsroom/cf/itemdetail.cfm?item_id¼8169
(accessed: 22.06.2015).
R. Wojtek (*)
Heuking Kühn Lüer Wojtek, Hamburg, Germany
e-mail: R.Wojtek@heuking.de

332 R. Wojtek
same time, substantial uncertainties surrounding the application of the competition

rules to such cooperative ventures” (p. 189).
Interoperator compensation is a core element of interoperability. The compen-
sation systems developed by the Universal Postal Union (UPU)—“terminal dues”
for letters as well as small packets up to 2 kg and “inward land rates” for parcels—
may be considered as the default background for any alternative arrangements
possibly resulting from the process of improving EU cross-border postal services.
Geradin (2012) demonstrated that rate agreements between postal operators,
whether organized inside or outside the UPU, are subject to and to be measured
against the rules of Art. 101 and 102 of the Treaty on the Functioning of the
European Union (TFEU). Geradin’s legal conclusions were recently reinforced by
a study of Copenhagen Economics (The Economics of Terminal Dues, 2014),
which identified several types of economic and competitive distortions flowing
from current UPU “terminal dues” provisions. Postal operators are thus placed in a
difficult position. They are strongly encouraged to cooperate more to improve
cross-border parcel services. At the same time, they may be threatened by the
competition rules if they cooperate too closely.
This chapter will look into the problems posed by the UPU inward land rates
(ILR) (Sect. 2) and describe the framework under EU competition law (Sect. 3),
followed by an analysis of other international rate agreements and their treatment
under EU and US competition law (Sect. 4), closing with conclusions (Sect. 5).
2 The UPU-System of Centrally Administered Rates
The UPU has created a system of payment between designated operators (DOs) for
the delivery of parcels originating in the territory of another DO, the ILR. The ILR
charged by the DO performing the delivery is based on the so-called base rate. The
base rate is calculated as 71.4 % of the ILR charged by the DO in 2004 (plus
inflation-linked adjustments), provided that the resulting amount is not less than the
global minimum base rate of 2.85 SDR per parcel and 0.28 SDR per kilogram,
corresponding to 4.25 SDR for a 5 kg parcel. The base rate is increased by a bonus
system for DOs offering special features, such as track and trace, home delivery,
predefined delivery standards, and use of the common internet-based customer
inquiry system.
To the extent that the 2004 ILR was cost based, the calculation of the current ILR
(71.4 % of 2004 plus inflation and possible bonuses) is, at least also to some extent,
cost based (Haller et al. 2013, p. 279).2 The reference to the 2004 ILR further
implies that the costs have not changed over the past years except for inflation.
2
Assuming that the ILR charged by a DO in 2004 amounted to 70 % of the comparable rate for
delivery of a domestic parcel, the new ILR amounts to 50 % of the domestic rate. This would be far
below the level which is considered acceptable for letters, cf. Campbell, sub. 3.
UPU Compensation Rates for Packages Under EU Competition Law: Are There. . . 333
Furthermore, there is always the global minimum base rate which does not reflect
the delivery costs of any particular DO.
A study commissioned by the UPU (“Market Research on International Letters
and Lightweight Parcels,” 2010) shows that in most countries the ILR differs
strongly from the prices charged by the postal operators for comparable domestic
parcels. In some countries, the ILR is above, in others well below the comparable
domestic rate (UPU 2010, pp. 60–61). The extent to which the ILR in any given
country differs from the local tariff for similar parcels varies broadly without
plausible explanation (Haller et al. 2013, p. 278). This finding was recently con-
firmed by a survey of the Postal Operations Council (POC) made in 2014 (POC
2014).
As shown by Copenhagen Economics (2014, p. 34 et seq.) and Campbell (2014,
p. 15 et seq.) for terminal dues, below-cost delivery rates constitute a barrier to
entry for private competitors or other DOs intending to enter the market for cross-
border delivery. Below-cost ILR can, therefore, lead to territorial separation of
markets for inward delivery. Conversely, where the ILR exceeds the cost of
delivery the DO is taking advantage of its particular position as participating DO
within the UPU system. In both cases, the lack of cost-orientation leads to distortion
of competition to the detriment of consumers.
3 EU Competition Law
Under EU competition law agreements or concerted practices between undertak-

ings that have as their object or effect the distortion of competition within the
internal market are generally prohibited unless they fall under a block exemption or
if the conditions for an individual exemption under Art. 101 (3) TFEU apply. In
addition, Art. 102 TFEU prohibits the abuse of a dominant position, for example, by
imposing unfair prices or applying dissimilar conditions to equivalent transactions
with other trading parties, thereby placing them at a competitive disadvantage.
3.1 Application of Art: 101 TFEU
All DOs exchanging parcels under the terms of the UPU are active in the same
product market, i.e., the delivery of inward cross-border postal items, but in
different geographic markets. The UPU provisions provide a horizontal agreement
enabling the DOs to exchange parcels at pre-agreed rates. Therefore, the prohibition
of Art. 101 TFEU may apply if the exchange is based on an agreement or concerted
practices between undertakings.
334 R. Wojtek
3.2 Agreement or Concerted Practices
The ILR system is part of an international treaty, the provisions of which are
transformed into national laws by acts of the national law makers. Neither the
UPU nor the ILR system within it constitutes an agreement between undertakings.
However, as Geradin (2012) has pointed out with respect to the provisions on
terminal dues for the inward delivery of letters (p. 6), representatives of the member
states’ DOs have been instrumental in negotiating the terms of the UPU as well as
the specific terms on inward delivery rates. This is particularly true where specific
provisions are drafted and agreed within the POC which is run by representatives of
DOs. Based on the practical evolution of the rate setting process at UPU level,
Geradin concludes that there are agreements between undertakings, namely the
DOs, or at least concerted practices.
Apart from the DOs participation in the rate setting process within the POC, the
DOs become involved in the application of the ILR system as they implement it
among themselves, exchange parcels on the basis of ILR, and make payments
accordingly. The European Court of Justice (ECJ) has held repeatedly that under-
takings engaging in price fixing or market sharing are acting in violation of EU
competition rules, even if the conduct is authorized by national law.3
An agreement or concerted practice which includes the exchange of parcels on
the basis of centrally fixed rates constitutes, on the face of it, a breach of the
prohibition of price fixing contained in Art. 101 TFEU. The Commission has
made it clear repeatedly that price fixing arrangements are considered as “hardcore”
violations. The relevant block exemptions exempting certain agreements from the
application of Art. 101 TFEU do not apply to agreements containing price fixing
clauses. By applying the ILR, the participating DOs forego the possibility to
negotiate cost-based rates. Consumers in the end pay the price because the cost of
delivery usually constitutes the major part of the total cost of transportation. If the
ILR exceeds the comparable local delivery rate, consumers will be overcharged; if
the ILR is below cost, competition may be deterred, which could lead to higher
prices for consumers in the long run. The rate agreement also has further anticom-
petitive effects (as outlined above), particularly the strengthening of geographic
market allocation among DOs and distortions of competition making it difficult for
competitors to enter the market. Therefore, the practice of applying ILR among
DOs squarely falls within the ambit of Art. 101 TFEU.
3.3 Exemption
The question is whether the ILR agreement meets the test of Art. 101 (3) TFEU,
which would render the prohibitions of Art. 101 TFEU inapplicable. Art.
3
ECJ, ECLI:EU:C:2003:430, judgment of 9.12.2003, case C-198/01.
101 (3) TFEU contains four conditions which must be fulfilled together: Firstly, the
agreement must lead to efficiency gains, and, secondly, the restrictions must be
indispensable for the attainment of such gains; thirdly, consumers must receive a
fair share of the resulting benefits and, fourthly, the agreement must not lead to
eliminating competition. In applying these four criteria to the ILR agreement, one
must not only consider the complexities of the ILR agreement but also the market
and the requirements for the inward delivery of international parcels as well as the
effect on competition considering the relatively recent transition from strictly
regulated to fully liberalized postal markets in Europe.4
4 Other International Agreements
In the postal industry, there is already some experience with centralized rate
agreements in liberalized market for letter mail. But it is not the only industry
experiencing a transition from strictly regulated to fully liberalized markets. It may,
therefore, be helpful to look at similar arrangements in other markets and their
treatment by the competition authorities.
4.1 International Exchange of Letters: REIMS II
A good example for the treatment of an international rate agreement is the agree-
ment on terminal dues for inward delivery of international letters (European
Commission 1999, REIMS II), which was re-notified to the European Commission
in 2001 and decided upon in 2003 (European Commission 2003, “REIMS II
Re-Notification”).5
In REIMS II, the DOs of 17 EEA countries agreed to replace the terminal dues
system for letters as set by the UPU with a new rate system reflecting a percentage
of the domestic tariffs for letter mail. By tying the tariff for inbound mail to the
domestic tariff, the parties to the REIMS agreement intended to align the terminal
dues closer to the actual cost of delivery (REIMS II Re-Notification, recital 35).
Since the original terminal dues as set by the UPU were extremely low, the REIMS
parties agreed to increase the terminal dues among themselves in three steps until
the new terminal dues reached 80 % of the respective domestic letter tariff. In
addition, the REIMS agreement provides that all parties to the agreement have
access to the other DOs generally available domestic rates, especially the rates for
4
Cf. Art. 2 of Dir. 2008/6/EC, the 3rd Postal Directive, European Community (2008); also
compare European Commission (1998), Notice on the application of the competition rules to the
postal sector.
5
Since then, the REIMS agreement has been continued and is currently in place as REIMS V.
336 R. Wojtek
bulk mail. In order to provide incentives for quality of delivery the system provides
for penalties if a DO does not meet certain quality standards. The penalties can be as
high as 50 % (REIMS II, Re-Notification, recital 42). In addition, the REIMS
agreement allows its parties to enter into further bilateral or even multilateral
agreements.
The EU Commission took a very critical view of REIMS II and accepted it only
after certain modifications were agreed. The Commission agreed, however, with the
objective of REIMS with respect to the improvement of the efficiency and quality
of delivery of international letter mail. The Commission also accepted the argument
that by tying the terminal dues to a percentage of the price of domestic letter mail,
the REIMS rates were more closely linked to actual costs of delivery which would,
in the eyes of the Commission, reduce substantially the risk of distortions of
competition by cross-subsidization (REIMS II Re-Notification, recital 119).
The main objection of the Commission to the agreement focused on the dis-
crimination of third party postal operators which had no access to the REIMS II
system. The Commission permitted the agreement under the condition that “each
REIMS II Party should provide to any third-party postal operator competing with
the REIMS II Parties for the provision of outgoing cross-border mail services in any
other REIMS II country, delivery of incoming cross-border mail in its country at
terminal dues and under conditions which are non-discriminatory as compared to
those that the REIMS II Party offers to REIMS II Party (ies) in the sender’s
country” (REIMS II Re-Notification, recital 171).
Subsequent REIMS agreements must be read in conjunction with the conditions
imposed by the Commission on REIMS II. The Commission’s decision shows that
the Commission is prepared to accept rate fixing among the members of the REIMS
II agreement provided that the rates are based on costs and are made available to
third parties. The expected efficiency gains may justify a rate agreement as long as
the agreement does not exclude competition or discriminate third parties (Baratta
2004, pp. 25, 28–29).
4.2 Air Transportation: IATA
Not unlike the postal sector, the sector of air transportation used to be state
controlled and still is highly regulated. Some national markets are still dominated
by national flag carriers owned by the state. Over the past few decades, the airline
industry has undergone liberalization but remains subject to tight regulation, for
example, with respect to ownership and control of European or US airlines
(Kociubiński 2014, p. 191). There are restrictions regarding the operations of flights
outside the respective home country and in most cases airlines are prohibited from
offering unconnected flights in third countries. This leads to a need for cooperation
between airlines for international flights and “interlining,” an arrangement where
one airline sells a ticket for a complete flight which is broken down into different
segments, each of which is handled by another airline.
Especially for international flights it was, therefore, necessary to develop a

system of cooperation which includes the acceptance of tickets from other airlines
as well as the calculation and distribution of the fare paid for the entire flight. The
International Air Transport Association (IATA) developed already in the 1940s a
system of tariff consultation between all member airlines on all major international
routes. These consultations resulted in the so-called conferences which dealt with
agreements on tariffs as well as on the use of airline slots and other technical aspects.
As liberalization of the airline industry began to take place, competition author-
ities in the USA and the EU started to look more closely into the IATA agreements
and to question the necessity, particularly of rate agreements, in order to attain the
objectives pursued by IATA. As a first step, in 1996, the EU Commission excluded
agreements on tariffs for air freight from the original, broader block exemption
regulation 1617/93/EC for air transport tariff conferences (Art. 1 of Regulation
1523/96/EC, European Commission 1996). Five years later, the Commission fired a
warning shot by issuing a statement of objections (European Commission, Press
Release, IP/01/694 of May 15, 2001). In the statement of objections, the Commis-
sion took the preliminary view that IATA cargo tariff consultations infringe
competition rules because they did not seem to be indispensable to provide cus-
tomers with efficient interlining services within the EEA. Within 5 months follow-
ing the statement of objections IATA agreed to end all activities involving the joint
setting of rates for air transport of freight within the EEA (European Commission,
Press Release, IP/01/1433 of October 19, 2001).
On the other side of the Atlantic, the US Department of Transportation had
granted IATA immunity from the prohibition of cartels under Sect. 1 of the
Sherman Act (DOT 2007, pp. 1, 6 et seq.). However, in 2003, when IATA filed
for the approval of certain amendments to its tariff conference on freight rates the
US Department of Justice pointed out that due to liberalization the airline sector had
undergone significant changes which opened the way to alternative solutions for
international transport and interlining (DOJ 2003, pp. 5–6). The DOJ further noted
that the new tariff proposal did not show any economic benefits but, in fact, led to an
increase in low density cargo rates up to 20 %. In the opinion of the DOJ, the IATA
proposal failed to show an objective need for joint cargo rates and did not demon-
strate that no less anticompetitive alternatives existed. Interestingly, the DOJ also
referred to the statement of objections sent to IATA by the EU Commission in 2001
with regard to freight conferences (DOJ 2003, p. 8).
The US Department of Transportation, after issuing a draft order in 2006 (DOT
2006, Order 2006-7-3), decided to terminate the exemption of IATA tariff confer-
ences on cargo and passengers with respect to the markets in the USA, the EU, and
Australia (DOT 2007, Order 2007-3-23). The order entered into force on June
30, 2007. On the same day, the EU block exemption relating to IATA passenger
tariff conferences terminated, too.6 Similar to the reasoning of the Department of
6
Cf. DOT (2007), p. 51 and Art. 1 b-d, Art. 4 of Regulation 1459/2006/EC, European
Commission (2006).
338 R. Wojtek
Justice, the Department of Transportation pointed out that the IATA failed to
demonstrate a need for general tariff conferences, particularly when airlines have
the possibility to enter into alliances. The Department of Transportation further
noted that the conferences were often more expensive than bilaterally agreed or
individually set fares. By setting bench mark prices and through the exchange of
information the IATA conferences reduced price competition to the detriment of
consumers.
The case of the IATA tariff conferences shows that competition authorities take
a closer look once the market opens up to competition. With competition in place
there are usually less restrictive alternatives available to the parties in order to reach
their objectives. At the same time, it becomes extremely difficult to justify the
objective need for a general rate agreement. The positions taken by the European
Commission and the US authorities on IATA agreements demonstrate that there is
no room for general tariff agreements if there are less restrictive means to improve
the interoperability of international operators.
4.3 Airline Alliances
From 2006 on the EU Commission conducted formal investigations against three

major airline alliances, Star Alliance, SkyTeam, and Oneworld. In all cases, the
Commission makes it clear that the alliances may not be used to exclude compet-
itors from the market:
4.3.1 Star Alliance
The Commission cleared the Agreement between Air Canada, Lufthansa, and
United Airlines, including the so-called A++ Agreement setting joint tariffs on
certain transatlantic routes and providing for revenue sharing among the members.7
The agreement was permitted under condition that the Star Alliance members
permit competitors to combine their own regular flights with flights offered by
the members of the alliance. Furthermore, passengers of a competitor offering a
combined flight with members of the alliance must have access to the same frequent
flyer benefits which the alliance offers to its own members.
In its decision, the Commission accepts the argument of efficiency gains to be
achieved by the alliance, provided that the agreement does not restrict competitors
which need to cooperate with members of the alliance in order to offer attractive
flight schedules.
7
Decision of May 23, 2013, case COMP/AT.39.595, European Commission (2013b).
4.3.2 SkyTeam
Another alliance of airlines including Air France and Delta, “SkyTeam,” was
investigated by the Commission since 2006. In its notice of 2007, the Commission
noted that it did not see material competition problems in an agreement covering
code sharing, frequent flyer programs, or lounge access (European Commission
2007, recitals 3 et seq.). However, to the extent that the agreement went further and
included certain city connection flights (“city pass”), the Commission noted that
such cooperation agreements covered all key parameters of Art. 101 and constitute
barriers to entry, given the slot constraints at main EU airports, the parties’
frequency advantage and hub dominance, and network effects from the parties’
joint frequent flyer programs and other incentive schemes (European Commission
2007, recital 6).
Subsequently, the parties gave certain commitments which included the avail-
ability of slot space and access to frequent flyer programs to competitors. In
addition, the parties agreed to conclude interlining agreements with new entrants
and to enter into prorate agreements for the so-called behind and beyond traffic on
intra-EU roads. Following these commitments, the Commission closed the case in
2012.8 It should be noted, however, that subsequently the Commission opened new
proceedings against certain members of the alliance which had entered into more
intensive forms of cooperation, including revenue sharing on transatlantic flights.9
4.3.3 Oneworld
A cooperation agreement between British Airways, American Airlines, and Iberia,

members of “Oneworld,” was subject to investigation of the Commission in 2009.
According to their agreement, the parties intended to jointly manage flight sched-
ules, capacity planning and pricing, and to share revenues on transatlantic routes.
The Commission closed its investigation after the three members entered into a
number of commitments including a slot commitment, a fare combinability agree-
ment, and a special prorate agreement, allowing competitors to obtain favorable
terms from the parties for connecting passengers from the parties’ short-haul flights
in Europe and North America with own transatlantic services (European
8
See European Commission, Press Release, IP/12/79 of 27.1.2012, “Antitrust: Commission opens
a probe into transatlantic joint venture between Air France-KLM, Alitalia, and Delta and closes
proceedings against eight members of Sky Team airline alliance,” available at: http://europa.eu/
rapid/press-release_IP-12-79_en.htm?locale¼fr (accessed: 22.06.2015).
9
Case COMP/37.984, which has been cleared by the Commission in May 2015, subject to
conditions comparable to the ones imposed in the Star Alliance and the Oneworld cases. See
European Commission, Press Release, IP/15/4966 of 15.5.2015, “Antitrust: Commission accepts
commitments by SkyTeam members Air France/KLM, Alitalia, and Delta on three transatlantic
routes,” available at: http://europa.eu/rapid/press-release_IP-15-4966_en.htm?locale¼en
(accessed: 22.06.2015).
340 R. Wojtek
Commission 2010b, p. 46 et seq.). Based on this commitment, a competitor is able

to fly its customers on a local (e.g., Manchester—London) flight operated by BA,
and then connect it to its own London—New York flight. Furthermore, the
Oneworld members agreed to grant passengers of competitor’s access to their
own frequent flyer programs on flights made in combination with the competitor’s
flight (Memo of July 14, 2010, MEMO/10/330).
4.3.4 Interim Result
The airline cases show a clear tendency of the competition authorities to allow
large-scale agreements such as IATA only if and when necessary to achieve the
desired efficiency objective of the cooperation. Upon liberalization of the airline
industry, new alternatives to a centrally administered system became available and
the general IATA agreement was no longer considered acceptable under competi-
tion rules. However, also the smaller alliances are subject to restrictions under
competition laws. With respect to such alliances, the competition authorities are
particularly sensitive to anticompetitive effects which tend to eliminate competi-
tion. For this reason, the Commission has imposed the condition that the alliance
members must allow competitor’s access to interlining and frequent flyer programs.
4.4 Maritime Transport and Liner Conferences
The market for maritime transport is still subject to regulation by states. The
International Maritime Organization (IMO) which exists under the roof of the
United Nations is in charge of standard-setting relating to international maritime
transport. The maritime industry is another example of a market where participants
tend to agree on routes and rates in the so-called liner conferences.
The case of the Trans-Atlantic Agreement (TAA) involved collective price
setting by the members to the conference. Although at that time, the liner confer-
ences were exempted from EU competition rules under block exemption regulation
4056/86/EC (European Community 1986), the Commission held that the TAA rate
agreement did not comply with the regulation as it included rates for land transport
of sea containers between ports. The Commission criticized that the agreement did
not contain any incentive to improve the quality of service (European Commission
1994, recital 465).
Following the Commission’s decision of 1994, the parties to the TAA concluded
a new agreement, the Trans-Atlantic Conference Agreement (TACA) which was,
however, not accepted by the Commission. A revised TACA which was eventually
accepted by the Commission includes an obligation of the members to refrain from
price fixing of land transport and to permit members of the conference to enter into
bilateral contracts with others outside the TACA system (European Commission
2002, p. 9)
In 2009, the Commission issued its regulation 906/2009/EC exempting cooper-

ation agreements between liner shipping companies from the application of Art.
101 (1) TFEU, provided that the cooperation does not allow price fixing (European
Commission 2009). The regulation points out that the consortia formed by liner
companies generally help to improve the productivity and quality of liner shipping
services through the economies of scale they allow in the operation of vessels and
utilization of port facilities. By contrast, unjustified limitation of capacity and sales,
as well as the joint fixing of freight rates or market allocation, are considered to be
contrary to the goal of efficiency and cannot be exempted. Furthermore, to qualify
for the exemption, the combined market share of the consortium members in the
relevant market must not exceed 30 % (Art. 5). The maritime cases confirm the
general practice of the Commission to allow certain industry agreements under the
condition that they lead to efficiency gains without excluding competitors from the
market.10
4.5 Austrian Freight Forwarding Agents SSK
A group of 30 Austrian freight forwarding agents formed a collective forwarding

conference (Speditions-Sammelladungs-Konferenz “SSK”) with the objective to
set common freight rates for cross-border collective shipments into Austria. The
members of the SSK argued that shippers would benefit from the system by offering
them equal price conditions, thereby promoting fair competition of its members.
The Austrian anti-trust court (Kartellgericht) approved the agreement in 1996 and
the SSK practiced the central administration of freight rates until 2010.
The EU Commission did not agree with this decision and started to make
unannounced inspections as early as 2007. Subsequently, the Austrian competition
authority conducted its own investigation and decided that the SSK rate system was
a cartel within the meaning of Art. 101 (1) TFEU and that its members should be
fined for breach of Art. 101. The European Court of Justice confirmed this decision
and explained that the members were subject to penalties regardless of the legal
advice received on the basis of the decision of the local court in 1996.11 In the end,
the participants to the SSK were fined to a total of EUR 17.5 million.12
10
Cf. the recent announcement of the EU, the USA, and Chinese competition authorities to
cooperate more closely in monitoring increased tie-ups among shipping groups; available at:
http://www.reuters.com/article/2015/06/18/antitrust-shipping-idUSL5N0Z444W20150618
(accessed: 22.06.2015).
11
ECJ, ECLI:EU:C:2013:404, judgement of 18.6.2013, Case C-681/11—Bundeswettbe-
werbsbeh€ orde vs. Schenker, recitals 30, 43 and ruling 1.
12
BWB, press release of 23.1.2015, case BWB/K-150, available at: http://www.bwb.gv.at/
Aktuell/Seiten/Geldbußenentscheidung-gegen-Spediteure.aspx (accessed: 22.06.2015).
342 R. Wojtek
5 Conclusions
The practice of the competition authorities with respect to other regulated industries
can be used as an indication of the possible outcome of a review of the ILR
agreement by competition authorities and ECJ.
The ILR agreement complies with EU competition law (and possibly US
competition law) only if its conditions are necessary to achieve its goals and if
these goals justify the restrictions of competition. The analysis of similar agree-
ments in the airline and maritime industries shows that competition authorities
become increasingly reluctant to approve global rate agreements which include all
of the main players at the same industry. The examples show that competition
authorities in Europe and in the USA have urged global associations of originally
state-owned or controlled enterprises to limit the scope of their agreements, with
respect to not only their contents but also regarding membership. In liberalized
markets, competition authorities have consistently held that there was no need for
all-encompassing agreements if the same goals could be reached in bilateral or
multilateral agreements. Since the market for parcel delivery is fully liberalized in
the EU as well as in the USA, there are serious questions if a rate agreement which
includes all of the originally state-owned postal operators would be acceptable in
today’s liberalized market.
The UPU agreement on parcel rates may lead to market foreclosure. The
Commission is particularly sensitive with respect to elements of foreclosure. The
aspect of foreclosure of existing or potential competitors becomes more important
as the combined market share of the parties increases. Consequently, both the EU
Commission and the US authorities have stressed that bilateral agreements or
smaller alliances are preferable to large-scale industry agreements. The rule of a
30 % combined share of all foreign parcels coming into a country, as similarly
expressed in the EU regulation on maritime liner agreements, may serve as a good
indicator of what is acceptable.13 If the combined market share of the participants
exceeds that ratio, they may have to resort to bilateral agreements or smaller
multilateral agreements which have less of an impact on the market. Particularly
in the B2C market DOs often hold significant market shares, at least in their
territories, due to their territorial delivery network and availability of post offices.
Apart from these concerns, the ILR agreement must have the objective to
improve the efficiency of cross-border parcel delivery to the benefit of consumers
and must provide for adequate incentives to achieve this objective. The system must
be cost based and not lead to distortions of competition. To the extent that the
system may lead to restrictions on competition, it must provide for competitors’
access at the same conditions.
13
Generally, for cooperation between competitors regulation (EU) No 1218/2010 (European
Commission (2010a), “Specialization Block Exemption”) provides for a threshold of 20 %
combined market share of all parties (Art. 3). Block exemptions are not at all available for the
so-called hardcore restrictions of competition, esp. price fixing.
The ILR system has as its stated objective the goal to improve efficiency and
quality. However, the system is unlikely to achieve this goal because it does not
state the minimum quality standards. There is no quality standard as to delivery
times or maximum loss or damage rates nor does the agreement include penalties if
such standards are not reached. In this respect, the ILR agreement offers less in
terms of quality improvement as compared to REIMS II which provides for
substantial penalties. The ILR rate must be paid regardless of the quality of the
delivery service.
Also with respect to the cost-criterion, the ILR agreement meets the test only
partly. While 71.4 % of the rate is based on the 2004 ILR, this applies only if this
rate exceeds the global minimum base rate. Thus, the global minimum base rate
constitutes the default rate which can be exceeded but not reduced even if the cost
of delivery is below the base rate. The failure to comply with the cost test works to
the detriment of consumers and consequently violates the basic principle of Art.
101 (3) TFEU which requires that efficiency gains must be passed on to consumers.
Additionally, any differences between domestic rates and the ILR charged by the
same post office for the delivery of an inbound parcel constitutes a discrimination
between domestic and foreign senders. Such price differences would be discrimi-
natory and, therefore, unlawful under Art. 101 and 102 TFEU unless it can be
shown that the differences are cost based and therefore justified.
The ILR agreement does not contain a provision allowing third parties access to
its system. This constitutes the most serious violation of the principles as the
Commission has expressed them in the REIMS II case as well as in the airline
cases. Access of competitors has become a fundamental principle without which the
Commission is unlikely to approve a rate agreement.
References
Baratta Rosario (2004) The REIMS II exemption decision: enhancing competition in the cross-
border mail market through third party access. Competition Policy Newsletter of the EU
Commission, No. 1, Spring 2004
Campbell James I Jr (2014) Estimating the effects of UPU terminal dues, 2014–2017. Paper
presented at the 8th Bi-annual postal economics conference on e-commerce. Digital Economy
and Delivery Services, Toulouse
Copenhagen Economics (2014) The economics of terminal dues. Final Report for the U.S. Postal
Regulatory Commission, September (The Economics of Terminal Dues)
Department of Justice (DOJ) of the United States of America (2003) Comments on Docket
OST-2003-14480 before the United States Department of Transportation, Washington, DC,
29 May 2003. http://www.justice.gov/atr/public/comments/201073.htm. Accessed 22 June
2015
Department of Transportation (DOT) of the United States of America (2006). Order 2006-7-3,
issued 5.7.2006 under Docket OST-2006-25307 on International Air Transport Association.
Order to show cause. http://apps.americanbar.org/antitrust/at-committees/at-tei/air-resources/
DOT_Order_to_Show_Cause_IATA_Immunity_Status.pdf. Accessed 22 June 2015
344 R. Wojtek
Department of Transportation (DOT) of the United States of America (2007). Order 2007-3-23,
issued 30.3.2007 under Docket OST-2006-25307 on International Air Transport Association,
Final Order
European Commission (1994) Decision of 19.10.1994, case IV/34.446—Trans-atlantic Agree-
ment (TAA), OJ 1994/L 376/1 (TAA)
European Commission (1996). Commission Regulation (EC) No 1523/96 of 24 July 1996
amending Regulation (EEC) No 1617/93 on the application of Article 85 (3) of the Treaty to
certain categories of agreements and concerted practices concerning joint planning and coor-
dination of schedules, joint operations, consultations on passenger and cargo tariffs on sched-
uled air services and slot allocation at airports, OJ 1996/L 190/11 of 31 July 1996
European Commission (1998) Notice from the Commission on the application of the competition
rules to the postal sector and on the assessment of certain State measures relating to postal
services, OJ 1998/C 39/2 of 6 Feb 1998
European Commission (1999) Decision of 15.09.1999, Case No. IV/36.748—REIMS II, OJ 1999/
L 275/17 of 26 Sept 1999 (REIMS II)
European Commission (2002) Decision of 14.11.2002, Case COMP/37.396/D2—Revised TACA,
OJ 2003/L 26/53 of 31 Jan 2003
European Commission (2003) Decision of 23.10.2003, case COMP/C/38.170—REIMS II
renotification, OJ 2004/L 56/76 of 24.2.2004 (REIMS II Re-Notification). 6 Feb 1998
European Commission (2006) Commission Regulation (EC) No 1459/2006 of 28 September 2006
on the application of Article 81(3) of the Treaty to certain categories of agreements and
concerted practices concerning consultations on passenger tariffs on scheduled air services
and slot allocation at airports. OJ 2006/L 272/3 of 3 Oct 2006
European Commission (2007). Notice published pursuant to Article 27(4) of Council Regulation
(EC) No 1/2003 in Case COMP/37.984—SkyTeam, OJ 2007/C 245/46 of 19 July 2010
European Commission (2009) Commission regulation (EC) No 906/2009 of 28 September 2009
on the application of Article 81(3) of the Treaty to certain categories of agreements, decisions
and concerted practices between liner shipping companies (consortia). OJ 2009/L 256/31 of
29 Sept 2009
European Commission (2010a) Commission regulation (EU) No 1218/2010 of 14 December 2010
on the application of Article 101(3) of the Treaty on the Functioning of the European Union to
certain categories of specialisation agreements. OJ 2010/L 335/43 of 18 Dec 2010 (Speciali-
zation Block Exemption)
European Commission (2010b) Decision of 14.07.2010, Case COMP/39.596—BA/AA/IB
European Commission (2012). Green paper. An integrated parcel delivery market for the growth
of e-commerce in the EU, 29 Nov 2012, COM (2012) 698 final (Green Paper)
European Commission (2013a) Communication from the Commission. A roadmap for completing
the single market for parcel delivery—build trust in delivery services and encourage online
sales. 16 Dec 2013, COM (2013) 886 final (Roadmap)
European Commission (2013b) Decision of 23.5.2013, Case COMP/AT.39.595—Continental/
United/Lufthansa/Air Canada, doc. C (2013) 2836 final
European Community (1986) Council Regulation (EEC) No 4056/86 of 22 December 1986 laying
down detailed rules for the application of Articles 85 and 86 of the Treaty to maritime
transport. OJ 1986/L 378/4 of 31 Dec 1986
European Community (2008) Directive 2008/6/EC of the European Parliament and of the Council
of 20 February 2008 amending Directive 97/67/EC with regard to the full accomplishment of
the internal market of Community postal services. OJ 2008/L 52/3 of 27 Feb 2008 (3rd Postal
Directive)
Geradin D (2012) Legal opinion on the compatibility of the proposed system of terminal dues with
EU Law
Haller A, Jaag C, Trinkner U (2013) Termination charges in the international parcel market. In:
Crew MA, Kleindorfer PR (eds) Reforming the postal sector in the face of electronic compe-
tition. Edward Elgar, Cheltenham, pp 277–293
Kociubiński J (2014) Consolidation or fragmentation? European competition law in the EU air

transport sector: a policy analyses. 33. Polish Yearbook of International Law, W. N. Scholar,
Warsaw, Poland, pp 189–204
Universal Postal Union (UPU) (2010) Market research on international letters an lightweight
parcels and express mail service items. http://www.upu.int/uploads/tx_sbdownloader/
reportStrategyConferenceMarketResearchEn.pdf. Accessed 22 June 2015
Universal Postal Union, Postal Operations Council (POC) (2014) Inward land rates (ILR) system
review—results of the survey on 5 kg parcel tariffs, Memorandum by the International Bureau,
document: CEP C 3 GRCP 2014.2-Doc 3a. http://documents.upu.int/Bodies/2014/CEP/CEP%
20C%203/CEP%20C%203%20GRCP/MEETING/CEP%20C%203%20GRCP%202014.2/Doc
%203a/EN/cep_c_3_grcp-2_d003a.pdf?Mobile¼1&Source¼%2F%5Flayouts%2Fmobile%
2Fview%2Easpx%3FList%3D9cab4832%252D53b8%252D40f5%252D9ca4%252D8566e0ab
0364%26View%3D895ea13c%252Db413%252D4ce7%252Db5cb%252D912051fecf08%
26RootFolder%3D%252FBodies%252F2014%252FCEP%252FCEP%2520C%25203%25
2FCEP%2520C%25203%2520GRCP%252FMEETING%252FCEP%2520C%25203%252
0GRCP%25202014%252E2%252FDoc%25203a%252FEN%26CurrentPage%3D1.
Accessed 22 June 2015
WIK-Consult (2014) Study for the European Commission DG Internal Market and Services,
Design and development of initiatives to support the growth of e-commerce via better
functioning parcel delivery systems in Europe. Bad Honnef, Germany

The1bst

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

The1bst

Uploaded by

Copyright:

Available Formats

Topics in Regulatory Economics and Policy

More information about this series at http://www.springer.com/series/6646

The Future of the Postal

Topics in Regulatory Economics and Policy

Library of Congress Control Number: 2015956529

Springer Cham Heidelberg New York Dordrecht London

Printed on acid-free paper

Springer International Publishing AG Switzerland is part of Springer Science+Business Media

Timothy J. Brennan Michael A. Crew

IBM Global Business Services

1 Price Cap Regulation and Declining Demand . . . . . . . . . . . . . . . . . 1

11 Compensation Fund Under EU Law: A Suitable

Bruno Basalisco Copenhagen Economics A/S, København K, Denmark

Michael A. Crew CRRI Distinguished Professor of Regulatory Economics,

Dariusz Nehrebecki Oxera Consulting LLP, London, UK

Timothy J. Brennan and Michael A. Crew

T.J. Brennan (*)

© Springer International Publishing Switzerland 2016 1

2 The Implications of Declining Demand for USPS

3 The Declining Demand Problem Under PCR

Fig. 3 The price adjustment to restore the ability to cover cost

4 Price Caps with Changing Demand

understand and assess the magnitude of this elasticity is to calculate it from a

ACðQÞ ¼ CðQÞ=Q ¼ F=Q þ M:

The PCR equation incorporating the adjustment is then

Table 2 Price adjustments for a 10 % fall in demand

5 Relevance Across Regulated Sectors

Table 4 Factors leading to declining demand in regulated sectors

6 Summary and Conclusion

elasticity of demand, can restore solvency to a price-cap-regulated firm. The

For natural monopolies with scale economies throughout, this expression is

Because Q[AC0 ] ¼ MC AC,

And, because MC0 ¼ 0.

2. A general derivation of the price adjustment formula with changes in demand

pk½Dð pÞ ¼ CðkDð pÞÞ; ð1Þ

where C is the total cost function.

Because the elasticity of demand ed ¼ D0 p/D, (3) can be rewritten as

MC=AC ¼ 1 þ eAC ; ð5Þ

Substituting (5) into (4) gives

Multiplying out the terms on the right gives

Subtracting dk/k and ed[dp/p] from both sides of (7) gives

1 > eAC eD : ð10Þ

This is a reasonable assumption. If it is not true, when demand falls, if scale

Armstrong M, Sappington D (2007) Recent developments in the theory of regulation. In:

Antonia Niederprüm, Christian M. Bender, and Alex Kalevi Dieke

© Springer International Publishing Switzerland 2016 19

2 Volume and Price Trends

3 Postal Price Caps in Europe

operations.2 If a positive value is determined for the X-factor, weighted average

Table 1 Price cap models in Europe

determined by decree without a fixed term and cannot be adapted by regulators’

Table 2 Price cap models with X-factors

A striking feature of many recently implemented price cap models is the

1 Introduction and Motivation

© Springer International Publishing Switzerland 2016 29

With Consignia’s announcement of pretax losses of £1.1 billion in June 2002,

2 Developing the Language of Universal Service

References to “universal service” in Parliamentary business, 1952-2004

Source: Millbank Systems, Historic Hansard (2015)

3 International Postal Legislation

3.1 European Postal Legislation: Postal Services Directives

3.2 Universal Postal Union

Universal Postal Service